Gregory Bateson Steps To An Ecology Of Mind, 1987 [608426]

STEPS TO AN
ECOLOGY OF MIND
COLLECTED ESSAYS IN ANTHROPOLOGY,
PSYCHIATRY, EVOLUTION, AND
EPISTEMOLOGY
Gregory Bateson
Jason Aronson Inc.
Northvale, New Jersey
London

Balinese Painting ( Ida Bagus Djati Sura; Batuan, 1937 )
[Analysis, p. 147]

Copyright ® 1972, 1987 by Jason Aronson Inc.
10 9 8 7 6 5 4 3 2 1
All rights reserved. Printed in the United States of America. No
part of this book may be used or reproduced in any manner
whatsoever without written permission from Jason Aronson Inc.
except in the case of brief quotations in reviews for inclusion in
a magazine, newspaper, or broadcast.
Library of Congress Cataloging-in-Publication Data
Bateson, Gregory.
Steps to an ecology of mind.
(Chandler publications for health sciences)
Reprint. Originally published: San Franciso :
Chandler Pub. Co., 1972. With new pref.
Collection of articles reprinted from various
periodicals.
"The published work of Gregory Bateson": p.
Includes index.
1. Anthropology, Cultural—Collected works.
2. Knowledge, Theory of—Collected works. 3. Psychiatry —
Collected works. 4. Evolution — Collected works.
I. Title. II. Series. [DNLM: 1. Anthropology, Cultural—
collected works. 2. Ecology — collected works.
Evolution — collected works. 4. Schizophrenic
Psychology—collected works. 5. Thinking—collected
works. GN 6 B329s 1972a]
GN6.B3 198730687-24133
ISBN 0-87668-950-0

CONTENTS
1971 PREF ACE vi
1987 PREF ACE viii
FORE WORD xi
INTRO DUCTION 1
Part I: Metal ogues ……………………………. …………… 12
Metaloque: Why Do Things Get in a Muddle? 13
Metalogue: Why Do Fr ench men? 19
Metalogue: About Ga mes and Being Seriou s 24
Metalogue: How Mu ch Do You Know? 31
Metalogue: Why Do Things Have Outlin es? 37
Metalogue: Why a Swan? 43
Metaloque: What Is an Instinct? 48
Part II: Form and P attern in Anthrop ology …… 70
Culture Contac t and Schis mogenes is 71
Expe riments in Thinking About Obse rved
Ethno logic al Mate rial 83
Morale and Nationa l Char acter 98
Bali: The Value S ystem of a Steady State 116
Style, Grace , and Information in Primitive Art 137
Comment on Pa rt II 162
Part III: Form and Pathology in Relations hip
………………………………………………… ………………….. 165
Social Planning and the Concept of
Deutero-L earni ng ………… …………………………….. 166
A Theory of Play and Fan tasy 183
Epid emiolog y of a Schizoph renia 199
Toward a Theory of Sch izophren ia 205
The Group D ynamics of Sch izophren ia 233
Minimal Requir ements for a Theory of
Schizophren ia 249
Double Bind, 1969 276
The Logica l Cat egori es of Learn ing a nd Co mmunication 284
The Cybernet ics of "Se lf": A Theory of Alcohol ism 315
Comment on Pa rt III 345

Part IV: Biology a nd E volution …………………… 347
On Empty-Headedness Among Biologi sts and
State Board s of Educat ion 348
The Role of Somatic Change i n Evolu tion 351
Proble ms in Cetacean and Other
Mammalian Communica tion* 369
A Re-exa mination of "Bateson' s Rule" 384
Comment on Part IV 404
Part V: Ep istemol ogy an d Eco logy ……………… 406
Cybernet ic Exp lanat ion 407
Redundanc y and Coding 419
Conscious Purpose versu s Natu re 434
Effects of Consc ious Purpo se on Human Adaptat ion 447
Form, Subs tance , and Difference 455
Comment on Part V 472
Part VI: Crisis i n the Eco logy of Mi nd ……….. 474
From Versailles to Cyberneti cs 475
Patho logies of Epist emology 484
The Roots of Ecolog ical Crisis 494
Ecolog y and Fl exibi lity in Urban C ivilization 499
THE PUB LISHED WORK OF
GREG ORY BATESO N 512

1971 PREFACE
I have been one of Gregory Bateson's students for three years
and I was able to help him select the essays which are here brough t
together for the first time in one volume. I believe that this is a very
important book, not only for those who are profe ssiona lly concerned
with the behavio ral sciences , biolog y, and philosoph y, but also and
especially for those of my generation — the generation born since
Hirosh ima—who are searching for a better unde rstand ing of
themselves a nd th eir world .
The central idea in this book is that we create the world that we
perceive, not because there is no reality outside our heads (the
Indoch inese war is wrong , we are destroying our ecosystem and
therefor e ourselves , whether we believe it or not), but because we
select and edit the reality we see to confor m to our beliefs abou t
what sort of world we live in. The man who believes that the re-
sources of the world are infinite, for example, or that if
something is good for you then the more of it the better, will not
be able to see his errors, because he will not look for evidence of
them.
For a man to change his basic, perception-determining beliefs
— what Bateson calls his epistemologica l premises—he must first
be-come aware that reality is not necessarily as he believes it to
be. This is not an easy or comfortable thing to learn, and most
men in history have probabl y been able to avoid thinking about it.
And I am not convinc ed that the unex amined life is never worth
leading. But sometimes the dissonance between reality and false
beliefs reaches a point when it beco mes impossib le to avoid the
awarenes s that the world no longer makes sense. Only then is it
possible for the mind to consid er radically differen t ideas and
perception s.
Specifically, it is clear that our cultural mind has come to such
a point. But there is danger as well as possib ility in our situation.
There is no guarantee that the new ideas will be an improve ment
over the old. Nor can we hope t hat the change will be smooth.
vi

Already there are psychic casual ties of the culture change . The
psychedel ics are a powerful educat ional tool. They are the surest
way to learn the arb itrarin ess of our ordinary percep tion. Many of us
have had to use them to find out how little we knew. Too many of
us have beco me lost in the labyrinth , have decided that if reality
doesn't mean what we thought it did then there is no meaning in it
at all. I know that place. I have been lost there myself. As far as I
know , there are only two ways out.
One is religious conversion. (I tried Taoism. Others are
choosing various versions of Hindu ism, Buddhi sm, and even
Christianity. And such times always produce a host of self-
proclaimed messiahs. Also, a few of those who study radical
ideologies do so for religious rather than political reasons .) This
solution may satisfy some, al-though there is always the danger of
satanism. But I think that those who choose ready-made systems
of belief lose the chance to do some truly creative think ing, and
perhaps nothing less wi ll save us.
This second way out— thinking things through and taking as lit-
tle as possible on faith— is the more difficult. Intellectual activity
— from science to poetry—has a bad reputa tion in my genera tion.
The blame falls on our so-called educ ation al system, which seems
designed to prevent its victims from learning to think, while
telling them that thinking is what you do when you study a
textbook. Also, to learn to think, you must have a teacher who can
think. The low level of what passes for think ing among most of the
American academic community can perhaps only be appreciated
by contrast with a man like Gregor y Bateson, but it's bad enough to
cause m any of our best minds to give up looking for better.
But the essence of all our problems is bad thinking, and the'
only medicine for that is better thinking. This book is a sample of
the best thinking I've found. I commend it to you, my brothers
and sisters of the new culture, in the hope that it will help us on our
journe y.
—Mark Engel Honolulu, Hawaii April 16, 1971
vii

1987 PREFACE
Gregor y Bate son was fond of quot ing Herac leitus: " Into th e same
river no man can step twice ," particular ly in his later work , in which
he was trying to define the nature of the interface between the realm
of mind and ph ysical reality, and to di scuss the wa y in which mental
process establishes landmarks or thresholds, meanings and
definitions in the world of flux. But a book is like a river, not in
the simple sense of water flowing by, but becau se the intellectual
contex t, like the reader, change s stead ily. Whether one is reading it
for the first time or returning after a lapse of years, Steps to an
Ecolog y of Mind is today not the same book as it was when first
published some fifteen years ago, and for most readers its
impact should be greater. We have changed and the broad
intellectual climate has changed. It would not be fair to say that
this is the more important publication, but it is certainly more
accessible. The increased accessibility of Gregory's thought
today has come about largely because of the steady influenc e of
these essays and other writers drawing on them in the interval, and
because, after recognizing the unity of this collection, Gregory
himself was able to write at a more general level.
The work of Gregor y Bateson has been widely read during this
interven ing period. Ever y ear now I hear of two or three confe rences
focused on some aspec t of his thought , sometimes within a single
discipline, sometimes across a wide r range, and his name crops up
more and more often. Even more signif icantly, many of the ideas
that were most importan t to him have become familia r notions that
we feel at home with. He was one of a group of thinke rs working
toward an understanding of communica tions , of the importance of
self-regula ting systems, and the causa l role of ideas, messages,
differenc es. This has made him a central figure in the growing
appreciation of the importance of looking at events and messages in
context and looking at systems holistically, whether we are concerned
with the health of the human body/mind or the biosphere. The
importance of epistemology is more and more widely understood. At
the same time, much of this familiarity is illusory. Strange or
viii

unsettling ideas are dealt with as the oyster deals with the bit of grit,
packaged in soothing ways, smoothed over. The risk for a reader of
Gregory Bateson in 1972 was that he or she would too readily say,
"This doesn't make sense. It's too obscure for me." The risk today is the
premature claim of understanding, the premature application.
I have had two surprising experiences going back over these arti-
cles: The first was the discovery of how many of the ideas that
seemed important in his later work were already here, although few
will have grasped them completely on first encounter. The second is
how much more still awaits discovery in these articles for one who
has become accustomed to Gregory's thought. Working with Gregory
and writing about him, wrestling together with new ideas . as they
came along, I am probably as much at home here as any of his
students and colleagues, and yet the rereading remains a discovery.
Most of the pieces in this volume are tight, intense, abstract ar-
guments, that Gregory and others labored to "unpack" over the
intervening years; and still there are surprises hidden within them that
become visible as the reader comes to move freely in the text.
Frequently , during his career, as his Introduction indicates,
Gregory felt as if he were speaking and writing in a foreign language.
People did not simply agree or disagree with him; they were
bewildered or intoxicated. Mark Engels, in his 1971 Preface, recog-
nized the analogy between the "mind expanding" experiences of drugs
and religious conversion and the kinds of intellectual change that could
be achieved by a pervasive reshaping of patterns of thought. In
retrospect it strikes me that intoxication and conversion were common
responses even to these abstract and difficult pieces—responses in
which a fraction of the argument was carried on a tide of intuitive
affirmation. Today, however, it is becoming increasingly possible to
come to grips with Gregory's thinking, to select, affirm, contest,
question. Throughout his life, he treasured the relationships in which
he found opportunities for intellectual grappling that went beyond
admiration adulation; critical reading is essential.
This new edition, then, invites readers into an encounter with the
work of Gregory Bateson that was only available to a few when the
collection first appeared. My advice to readers would be to hang on to
the challenge as well as the affirmation. We have not as a civilization
achieved those epistemological shifts that may some day enable nuclear
ix

disarmament, ecological responsibility, and new approaches to both
education and healing that will value and enhance the complexity of
persons in their familial and social setting. In these and in Gregory's
later books (Mind and Nature: A Necessary Unity, Dutton 1979, and,
jointly with me, Angels Fear: Toward an Epistemology of the
Sacred , Macmillan, 1987) the intellectual tools are offered. Today they
will come more readily to hand, be easier to balance and handle in a
disciplined manner than they were in the early 1970s, be more
accessible to practice and skill. But still there remains the challenge of
using the tools in such a way that they be-come a part of the user. And
still the tasks for which these tools have been shaped largely remain to
be done, more urgent today than ever.
—Mary Catherine Bateson Cambridge, Mass. August 1987
x

FOREWORD
Some men seem able to go on working steadily with little
success and no reassurance from outside. I am not one of these. I have
needed to know that somebody else believed that my work had
promise and direction, and I have often been surprised that others
had faith in me when I had very little in myself. I have, at times, even
tried to shrug off the respons ibility which their continued faith
imposed on me by thinking , "But they don't really know what I am
doing. How can they know when I myself do not?"
My first anthropological field work among the Baining of New
Britain was a failure, and I had a period of partial failu60re in research
with dolphins . Neither of these failures has ever been held against
me.
I therefore have to thank many people and institutions for backing
me, at times when I did not consider myself a good bet.
First, I have to thank the Counci l of Fellows of St. John's
College, Cambridge, who elected me to a Fellowship immediately
after my failure among the Baining.
Next, in chronological order, I owe a deep debt to Margaret Mead,
who was my wife and very close co-worker in Bali and New Guinea,
and who since then has continued as a friend and professional
colleague.
In 1942, at a Macy Foundation conference, I met Warren
McCulloch and Julian Bigelow, who were then talking excitedly
about "feedback." The writing of Naven had brought me to the very
edge of what later became cybernetics, but I lacked the concept of
negative feedback. When I returned from overseas after the war, I
went to Frank Fremont-S mith of the Macy Foundation to ask for a
conference on this then-mysterious matter. Frank said that he had just
arranged such a conference with McCulloch as chair-man. It thus
happ ened that I was privileged to be a member of the famous Macy
Conferences on Cybernetics. My debt to Warren McCulloch, Norbert
Wiener, John von Neumann, Evelyn Hutchinson , and other members
of these conferen ces is eviden t in everything that I have written since
World War II.
xi

In my first attempts to synthesize cybernetic ideas with anthropo-
logical data, I had the benefit of a Guggenheim Fellowship.
In the period of my entry into the psychiat ric field, it was Jurgen
Ruesch, with whom I worked in the Langley Porter Clinic, who ini-
tiated me into many of the curious features of the psychiatric world.
From 1949 t o 1962 , I had the title of "Ethno logist" in the Veterans
Administration Hospital at Palo Alto, where I was given singular
freedom to study whatever I thought interesting. I was protected from
outside demands and given this freedom by the director of the
hospital, Dr. John J. Prusmack.
In this period, Bernard Siegel suggested that the Stanford Uni-
versity Press repub lish my book, Naven, which had fallen flat on its
face when first published in 1936; and I was lucky enough to get film
footage of a sequence of play between otters in the Fleishhacker Zoo
which seemed to me of such theoretical interest as to justify a small
research program.
I owe my first research grant in the psychiatric field to the late
Chester Barnard of the Rockefeller Foundation, who had kept a copy
of Naven for some years by his bedside. This was a grant to study
"the role of the Paradoxe s of Abstra ction in Communica tion."
Under this grant, Jay Haley, John Weakland, and Bill Fry joined
me to form a small research team within the V.A. Hospital.
But again there was failure. Our grant was for only two years,
Cheste r Barna rd had retired, and in the opinion of the Foundation
staff we did not have enough results to justify renewal. The grant ran
out, but my team loyally stayed with me without pay. The work
went on, and, a few days after the end of the grant, while I was writ-
ing a despera te letter to Norbe rt Wiener for his advice about wher e
to get the next grant, the double bind hypothesis fell into place.
Finally Frank Fremont-Smith and the Macy Foundation saved us.
After that there were grants from the Foundations Fund for Psy-
chiatry and from the National Institute of Mental Health.
Gradual ly it appeared that for the next advances in the study of
logical typing in communication I should work with animal material,
and I started to work with octopus . My wife, Lois, worked with me,
and for over a year we kept a dozen octopuses in our living room.
This preliminary work was promising but needed to be repeated and
extended under better conditions. For this no grants were available.
xii

At this point, John Lilly came forward and invited me to be the
director of his dolphin laboratory in the Virgin Islands. I worked there
for about a year and became interested in the probl ems of cetacean
communica tion, but I think I am not cut out to administer a
labora tory dubious ly funded in a place where the logistics are intol-
erably difficult.
It was while I was struggling with these problems that I received a
Career Development Award under the National Institute of Mental
Health. These awards were administered by Bert Boothe, and I owe
much to his continued faith and interest.
In 1963, Taylor Pryor of the Oceanic Foundation in Hawaii invited
me to work in his Oceanic Institute on cetacean and other problems of
animal and human communication. It is here that I have written more
than half of the present book, including the whole of Part V.
While in Hawaii , I have al so been working rec ently with the Cul-
ture Learn ing Institute of the East-West Cente r in the Unive rsity of
Hawaii, and owe some theoretical insights regarding Learning III to
discussions held in that Institute.
My debt to the Wenne r-Gren Foundation is evident from the fact
that the book contains no less than four position pape rs written for
Wenner-Gren conferences. I wish also to thank personally Mrs. Lita
Osmundsen, the Director of Research of that Foundation.
Many also have labored along the road to help me. Most of these
cannot be mentioned here, but I must particula rly thank Dr. Vern
Carroll, who prepared the bibliography, and my secretary, Judith Van
Slooten, who labored with accuracy through long hours in preparing
this book for press.
Finally there is the debt that every man of scien ce owes to the gi-
ants of the past. It is no mean comfort, at times when the next idea
cannot be found and the whole enterprise seems futile, to remember that
greater men have wrestled with the same problems. My personal
inspiration has owed much to the men who over the last 200 years
have kept alive the idea of unity between mind and body: Lamarck,
the founder of evolutionary theory, miserable, old, and blind, and
damned by Cuvier, who believed in Special Creation; William Blake,
the poe t and pa inter, who saw " through his eyes, no t with them," and
knew more abou t what it is to be human than any other man;
Samuel Butler, the ablest contemporar y critic of Darwinian
xiii

evolu tion and the first analyst of a schizophrenog enic family; R. G.
Collingwood, the first man to recognize—and to analyze in crys-
talline prose—the nature of context; and William Bateson, my father,
who was certainly ready in 1894 to receive the cybernetic ideas.
Selection and Arrangement o f Items
The book contains almost everything that I have written, with the
excep tion of items too long to be included , such as books and ex-
tensive analyses of data; and items too trivial or ephe meral, such as
book reviews and controversial notes. A complete personal bibliog-
raphy is appended.
Broadl y, I have been conce rned with four sorts of subject matter:
anthropology, psychiatry, biological evolution and genetics, and the
new epistemology which comes out of systems theory and ecology.
Essays on th ese subj ects m ake up Pa rts II, III, IV, and V of the book ,
and the order of these parts corresponds to the chronolog ical order
of four overl apping period s in my life in which these subjec ts have
been central to my thinking. Within each part, the essays are in
chronological order.
I recognize that readers are likely to attend most carefully to those
parts of the book dealing with their particular subjects. I have
therefor e not edited out some repetition. The psychiatrist interes ted
in alcoholism will encounter in "The Cybernetics of `Self' " ideas
which appear again in more philosoph ic dress in "Form, Subs tance ,
and Difference."
Ocean ic Institute, Hawaii Apra 16, 1971
xiv

INTRODUCTION
The Science of Mind and O rder*
The title of this book of collected essays and lectures is intended
precis ely to define the contents. The essays, spread over thirty-five
years, combine to propose a new way of thinking abou t ideas and
abou t those aggrega tes of ideas which I call "minds." This way of
thinking I call the "ecolog y of mind," or the ecolog y of ideas. It is a
science which does not yet exist as an organized body of theory or
knowledg e.
But the definition of an "idea" which the essays combine to pro-
pose is much wider and more formal than is conv ention al. The es-
says must speak for themselves, but here at the beginn ing let me
state my belief that such matters as the bilateral symmetry of an animal,
the patterned arrang ement of leaves in a plant, the escalation of an
armaments race, the processes of courtship , the nature of play, the
grammar of a sentence, the mystery of biolog ical evolution, and the
contemporar y crises in man's relationship to him envi-
ronment, can only be under stood in terms of such an ecolog y of
ideas as I propose.
The question s which the book raises are ecolog ical: How do
ideas interact? Is there some sort of natur al selection which
determines the surviva l of some ideas and the extinction or death of
others ? What sort of econo mics limits the multiplicity of ideas in a
given region of mind? What are the necessar y cond itions for
stability (or surviva l) of such a system or sub system?
Some of these questions are touched upon in the essays, but the
main thrust of the book is to clear the way so that such questions can be
meaningfully asked.
It was only in late 1969 that I became fully conscious of what I had
been doing. With the writing of the Korzybski Lecture, "Form,
* This essay, written in 1971, has not been published else- where.
1

Substance, and Difference," I found that in my work with primitive
peoples, schizophrenia, biological symmetry, and in my discontent with
the conventional theories of evolution and learning, I had identified a
widely scattered set of bench marks or points of reference from which a
new scientific territory could be defined. These bench marks I have
called "steps" in the title of the book.
In the nature of the case, an explorer can never know what he is
exploring until it has been explored. He carries no Baedeker in his
pocket, no guidebook which will tell him which churches he should
visit or at which hotels he should stay. He has only the ambiguous
folklore of others who have passed that way. No doubt deeper levels of
the mind guide the scientist or the artist toward experiences and
thoughts which are relevant to those problems which are somehow his,
and this guidance seems to operate long before the scientist has any
conscious knowledge of his goals. But how this happens we do not
know.
I have often been impatient with colleagues who seemed unable to
discern the difference between the trivial and the profound. But when
students have asked me to define that difference, I have been struck
dumb. I have said vaguely that any study which throws light upon the
nature of "order" or "pattern" in the universe is surely nontrivial.
But this answer only begs the question.
I used to teach an informal course for psychiatric residents in the
Veterans Administration Hospital at Palo Alto, trying to get them to
think some of the thoughts that are in these essays. They would attend
dutifully and even with intense interest to what I was saying, but every
year the question would arise after three or four sessions of the class:
"What is this course all about?"
I tried various answers to this question. Once I drew up a sort of
catechism and offered it to the class as a sampling of the questions
which I hoped they would be able to discuss after completing the
course. The questions ranged from "What is a sacrament?' to "What is
entropy?" and "What is play?"
As a didactic maneuver, my cathechism was a failure: it silenced the
class. But one question in it was useful:
A certain mother habitually rewards her small son
with ice cream after he eats his spinach. What
2

additional information would you need to be able to
predict whether the child will: a. Come to love or hate
spinach, b. Love or hate ice cream, or c. Love or hate
Mother?
We devoted one or two sessions of the class to exploring the many
ramification s of this question, and it became clear to me that all the
needed additional information conce rned the context of the mother's
and son's behavior. In fact, the phenomenon of context and the closely
related pheno menon of "meaning" defined a division between the
"hard" s cienc es and the sort of science which I was trying to build.
Gradually I discovered that what made it difficult to tell the class
what the course was about was the fact that my way of thinking was
different from theirs. A clue to this difference came from one of the
students. It was the first session of the class and I had talked about the
cultural differences between England and America—a matter which
should always be touched on when an Englishman must teach
Americans about cultural anthropology. At the end of the session, one
resident came up. He glanced over his shoulder to be sure that the
others were all leaving, and then said rather hesitantly, "I want to ask a
question." "Yes." "It's—do you want us to learn what you are telling
us?" I hesitated a moment, but he rushed on with, "Or is it all a sort of
example, an illustration of something else?" "Yes, indeed!"
But an example of what?
And then there was, almost every year, a vague complaint which
usually came to me as a rumor. It was alleged that "Bateson knows
something which he does not tell you," or "There's something be-hind
what Bateson says, but he never says what it is."
Evidently I was not answering the question, "An example of what?"
In desperation, I constructed a diagram to describe what I conceive
to be the task of the scientist. By use of this diagram, it became clear that
a difference between my habits of thought and those of my students
sprang from the fact that they were trained to think and argue
inductively from data to hypotheses but never to test hypotheses
against knowledge derived by deduction from the fundamentals of
science or philosophy.
The diagra m had three columns. On the left, I listed various sorts
of uninterpreted data, such as a film record of human or animal
3

behavio r, a description of an expe riment, a description or pho-
tograph of a beetle's leg, or a recorded human utterance . I stressed
the fact that "data" are not events or objec ts but always records or
descriptions or memories of even ts or objects. Alwa ys there is a
transfor mation or recoding of the raw event which interv enes
between the scientist and his object. The weight of an object is meas-
ured against the weight of some other object or registered on a me-ter.
The human voice is transformed into variable magnetizations of tape.
Moreover, always and inevitably, there is a selection of data because
the total universe, past and present, is not subject to observation from
any given observer's position.
In a strict sense, therefore, no data are truly "raw," and every
record has been somehow subjected to editing and transformation
either by man or by his instruments.
But still the data are the most reliable source of information, and
from them the scientist must start. They provide his first inspiration and
to them he must later return.
In the middle column, I listed a number of imperfectly defined
explanatory notions which are commonly used in the behavioral
sciences—"ego," "anxiety," "instinct," "purpose," "mind," "self,"
"fixed action pattern," "intelligence," "stupidity," "maturity," and the
like. For the sake of politeness, I call these "heuristic" concepts; but, in
truth, most of them are so loosely derived and so mutually irrelevant
that they mix together to make a sort of conceptual fog which does
much to delay the progress of science.
In the right-hand column, I listed what I call "fundamentals."
These are of two kinds: propositions and systems of propositions
which are truistical, and propositions or "laws" which are generally
true. Among the truistical propositions I included the "Eternal Verities"
of mathematics where truth is tautologically limited to the do-mains
within which man-made sets of axioms and definitions obtain: "If
numbers are appropriately defined and if the operation of addition is
appropriately defined; then 5 + 7 = 12." Amo ng propositions which I
would describe as scientifically or generally and empirically true, I
would list the conservation "laws" for mass and energy, the Second
Law of Thermodynami cs, and so on. But the line between tautological
truths and empirical generalizations is not sharply definable, and,
among my "fundamentals," there are many propositions whose truth no
4

sensible man can doubt but which can-not easily be classified as either
empirical or tautological. The "laws" of probability cannot be stated so
as to be understood and not be believed, but it is not easy to decide
whether they are empirical or tautological; and this is also true of
Shannon's theorems in Information Theory.
With the aid of such a diagr am, much can be said abou t the
whole scientific endeavo r and abou t the position and direction of
any particular piece of inquiry within it. "Exp lanation" is the
mapping of data onto funda mentals, but the ultimate goal of science
is the increase o f fund amental knowledge .
Many investigators, especially in the behavioral sciences, seem to
believe that scientific advance is predominantly inductive and should
be inductive. In terms of the diagram, they believe that progress is
made by study of the "raw" data, leading to new heuristic concepts.
The heuristic concepts are then to be regarded as "working
hypotheses" and tested against more data. Gradually, it is hoped, the
heuristic concepts will be corrected and improved until at last they are
worthy of a place in the list of fundamentals. About fifty years of work
in which thousands of clever men have had their share have, in fact,
produced a rich crop of several hundred heuristic concepts, but, alas,
scarcely a single principle worthy of a place in the list of
fundamentals.
It is all too clear that the vast majority of the concepts of contem-
porary psychology, psychiatry, anthropology, sociology, and eco-
nomics are totally detached from the network of scientific fun-
damentals.
Moliere, long ago, depicted an oral doctoral examination in which
the learned doctors ask the candidate to state the "cause and reason"
why opium puts people to sleep. The candidate triumphantly answers
in dog Latin, "Because there is in it a dormitive principle (virtus
dormitiva)."
Characteristically, the scientist confronts a complex interactive
system— in this case, an interaction between man and opium. He
observes a change in the system — the man falls asleep. The scientist
then explains the change by giving a name to a fictitious "cause," lo-
cated in one or other component of the interacting system. Either the
opium contains a reified dormitive principle, or the man contains a
5

reified need for sleep, an adormitosis, which is "expressed" in his
response to opium.
And, characteristically, all such hypotheses are "dormitive" in the
sense that they put to sleep the "critical faculty" (another reified
fictitious cause) within the scientist himself.
The state of mind or habit of though t which goes from data to
dormitive hypothesi s and back to data is self-reinfo rcing . There is,
among all scientists, a high value set upon prediction, and, indeed, to
be able to predi ct pheno mena is a fine thing. But predi ction is a
rather poor test of an hypothesis, and this is especially true of "dor-
mitive hypotheses ." If we assert that opium contains a dormitive
principle, we can then devot e a lifetime of research to studying the
characteristics of this princip le. Is it heat-s table? In whi ch fraction of
a distillate is it located? What is its molecular formula? And so on.
Many of these questions will be answerable in the laboratory and will
lead on to derivative hypotheses no less "dormitive" than that from
which we started.
In fact, the multiplication of dormitive hypotheses is a symptom of
excessive preference for induction, and this preference must al-ways
lead to something like the present state of the behavioral sciences— a
mass of quasi-theoretical speculation unconnected with any core of
fundamental knowledge.
In contrast, I try to teach students— and this collection of essays is
very much concerned with trying to communicate this thesis—that in
scientific research you start from two beginnings, each of which has its
own kind of authority: the observations cannot be denied, and the
fundamentals must be fitted. You must achieve a sort of pincers
maneuver.
If you are surveying a piece of land, or mapping the stars, you have
two bodies of knowledge, neither of which can be ignored. There are
your own empirical measurements on the one hand and there is
Euclidean geometry on the other. If these two cannot be made to fit
together, then either the data are wrong or you have argued wrongly
from them or you have made a major discovery leading to a revision of
the whole of geometry.
The would-be behavioral scientist who knows nothing of the basic
structure of science and nothing of the 3000 years of careful philosophic
and humanistic thought about man — who cannot define either entropy
6

or a sacrament —had better hold his peace rather than add to the existing
jungle of half-baked hypotheses.
But the gulf between the heuristic and the fundamental is not solely
due to empiricism and the inductive habit, nor even to the seductions of
quick application and the faulty educational system which makes
professional scientists out of men who care little for the fundamental
structure of science. It is due also to the circumstance that a very large
part of the fundamental structure of nineteenth-century science was
inappropriate or irrelevant to the problems and phenomena which
confronted the biologist and behavioral scientist.
For at least 200 years, say from the time of Newton to the late
nineteenth century, the dominant preoccupation of science was with
those chains of cause and effect which could be referred to forces and
impacts. The mathematics available to Newton was preponderantly
quantitative, and this fact, combined with the central focus upon forces
and impacts, led men to measure with remarkable accuracy quantities of
distance, time, matter, and energy.
As the measurements of the surveyor must jibe with Euclidean
geometry, so scientific thought had to jibe with the great conservative
laws. The description of any event examined by a physicist or chemist
was to be founded upon budgets of mass and energy, and this rule gave
a particular kind of rigor to the whole of thought in the hard sciences.
The early pioneers of behavioral science not unnaturally began their
survey of behavior by desiring a similar rigorous base to guide their
speculations. Length and mass were concepts which they could hardly
use in describing behavior (whatever that might be), but energy seemed
more handy. It was tempting to relate "energy" to already existing
metaphors such as "strength" of emotions or character or "vigor." Or to
think of "energy" as somehow the opposite of "fatigue" or "apathy."
Metabolism obeys an energy budget (within the strict meaning of
"energy"), and energy expended in behavior must surely be included in
this budget; therefore it seemed sensible to think of energy as a
determinant of behavior.
It would have been more fruitful to think of lack of energy as pre-
ventive of behavior, since in the end a starving man will cease to be-
have. But even this will not do: an amoeba, deprived of food, be-comes
for a time more active. Its energy expenditure is an inverse function of
energy input.
7

The nineteeth-century scientists (notably Freud) who tried to es-
tablish a bridge between behavioral data and the fundamentals of
physical and chemical science were, surely, correct in insisting upon the
need for such a bridge but, I believe, wrong in choosing "energy" as the
foundation for that bridge.
If mass and length are inappropriate for the describing of behavior,
then energy is unlikely to be more appropriate. After all, energy is Mass
x Velocity2, and no behavioral scientist really insists that "psychic
energy" is of these dimensions.
It is necessary, therefore, to look again among the fundamentals for
an appropriate set of ideas against which we can test our heuristic
hypotheses.
But some will argue that the time is not yet ripe; that surely the
fundamentals of science were all arrived at by inductive reasoning from
experience, so we should continue with induction until we get a
fundamental answer.
I believe that it is simply not true that the funda mentals of sci-
ence began in induct ion from expe rienc e, and I sugge st that in the
search for a bridgehead among the fundamentals we should go back to
the ve ry beginn ings of sc ientific and phi losophic t hought ; certainly to
a period before science, philosophy, and religion had be-come separate
activities separately pursued by professionals in separate disciplines.
Consider, for example, the central origin myth of Judaeo-Christian
peoples. What are the fundamental philosophic and scientific problems
with which this myth is concerned?
In the beginning God created the heaven and the earth.
And the earth was without form, and void; and darkness was
upon the face of the deep. And the Spirit of God moved upon
the face of the waters.
And God said, Let there be light: and there was light. And God
saw the light, that it was good: and God divided the light from
the darkness. And God called the light Day, and the darkness he
called Night. And the evening and the morning were the first
day.
8

And God said, Let there be a firmament in the midst of the
waters, and let it divide the waters from the waters. And God
made the firmament, and divided the waters which were under
the firmament from the waters which were above the
firmament: and it was so. And God called the firmament
Heaven. And the evening and the morning were the second
day.
And God said, Let the waters under the heaven be
gathered together unto one place, and let the dry land
appear: and it was so. And God called the dry land Earth;
and the gathering together of the waters called he Seas: and
God saw that it was good.
Authorized version
Out of these first ten verse s of thunde rous prose, we can draw
some of the premises or funda mentals of ancient Chalde an though t
and it is strang e, almost eerie, to note how m any of the funda mentals
and proble ms of modern science are foreshadowed in the ancient
docu ment.
(1) The problem of the origin and nature of matter is summarily
dismissed .
(2) The passage deals at length with the proble m of the origin of
order.
(3) A separ ation is thus genera ted between the two sorts of prob-
lem. It is possible that this separa tion of proble ms was an error, but
—error or not— the separa tion is maintained in the funda mentals of
modern science. The conserva tive laws for matter and energy are
still separate from the laws of order, negat ive entropy, and
information.
(4) Orde is seen as a matter of sorting and dividing. But the essential
notion in all sorting is that some difference shall cause some other
difference at a later time. If we are sorting black balls from white balls,
or large balls from small balls, a difference among the balls is to be
followed by a difference in their location—balls of one class to one sack
and balls of another class to another. For such an operation, we need
9

something like a sieve, a threshold, or, par excellence, a sense organ.
It is understandable, therefore, that a perceiving Entity should have
been invoked to perform this function of creating an otherwise
improbable order.
(5) Close ly linked with the sorting and dividing is the mystery of
classification, to be followed later by the extraord inary human
achievement o f naming.
It is not at all clear that the various components of this myth are all
produc ts of inductive reasoning from experience . And the mat-ter
beco mes still more puzzling when this orig in myth is co mpared with
others which embody different funda mental premises.
Among the Iatmul of New Guine a, the central origin myth, like
the Genesis story, deals with the question of how dry land was separated
from wate r. They say that in the beginning the crocod ile
Kavwokmali paddled with his front legs and with his hind legs; and his
paddl ing kept the mud suspended in the wate r. The great culture
hero, Keve mbuangga, came with his spear and killed Kavwok mali.
After that the mud settled and dry land was formed. Keve mbuangga
then stamped with his foot on the dry land, i.e., he proudl y
demonstra ted "that it was good."
Here there is a stronger case for deriving the myth from exper i-
ence combined with induct ive reason ing. After all, mud does re-
main in suspension if randomly stirred and does settle when the stir-ring
ceases. Moreove r, the Iatmul people live in the vast swamps of the
Sepik Rive r valley where the separ ation of land from water is
imperfect. It is under standab le that they might be interested in the
differen tiation of land f rom wate r.
In any case, the Iatmul have arrived at a theory of order which is
almost a precise converse of that of the book of Genes is. In Iatmul
though t, sorting will occur if rando mization is prevented . In Gene-
sis, an agen t is invoked to do the sorting and divid ing.
But both cultures alike assume a fundamental division between the
probl ems of material creation and the proble ms of order and
differen tiation.
Return ing now to the question of whether the funda mentals of
science and/or philosoph y were , at the primitive level, arrived at by
induc tive reasoning from empirica l data, we find that the answer is
not simple. It is difficult to see how the dichoto my between
10

substance and form could be arrived at by inductive argument. No
man, after all, has ever seen or experienced formless and unsorted
matter; just as no man has ever seen or experienced a "random" event.
If, therefore, the notion of a universe "without form and void" was
arrived at by induction, it was by a monstrous—and perhaps
erroneous — jump of extrapolation.
And even so, it is not clear that the starting point from which the
primitive philosophers took off was observation. It is at least equally
likely that dichoto my between form and substan ce was an uncon-
scious deduct ion from the subject-predicate relation in the structure of
primitive langu age. This, howeve r, is a matter beyond the reach of
useful specu lation.
Be that as it may, the central—but usually not explicit — subject
matter of the lectures which I used to give to psychiat ric residen ts
and of these essays is the bridge between behav ioral data and the
"funda mental s" of sc ience and ph ilosoph y; and m y critical comments
above about the metaphoric use of "energy" in the behavioral science s
add up to a rather simple accusat ion of many of my colleagues, that
they have tried to build the bridge to the wrong half of the ancient
dichoto my between form and substanc e. The conservative laws for
energy and matter concern substance rather than form. But mental
process , idea s, communicat ion, organiz ation , differentiation , pattern,
and so on, are matters of form rather than subst ance.
Within th e bod y of funda mentals, that h alf which d eals wi th for m
has been dramatica lly enriched in the last thirty years by the
discover ies of cyberne tics and systems theory. This book is con-
cerned with building a bridge between the facts of life and behavior and
what we know t oday of the nature of pa ttern and order.
11

Part I: Metalogues
DEFINITION: A metalogue is a conversation about some
problematic subject. This conversation should be such that not
only do the participants discuss the problem but the structure of the
conversation as a whole is also relevant to the same subject. Only
some of the conversations here presented achieve this double
format.
Notabl y, the history of evolu tiona ry theor y is inevitably a
metalogu e between man and nature, in which the creation and
interac tion of ideas must necessa rily exemplif y evolutionar y
proce ss.

Metaloqu e: Why Do Things Get in a
Muddle?*
Daught er: Dadd y, why do things g et in a muddle?
Father:What do you mean? Things? Mudd le?
D:Well, people spend a lot of time tidying thing s, but they neve r
seem to spend time muddling them. Things just seem to get in a
muddle by themselves . And then peop le have to tidy them up
again.
F:But do your things g et in a muddle if you don' t touch t hem?
D:No—not if nobod y touche s them. But if you touch them—or if
anybody touches them—the y get in a muddle and it's a worse
muddle if it isn't me.
F:Yes—that's why I try to keep you from touching the things on m y
desk. Because my things get in a worse muddle if they are
touched b y somebod y who i sn't me.
D:But do peop le always muddle other people's things ? Why do
they, Dadd y?
F:Now, wait a minute. It's not so simple. First of all, what do you
mean by a muddle?
D:I mean—so I can't find things , and so it looks all muddled up.
The way it is when no thing is straigh t
F:Well, but are you sure you mean the same thing by muddle that
anybody else would m ean?
D:But, Dadd y, I'm sure I do—because I'm not a very tidy person
and if I say thing s are in a muddle, then I'm sure everybody else
would agree with me.
F: All right—bu t do you think you mean the same thing by "tidy"
that. other people would ? If your mummy makes your things
tidy, do you know where to find them?
D: Hmm . . . sometimes—becaus e, you see, I know where she puts
things when she tidies up
F: Yes, I try to keep her away from tidying my desk, too. I'm sure
that she and I don' t mean the same thing b y "tidy."
* Written in 1948; not previousl y published.
13

D: Dadd y, do y ou and I m ean the sa me thing b y "tidy?" F: I doubt i t,
my dear— I doubt it.
D: But, Dadd y, isn't that a funny thing—tha t everybody means the
same when they say "muddled" but ever y-body means something
differen t by "tidy." But "tidy" is the opposi te of "muddled ," isn't
it?
F: Now we begin to get into more difficult quest ions. Let's start
again from the beginning . You said "Why do things always get in
a muddle?" Now we have made a step or two—and let's change
the question to "Why do things get in a state which Cath y calls
'not tidy?' " Do y ou see wh y I want to make th at change?
D: … Yes, I think so—bec ause if I have a specia l meaning for "tidy"
then some of other peop le's "tidies" will look like muddles to me
—even if we do agree about most of what we ca ll muddles
F: That's right. Now—let's look at what you call tidy. When your
paint box is put in a tidy place, wher e is it? D: Here on the end of
this shelf.
F:Okay—now if it were anywhere e lse?
D:No, t hat would no t be tidy.
F:What abou t the other end of the she lf, here? L ike this?
D:No, that's not where it belongs , and anyhow it would have to be
straigh t, not all crooked t he wa y you put it.
F:Oh—in t he right place a nd straight.
D:Yes.
F:Well, that means that there are only very few places which are
"tidy" for your paint box
D:Only one place—
F: No—ve ry few places , because if I move it a little bit, like this, it
is still tidy.
D: All right—bu t very, very few p laces .
F: All right, very, very few places. Now what about the teddy bear
and your doll, and the Wizard of Oz and your swea ter, and your
shoes? It's the same for all the things , isn't it, that each thing has
only a very, very few places which a re "tidy" for that thing?
D: Yes, Dadd y—but the Wizard of Oz could be any-where on that
shelf. And Dadd y—do you know what? I hate, hate it when my
books g et all mixed up with your books and Mummy's books.
F : Yes, I know . (Pause )
14

D: Dadd y, you didn' t finish. Why do my things get the way I say
isn't tidy?
F: But I have finish ed—it 's just becaus e there are more ways which
you call "un tidy" than there are wa ys which y ou call "tidy."
D: But t hat isn't a reason wh y
F: But, yes, it is. And it is the real and only and very importan t
reason .
D: Oh, Dadd y! Stop i t.
F: No, I'm not fooling. That is the reason, and all of science is
hooked up with that reason. Let's take an-othe r example. If I put
some sand in the bottom of this cup and put some sugar on the
top of it, and now stir it with a teaspoon , the sand and the sugar
will get mixed up , won't they?
D: Yes, but, Dadd y, is it fair to shift over to talking about "mixed
up" when we s tarted wi th "muddled up?"
F: Hmm . . . I wonder . .. but I think so—Y es—bec ause let's say we
can find somebod y who thinks it is more tidy to have all the sand
unde rneath all the sugar . And i f you like I'll say I want i t that way
D: Hmm…
F: All right— take another example. Sometimes in the movies you
will see a lot of letters of the alphab et all scattered over the
screen , all higgledy-piggledy and some even upside down. And
then something shakes the table so that the letters start to move,
and then as the shaking goes on, the letters all come together to
spell the title of the film.
D: Yes, I've seen that—the y spelled DONALD.
F: It doesn't matter what they spelled. The point is that you saw
something being shaken and stirred up and in-stead of getting
more mixed up than before, the letters came together into an
order, all right way up, and spelled a word—the y made up
something whi ch a lot of peopl e would agree is sense.
D: Yes, Dadd y, but you know .. .
F: No, I don't know; what I am trying to say is that in the real world
things never happen that way. It's only in the m ovies.
D: But, Dadd y .. .
F: I tell you it's only in the movies that you can shake things and
they seem to take on m ore orde r and sense t han th ey had befor e ..
.
15

D: But , Dadd y .. .
F: Wait till I've finished this time . . . And they make it look like that
in the movies by doing the whole thing backwards . They put the
letters all in order to spell DONALD and then they start the
camera and t hen t hey start shaking the table.
D: Oh, Dadd y—I knew that and I did so want to tell you that—and
then when they run the film, they run it backward s so that it
looks as though things had happened forwards . But really the
shaking happened back -wards. And they have to photogr aph it
upside down . .. Why do they, Dadd y?
F: Oh God.
D: Why do they have to fix the camera ups ide down, Dadd y?
F: No, I won't answer that question now becaus e we're in the middle
of the quest ion abou t muddles.
D: Oh—al l right, but don't forget, Dadd y, you've got to answer that
quest ion about the camera anothe r day. Don't forget! You won't
forget, will you, Dadd y? Be-cause I may not remember. Please,
Dadd y.
F: Okay—but another day. Now, where were we? Yes, about things
never happen ing backwa rds. And I was trying to tell you why it
is a reason for things to hap-pen in a certain way if we can show
that that way has more wa ys of happen ing than s ome other way.
D: Dadd y—don't begin talking nonsens e.
F:I'm not talking nonsense . Let's start again. There's only one way
of spelling DONALD. Agre ed?
D: Yes.
F: All right. And there are millions and millions and mil-lions of
ways of scattering s ix letters on the table. Agreed?
D: Yes. I suppos e so. Can some of these be upside down?
F: Yes—just in the sort of higgledy-piggled y muddle they were in in
the film. But there could be million s and millions and million s of
muddles l ike that, couldn't there? And only one DO NAL D?
D: All right— yes. But, Dadd y, the same letters might spell OLD
DAN.
F: Neve r mind. The movie peop le don't want them to spell OLD
DAN. They only want DO NAL D.
D: Why do they?
F: Da mn the m ovie peopl e.
16

D: But y ou mentioned them first, Dadd y.
F: Yes—but that was to try to tell you why things h appen t hat way in
which there are most ways of their happening. And now it's your
bedtime.
D: But, Dadd y, you neve r did finish telling me why things happen
that way—the wa y that has most wa ys.
F: All right. But don't start any more hares running—on e is quite
enough . Anyhow, I am tired of DONAL D, let's take anothe r
example. Let's take tossing penni es.
D: Dadd y? Are you still talking abou t the same question we started
with? "Wh y do things get in a muddle?"
F: Yes.
D: Then, Dadd y, is what you are trying to say true about penn ies,
and about DONAL D, and about sugar and sand, and about my
paint box, and a bout p ennies ?
F: Yes—th at's right.
D: Oh—I was just wondering , that's all.
F: Now, let's see if I can get it said this time. Let's go back to the
sand and the sugar, and let's suppose that somebody says that
having the sand a t the bo ttom is "tidy" or "order ly."
D: Dadd y, does somebod y have to say something like that before
you can go on to talk about how things are going to get mixed up
when y ou stir them?
F:Yes—that's just the point. They say what they hope will happen
and then I tell them it won't happen because there are so many
other things that might happen. And I know that it is more likely
that one o f the many things will happen and not on e of the few.
D: Daddy, you're just an old bookmaker, backing all the other
horses against the one horse that I want to bet on.
F: That's right, my dear. I get them to bet on what they call the
"tidy" way—I know that there are infinitely many muddled
ways—so things will always go toward muddle and mixedness.
D: But why didn't you say that at the beginning, Daddy? I could
have understood that all right.
F: Yes, I suppose so. Anyhow, it's now bedtime.
D: Daddy, why do grownups have wars, instead of just fighting the
way children do?
17

F: No—bedtime. Be off with you. We'll talk about wars another
time.
18

Metalogu e: Why Do Fren chmen?*
Daught er: Dadd y, why do French men wave th eir arms about?
Father: What do you mean?
D: I mean when t hey talk. Why do th ey wave t heir arms and all that?
F: Well—wh y do you smile? Or why do you stamp your foot
sometimes?
D: But that's not the same thing , Dadd y. I don't wave my arms abou t
like a French man does. I don't believe they can stop doing it,
Dadd y. Can t hey?
F: I don't know—the y might find it hard to stop. . . . Can you stop
smiling?
D: But Dadd y, I don't smile all the time. It's hard to stop when I feel
like smiling. But I don' t feel like it all the time. And t hen I stop.
F: That's true—bu t then a French man doesn't wave his arms in the
same way all the t ime. So metimes he waves the m in one wa y and
sometimes in another—and sometimes, I think, he stops waving
them.
* * *
F: What do you think? I mean, what does it make you think when a
French man waves h is arms?
D: I think it looks silly, Dadd y. But I don't suppose it looks like that
to another French man. They cannot all look silly to each other.
Becau se if they did, they would s top it. Wouldn't they?
F: Perhaps—bu t that is not a very simple quest ion. What else do
they make you think?
D: Well—the y look all exc ited .. .
F: All right—"s illy" and "excited."
D: But are they really as excited as the y look? If I were as excited as
that, I would want to danc e or sing or hit somebod y on the nose
* This metalogue is reprinted from Impulse 1951, an annual of contemporar y
dance , by permission of Impulse Publications , Inc. It has also appeared in ETC.: A
Re-view of General Semantics, Vol. X, 1953.
19

… but they just go on waving their arms. They can't be really
excited.
F:Well—are they really as silly as they look to you? And anyhow,
why do you sometimes want to danc e and sing and punch
somebod y on the no se?
D:Oh. So metimes I just feel like that.
F:Perhaps a French man just feels "like that" when he waves his
arms about.
D: But he couldn' t feel like that all the time, Dadd y, he just couldn' t.
F: You mean—the French man surely does not feel when he waves
his arms exactly as you would feel if you waved yours. And
surely you are right.
D: But , then, how does he feel?
F: Well—l et us suppos e you are talking to a Frenchman and he is
waving his arms about , and then in the middle of the
conver sation, after something that you have said, he suddenl y
stops waving his arms, and just talks. What would you think
then? That he had j ust stopped being s illy and e xcited?
D: No . . . I'd be frightened . I'd think I had said something that hurt
his feelings and perhaps he might b e really angry.
F: Yes—and y ou might be right.
* * *
D: All right— so they stop waving their arms when they start being
angry.
F: Wait a minute. The quest ion, after all, is what does one
French man tell another French man by waving his arms? And we
have part of an answer—he tells him something abou t how he
feels about t he other guy. He tells him he is not seriou sly angry—
that he is willing and a ble to be what you call "silly."
D: But—no—tha t's not sensib le. He cannot do all that work so that
later he will be able to tell the other guy that he is angry by just
keeping hi s own arms still. Ho w does he know that he is going to
be angry later on?
F: He doesn't know . But , just in case .. .
D: No, Dadd y, it doesn' t make sense . I don't smile so as to be able to
tell you I am angry by not smiling l ater on .
20

F: Yes—I think that that is part of the reason for smiling. And there
are lots of peop le who smile in order to tell you that they are not
angry—when they really are.
D: But that's differen t, Dadd y. That's a sort of telling lies with one's
face. Like playing poke r.
F: Yes.
F: Now where are we? You don't think it sensible for Frenchmen to
work so hard to tell each other that they are not angry or hurt.
But after all what is most conversa tion abou t? I mean, among
Americans?
D: But, Dadd y, it's abou t all sorts of things—bas eball and ice cream
and gardens and games. And peop le talk about other peop le and
abou t themselves and about what t hey got for Chr istmas.
F: Yes, yes—but who listens? I mean—all right, so they talk abou t
basebal l and gard ens. But are they exchanging information? And,
if so, what information?
D: Sure—when you come in from fishing, and I ask you "did you
catch anything?" and you say "nothing ," I didn't know that you
wouldn' t catch anything t ill you told me.
F: Hmm.
* * *
F: All right-s o you mention my fishing—a matter about which I am
sensitive—and then there is a gap, a silence in the conve rsation
—and that silence tells you that I don't like cracks abou t how
many fish I didn' t catch. It's just like the French man who stops
waving his arms about when he is hurt.
D: I'm sorry, Dadd y, but you did say .. .
F: No—wai t a minute— let's not get confused by being sorry—I
shall go out fishing again tomorrow and I shall still know that I
am unlikely to catch a fish .. .
D: But, Dadd y, you said all convers ation is only telling
other people that you are not angry with them .. .
F: Did I? No—not all conver sation, but much of it. Some-times if
both people are willing to listen carefu lly, it is possible to do
more than exchange greetings and good wish es. Even to do more
21

than exchange information . The two peop le may even find out
something which neither of them knew be fore.
* * *
F: Anyhow, most conver sations are only about wheth er peop le are
angry or something . They are busy telling each other that they
are friendly—which is sometimes a lie. Afte r all, what happ ens
when they canno t think of anything to say? They all feel
unco mfortable.
D: But wouldn't that be information , Dadd y? I mean—info rmation
that they are no t cross?
F: Surely, yes. But it's a different sort of information from "the cat is
on the m at."
* * *
D: Dadd y, why cannot peop le just say "I am not cross at you" and
let it go at that?
F: Ah, now we are getting to the real probl em. The point is that the
messages which we exch ange in gesture s are really not the same
as any translation of those ges tures into words .
D: I don' t unde rstand .
F: I mean—that no amount of telling somebod y in mere words that
one is or is not angry is the same as what one might tell them by
gestu re or tone o f voice.
D: But, Dadd y, you cannot have word s withou t some tone of voice ,
can y ou? Even i f somebod y uses as l ittle tone as h e can, t he oth er
people wi ll hear that he i s hold ing hi mself b ack—and th at wil l be
a sort of tone, won't it?
F: Yes—I suppo se so. After all that's what I said just now about
gestu res— that the French man can say something special by
stopping his gestures.
* * *
F: But then, what do I mean by saying that “mere words” can never
conve y the same message as gestu res—i f there are no "mere
words"?
D: Well, the words might be wr itten.
22

F:No—tha t won't let me out of the difficulty. Because written
words still have some sort of rhythm and they still have
overtones . The poin t is that no mere words e xist. There are only
words with either gestu re or tone of voice or something of the
sort. But , of course, gestures withou t words are common enough.
* * *
D: Dadd y, when they teach us French at schoo l, why don't they
teach us to wave our h ands?
F: I don't know . I'm sure I don't know . That is probabl y one of the
reasons why people find learning languages so difficult.
* * *
F: Anyhow, it is all nonsense . I mean, the notion that langu age is
made of words is all nonsense—and when I said that gestu res
could not be translated into "mere words," I was talking
nons ense, becau se there is no such thing as "mere words ." And
all the syntax and grammar and all that stuff is nonsense . It's all
based on the idea that "mere" words exist—and there are none.
D: But, Dadd y .. .
F: I tell you—we have to start all over again from the beginning and
assume that language is first and fore-most a system of gestures.
Animals after all have only gestu res and tones of voice—and
words were inven ted later. Much later. And after that they
invented schoo l-masters.
D: Dadd y?
F: Yes.
D: Would it be a good thing if peop le gave up words and went back
to only using gestures ?
F: Hmm. I don't know . Of course we would not be able to have any
conv ersations like this. We could only bark, or mew, and wave
our arms about , and laugh and grunt and weep. But it might be
fun— it would make life a sort of ballet—with dance rs making
their own m usic.
23

Metalogue: Abou t Games and Being
Serious*
Daughter : Dadd y, are these convers ation s serious ?
Fathe r: Certa inly they are.
D: They're no t a sort of game that you play with me?
F: God forbid . . . but they are a sort of game that we play together.
D: Then they're not serious !
* * *
F: Suppose you tell me what you would unde rstand by the words
"serious" and a "game."
D: Well . . . if you're . . . I don' t know .
F:If I am what?
D: I mean . . . the conversa tions are serious for me, but if you are
only playing a game .. .
F: Steady now. Let's look at what is good and what is bad abou t
"playing" and "games." First of all, I don't mind —not much—
about winn ing or losing. When your questions put me in a tight
spot, sure, I try a little harder to think straight and to say clearly
what I mean. But I don't bluff and I don't set traps. There is no
temptation to cheat.
D:That's just it. It's not serious to you. It's a game. Peop le who
cheat just don't know how to play. They treat a ga me as though it
were s erious.
F:But i t is serious.
D: No, it isn't—not for you it isn't.
F: Bec ause I don't even wan t to che at?
D: Yes—pa rtly that.
* This metalo gue is reprinted by permission from ETC.: A Revie w of Gener al
Semantics, Vol. X, 1953 .
24

F: But do you want to cheat and bluff all the time? D: No—of
course not.
F: Well then?
D:Oh—Dadd y—you'll never unders tand.
F:I gues s I neve r will.
F: Look , I scored a sort of debating point just now by forcing you to
admit that you don't want to cheat—and then I tied onto that
admission the conclu sion that therefo re the conversa tions are not
"seriou s" fo r you either. Was that a sort of ch eating?
D: Yes—sor t of.
F:I agree—I think it was . I'm sorry.
D:You see, Dadd y—if I cheated or wanted to cheat , that would
mean that I was not serious abou t the things we talk about. It
would mean th at I was only playing a game with you.
F:Yes, that makes sen se.
* * *
D: But i t doesn' t make sense, Dadd y. It's an awful muddle.
F: Yes—a muddle—but still a sort of sense .
D: How, Dadd y?
* * *
F:Wait a minute. This is difficult to say. First of all—I think that we
get somewhere with the se conv ersations . I enjoy them very much
and I think you do. But also, apart from that, I think that we get
some ideas straigh t and I think that the muddles help. I mean—
that if we both spoke logically all the time, we would neve r get
anywhere. We would only parrot all the old cliches that
everybody has repeated for hundred s of years.
D: What is a cliche, Dadd y?
F: A cliche? It's a French word , and I think it was origina lly a
printer's word. When they print a sentence they have to take the
separa te letters and put them one by one into a sort of grooved
stick to spell out the sentence . But for words and sentences
which peopl e use often, the printer keep s little sticks of letters
ready made up. And these ready-made sentenc es are called
cliches.
25

D: But I've forgotten now what you were saying about cliches,
Daddy.
F:Yes—it was about the muddles that we get into in these talks
and how getting into muddles makes a sort of sense. If we
didn't get into muddles, our talks would be like playing rummy
without first shuffling the cards.
D:Yes, Daddy—but what about those things—the ready-made
sticks of letters?
F: The cliches? Yes—it's the same thing. We all have lots of ready-
made phrases and ideas, and the printer has ready-made sticks
of letters, all sorted out into phrases. But if the printer wants to
print something new—say, something in a new language, he
will have to break up all that old sorting of the letters. In the
same way, in order to think new thoughts or to say new things,
we have to break up all our ready-made ideas and shuffle the
pieces.
D: But, Daddy, the printer would not shuffle all the letters? Would
he? He wouldn't shake them all up in a bag. He would put them
one by one in their places—all the a's in one box and all the b's
in another, and all the commas in another, and so on.
F: Yes—that's right. Otherwise he would go mad trying to find an a
when he wanted it.
* * *
F:What are you thinking?
D: No—i t's on ly that there are so many ques tions . F: For example?
D: Well, I see what you mean about our getting into muddles. That
that makes us say new sorts of things . But I am thinking abou t
the printer. He has to keep all his little letters sorted out even
though he breaks up all the ready-made phrases. And I am
wondering' about our muddles . Do we have to keep the little
pieces of our thought in some sort of order— to keep from going
mad?
F:I think so— yes—but I don't know what sort of order. That
would be a terribly hard question to answer . I don't think we
could get an answer t o that question today.
26

* * *
F:You said there were "so many quest ions. " Do y ou have ano ther?
D:Yes—abou t games and being serious. That's what we started
from, and I don't know how or why that led us to talk about our
muddles. The way you confuse everything—it 's a sort of
cheating.
F:No, absolutely not.
* * *
F:You brought up two quest ions. And r eally there are a lot more . . .
We started from the quest ion a bout t hese conve rsations—ar e they
serious ? Or are they a sort of game? And you felt hurt that I
might be playing a game, while you were serious. It looks as
though a conver sation is a game if a person takes part in it with
one set of emotion s or ideas—but not a "game" if his ideas or
emotions a re different.
D: Yes, it's if your ideas about the conver sation are different from
mine .. .
F: If we both had the game idea, it would be all right? D: Yes—of
course.
F: Then it seems to be up to me to make clear what I mean by the
game idea. I know that I am serious—wha tever that means—
abou t the things that we talk abou t. We talk about ideas. And I
know that I play with the ideas in order to unde rstand them and
fit them togeth er. It's "play" in the same sense that a small child
"plays" with blocks . . . And a child with building blocks is
mostly very serious about his "play."
D: But i s it a game, Dadd y? Do you play against me?
F:No. I think of it as you and I playing together agains t the
building blocks—the ideas. Sometimes competing a bit—but
competing as to who can get the next idea into place. And
sometimes we attack each other's bit of building , or I will try to
defend my built-up ideas from your criticism. But always in the
end we are working together to build the ideas up so that they
will stand .
27

* * *
D:Dadd y, do our talks have rules? The difference between
a game and j ust playing is that a game has rules.
F:Yes. Let me think abou t that. I think we do have a sort of rules …
and I think a child playing with blocks has rules. The blocks
themselves make a sort of rules. They will balance in certain
positions and they will not balance in other positions. And it
would be a sort of cheating if the child used glue to make the
blocks stand up in a position from which they would otherwise
fall.
D: But what rules do we have?
F: Well, the ideas that we play with bring in a sort of rules. There
are rules about how ideas will stand up and sup-port each
other. And if they are wrongly put together the whole building
falls down.
D: No glue, Daddy?
F: No—no glue. Only logic.
* * *
D: But you said that if we always talked logically and did not get
into muddles, we could never say anything new. We could only
say ready-made thing s. What did you call those things?
F:Clich es. Yes. Glue i s wha t cliche s are stuck together with.
D: But you said "logic ," Dadd y.
F: Yes, I know . We're in a muddle again . Only I don't see a way out
of this particular muddle.
* * *
D: How did we ge t into it, Dadd y?
F: All right, let's see if we can retrace our steps. We were talking
about the "rules" of these conver sations. And I said that the ideas
that we p lay with hav e rules of logic .. .
D: Dadd y! Wouldn' t it be a good thing if we had a few more rules
and obeyed them more carefully? Then we might not get into
these dreadful muddles .
28

F: Yes. But wait. You mean that I get us into these muddles becaus e
I cheat a gains t rules which we don't h ave. Or pu t it this wa y. That
we might have rules which would stop us from getting into
muddles— as long a s we obe yed them.
D: Yes, Dadd y, that's what the rules of a game are for.
F:Yes, but do you want to turn these conve rsations into that sort of
a game? I'd rather play canast a—which i s fun too .
D: Yes, that's right. We can play canasta whenever we want to. But
at the moment I would rather play this game. Only I don't know
what sort of a game this is. Nor what so rt of rules it has.
F: And y et we hav e been playing for some time.
D:Yes. And i t's been fun.
F:Yes.
* * *
F:Let's go back to the question which you asked and which I said
was too difficult to answer today. We were talking about the
printer breaking up his cliches , and you said that he would still
keep some sort of order among his letters— to keep from going
mad. And then you asked "What sort of order should we cling to
so that when we get into a muddle we do not go mad?" It seems
to me that the "rules" of the game is only an-other name for that
sort of order.
D:Yes—and cheating is what g ets us into muddles.
F: In a sense, yes. That's right. Excep t that the whol e point of the
game is that we do get into muddles, and do come out on the
other side, and if there were no muddles our "game" would be
like canas ta or chess—and that is not how we wan t it to be.
D: Is it you that make the rules, Dadd y? Is that fair?
F: That, daughte r, is a dirty crack. And probably an unfair one. But
let me accept it at face value. Yes, it is I who make the rules—
after all, I do no t want us to go m ad.
D:All right. But , Dadd y, do you also ch ange t he rules? Sometimes?
F:Hmm, another dirty crack. Yes, daugh ter, I change them
constantly. Not a ll of them, but some of them.
D: I wish y ou'd tell me when y ou're going to change them!
29

F: Hmm—yes—again. I wish I could. But it isn't like that. If it were
like chess or canas ta, I could tell you the rules, and we could , if
we wanted to, stop playing and discuss the rules. And then we
could start a new game with the new rules. But what rules would
hold us between the two games? Whil e we were discus sing the
rules?
D:I don't unders tand.
F:Yes. The point is that the purpo se of these conversations is to
discov er the "rules." It's like life—a game whose purpose is to
discover the rules, which rules are always changing and always
undiscoverable.
D:But I don' t call that a game, Dadd y.
F:Perhaps not. I would call it a game, or at any rate "play." But it
certainly is not like chess or canasta. It's more like what kittens
and puppies do. Perhaps. I don't know.
* * *
D:Dadd y, why do kittens and puppies play?
F:I don't know—I don't know .
30

Metalogue: How Much Do You Know?*
Daught er: Dadd y, how much do you know?
Father: Me? Hmm—I have a bout a pound of knowl edge.
D:Don't be silly. Is it a pound sterling or a pound weight ? I mean
really how m uch do you know?
F: Well, my brain weigh s abou t two pounds and I suppose I use
abou t a quarter of it—o r use it at about a quarter efficiency. So
let's say half a pound.
D: But do you know more than Johnn y's dadd y? Do you know more
than I do?
F: Hmm—I once knew a little boy in Engl and who asked his father,
"Do fathers always know more than sons? " and the father said,
"Yes." The next question was, "Dadd y, who invent ed the steam
engine?" and the father said, "James Watt." And then the son
came back with " —but wh y didn't James Watt's father invent it?"
* * *
D: I know . I know more than that boy because I know why James
Watt's father didn' t. It was because some-body else had to think
of something else before anybody could make a steam engine. I
mean something like—I don't know—but there was somebod y
else who had to discove r oil before anybody could make an
engine.
F: Yes—that makes a differenc e. I mean, it means that knowledge is
all sort of knitted together, or woven, like cloth, and each piece
of knowledge is only meaningfu l or usefu l because of the other
pieces—and . . .
D: Do you think we ought to measure it by the yard?
F: No. I don 't.
D: But t hat's how we bu y cloth.
* This metalogue is reprinted by permission from ETC.:A Revie w of Gener al
Semantics, Vol. X, 1953. 91
31

F: Yes. But I didn' t mean that it is cloth. Only it's like it—and
certainly would not be flat like cloth—bu t in three dimensions—
perhap s fou r dimensions.
D: What do you mean, Dadd y?
F:I really don't know , my dear. I was ju st trying to think.
F:I don't think we are doing very well this morning. Sup-pose we
start out on another tack . What we have to think abou t is how the
pieces of knowledge are woven togethe r. How they help each
other.
D: How do they?
F: Well—i t's as if sometimes two facts get added together and all
you have is just two facts. But sometimes instead of just adding
they multiply—and y ou get four facts.
D:You canno t multipl y one by one and get four. You know you
can't.
F:Oh.
* * *
F: But yes I can, too. If the things to be multipl ied are pieces of
knowledge or facts or something like that. Because every one of
them is a doubl e something.
D:I don't unders tand.
F:Well—at least a doub le something .
D: Dadd y!
F:Yes—take the game of Twenty Quest ions. You think of
something . Say you think of "tomorrow ." All right. Now I ask "Is
it abstract?" and y ou say "Yes." Now from your "yes" I hav e got a
double bit of information . I know that it is abstract and I know
that it isn't concrete. Or say it this way—from your "yes" I can
halve the number of possibilities of what the thing can be. And
that's a multipl ying by one ove r two .
D:Isn't it a division?
F:Yes—it 's the same thing. I mean—al l right—i t's a multiplication
by .5. The important thing is that it's not just a subtra ction or an
addition.
D:How do you know it isn't?
32

F:How do I know it?—W ell, suppose I ask another question which
will halve the possibilities among the abstractions. And then
another. That will have brought down the total possibilities to an
eighth of what they were at the beginning . And two times two
times two is eight.
D:And two and two and two i s only six.
F:That's right.
D:But, Dadd y, I don't see—wha t happen s with Twenty Quest ions?
F: The point is that if I pick my question s properly I can decide
between two times two times two times two twent y times over
things—220 things. That's over a mil-lion things that you might
have thought of. One question is enough to decide between two
things ; and two questions will decide between four thing s—and
so on.
D: I don't like arithmetic , Dadd y.
F: Yes, I know . The working it out is dull, but some of the ideas in it
are amusing. Anyhow, you want ed to know how to measure
knowledg e, and if you start measuring things that alwa ys leads to
arithmetic.
D: We haven't measured any knowledg e yet.
F:No. I know . But we have made a step or two toward knowing
how we would measure it if we wanted to. And that means we
are a little nearer to knowing what knowledge i s.
D: That would be a funny sort of knowledg e, Dadd y. I mean
knowing about knowledge—would we measure that sort of
knowing the same way?
F: Wait a minute—I don't know—tha t's really the $64 Quest ion on
this subject. Because—we ll, let's go back to the game of Twenty
Ques tions . The point that we never mention ed is that those
question s have to be in a certain order. First the wide gene ral
question and then the detailed question. And it's only from
answers to the wide questions that I know which detailed
question s to ask. But we coun ted them all alike. I don 't know . But
now y ou ask m e if knowing abou t knowl edge would be m easured
the same way as other knowledge . And the answer must surely be
no. You see, if the early quest ions in the game tell me what
question s to ask later, then they must be partly questions about
knowing. They're exp loring the bus iness of knowing .
33

D: Dadd y—has anybody ever measured how m uch any-body knew .
F: Oh yes. Often . But I don't quite know what the answers meant.
They do it with examinations and tests and quizzes, but it's like
trying to find out how big a piece of paper is by throwing stones
at it.
D: How do you mean?
F:I mean—if you throw stone s at two pieces of pape r from the
same distan ce and you find that you hit one piece more often
than the other, then probabl y the one that you hit most will be
bigge r than the other. In the same way, in an examination you
throw a lot of question s at the students, and if you find that you
hit more piece s of knowledge in one student than in the other s,
then you think t hat student must know m ore. That's the idea.
D: But could one measure a piece of paper that way?
F:Surel y one could. It might even be quite a good way of doing it.
We do measure a lot of thing s that way. For example, we judge
how strong a cup of coffee is by looking to see how black it is—
that is, we look to see how much light is stopped. We throw light
waves a t it instead of stones , it's the same idea.
D: Oh.
* * *
D: But then—wh y shou ldn't we m easure knowledg e that way?
F: How? By quizzes ? No—God forbid. The trouble is that that sort
of measuring leaves out your point—tha t there are different sorts
of knowl edge—and that there's knowing abou t knowledg e. And
ought one t o give h igher marks to the studen t who can answer the
widest question? Or perhap s there should be a different sort of
marks for each d ifferent sort of ques tion.
D:Well, all right. Let's do that and then add the marks togethe r and
then .. .
F:No— we couldn't add them together. We might multiply or divide
one s ort of marks b y anothe r sort but we couldn' t add th em.
D:Why not, Dadd y?
F:Because—be cause we couldn 't. No wonder you don't like
arithmetic i f they don' t tell you that sort of thing at school—What
34

do they tell you? Golly—I wonder what the teach ers think
arithmetic is about .
D: What is it about, Dadd y?
F: No. Let's stick to the question of how to measure knowl edge—
Arithmetic is a set of tricks for thinking clearly and the only fun
in it is just its clarity. And the first thing abou t being clear is not
to mix up ideas which are really different from each other. The
idea of two oranges is really different from the idea of two miles.
Becau se if you add t hem together you only get fog in your head.
D: But, Dadd y, I can't keep ideas sep arate. Ought I to do that?
F: No— No— Of course not. Combine them. But don't add them.
That's all. I mean—if the ideas are numbers and you want to
combine two different sorts, the thing to do is to multiply them
by each other. Or divide them by each other. And then you'll get
some new sort of idea, a new sort of quant ity. If you have miles
in your head , and you have hours in your head , and you divide
the m iles b y the hours , you get "miles per hour"— that's a speed .
D:Yes, Dadd y. What would I get if I multiplied them?
F:Oh—er— I suppos e you'd get mile-hours . Yes. I know what they
are. I mean, what a mile-hou r is. It's what you pay a taxi drive r.
His meter measure s miles and he has a clock which measures
hours, and the meter and the clock work together and multiply
the hours by the miles and then it multiplies the mile-hou rs by
something e lse which m akes mile-hours into do llars.
D: I did a n expe riment once .
F:Yes?
D:I wanted to find out if I could think two though ts at the same
time. So I thought "It's summer" and I thought "It's winte r." And
then I tried to think the two though ts togethe r.
F:Yes?
D:But I found I wasn' t having two though ts. I was only having one
thought about having two t houghts .
F:Sure, that's just it. You can't mix though ts, you can only combine
them. And in the end, that means you can't count them. Because
counting is really only adding things together. And you mostly
can't do that.
D:Then really do we only have one big thought which has lots
of branches and lots and lots of branches?
35

F:Yes. I think so. I don't know. Anyhow I think that is a clearer
way of saying it. I mean it's clearer
than talking about bits of knowledge and trying to count them.
* * *
D: Dadd y, why don't you use the other three-qua rters of your brain ?
F: Oh, yes—that— you see the troub le is that I had school -teachers
too. And they filled up about a quarter of my brain with fog. And
then I read newspaper s and listened to what other peop le said,
and t hat filled up anothe r qua rter with fog.
D: And t he other quarter, Dadd y?
F: Oh—th at's fog that I made for myself when I was trying to think.
36

Metalogu e: Why Do Things Have
Outlines?*
Daught er: Dadd y, why do things h ave ou tlines?
Father: Do t hey? I don' t know . What so rt of things do you mean?
D: I mean when I draw thing s, why do they have outlines ? F: Well,
what about other sorts of things—a flock of sheep? or a
conv ersation? Do they have outlines?
D: Don't b e silly. I can't draw a conve rsation. I mean things.
F: Yes—I was trying to find out just what you meant. Do you mean
"Why do we give thing s outlines when we draw them?" or do
you mean tha t the t hings h ave out -lines whether we dr aw them or
not?
D: I don't know , Dadd y. You tell me. Which do I mean?
F: I don't know , my dear. There was a very angry artist once who
scribb led all sorts of things down, and after he was dead they
looked in his books and in one place they found he'd written
"Wise men see outlines and therefo re they draw them" but in
another place he'd written "Mad men see outlines and therefor e
they draw them."
D: But which does he m ean? I don't under stand .
F: Well, William Blake— that was h is na me—was a grea t artist and a
very angry man. And sometimes he rolled up his ideas into little
spitballs so that he could throw them at people.
D:But what was he mad about, Daddy?
F:But what was he mad about? Oh, I see—you mean "angry." We
have to keep those two meanings of "mad" clear if we are going
to talk about Blake. Because a lot of people thought he was mad
—really mad—crazy. And that was one of the things he was
mad-angry about. And then he was mad-angry, too, about some
artists who painted pictures as though things didn't have out-
lines. He called them "the slobbering school."
D: He wasn't very tolerant, was he, Daddy?
* Reprinted by permission from ETC.: A Revie w of General Semantics, Vol. X I,
1953 .
37

F:Tolerant? Oh, God. Yes, I know—that's what they drum into
you at school. No, Blake was not very tolerant. He didn't even
think tolerance was a good thing. It was just more slobbering.
He thought it blurred all the outlines and muddled everything—
that it made all cats gray. So that nobody would be able to see
anything clearly and sharply.
D: Yes, Daddy.
F: No, that's not the answer. I mean "Yes, Daddy" is not the answer.
All that says is that you don't know what your opinion is—and
you don't give a damn what I say or what Blake says and that
the school has so befuddled you with talk about tolerance that
you can-not tell the difference between anything and anything
else.
D: (Weeps.)
F: Oh, God. I'm sorry, but I was angry. But not really angry with
you. Just angry at the general mushiness of how people act and
think—and how they preach muddle and call it tolerance.
D: But, Daddy
F: Yes?
D: I don't know. I don't seem able to think very well. It's all in a
muddle.
F:I'm sorry. I suppose I muddled you by starting to let off steam.
* * *
D: Dadd y? F: Yes?
D: Why is that something to get angry about?
F:Is what something to get angry about?
D: I mean—about whether things have outlines. You said William
Blake got angry about it. And then you get angry about it. Why
is that, Daddy?
F: Yes, in a way I think it is. I think it matters. Perhaps in a way, is
the thing that matters. And other things only matter because
they are part of this.
D: What do you mean, Daddy?
F: I mean, well, let's talk about tolerance. When Gentiles want to
bully Jews because they killed Christ, I get intolerant. I think
38

the Gentiles are being muddle-headed and are blurring all the
outlines. Because the Jews didn't kill Christ, the Italians did it.
D: Did they, Daddy?
F: Yes, only the ones who did are called Romans today, and we
have another word for their descendants. We call them Italians.
You see there are two muddles and I was making the second
muddle on purpose so we could catch it. First there's the
muddle of getting the history wrong and saying the Jews did it,
and then there's the muddle of saying that the descendants
should be responsible for what their ancestors didn't do. It's all
slovenly.
D: Yes, Daddy.
F: All right, I'll try not to get angry again. All I'm trying to say is
that muddle is something to get angry about. D: Daddy?
F: Yes?
D: We were talking about muddle the other day. Are we really
talking about the same thing now?
F: Yes. Of course we are. That's why it's important—what we said
the other day.
D: And you said that getting things clear was what Science was
about.
F: Yes, that's the same thing again.
* * *
D: I don't seem to unders tand it all very well. Ever ything seems to
be everything else, and I get lost in it.
F: Yes, I know it's difficult. The point is that our conve rsations do
have an outline, somehow—if only one cou ld see it clearly.
* * *
F: Let's think about a real conc rete out-and-out muddle, for a
change, and see if that will help. Do you remember the game of
croquet in Alice in Wonder land?
D: Yes—with flamingos?
F: That's right.
D: And porcup ines for balls?
39

F:No, hedg ehogs. They were hedgehogs . They don't have
porcup ines i n Eng land.
D: Oh. Was it in England , Dadd y? I didn't know .
F: Of cours e it was in England . You don't have duch esses in
America either.
D: But there's the Duchess of Windso r, Dadd y.
F: Yes, bu t she doe sn't have qu ills, not like a real porcupine.
D: Go on about Alice and don't be silly, Dadd y.
F: Yes, we were talking about flamingos. The point is that the man
who wrote Alice was think ing about the same things that we are.
And he amused himself with little Alice by imagining a game of
croque t that would be all muddle, just absolute muddle. So he
said they should use flamingos as mallets becaus e the flamingos
would bend their necks so the player wouldn't know even
whether his mallet would hi t the ball or how it would hit the ball.
D: Anyhow the ball might walk away of its own accord because it
was a hedgehog .
F: That's right. So that it's all so muddled that nobod y can tell at all
what's going to happen.
D: And t he hoop s walked around , too, becau se they were soldiers.
F: That's right—eve rything could m ove and nobod y could tell how it
would m ove.
D: Did everything have to be alive so as to make a complete
muddle?
F: No—he could have made it a muddle by . . . no, I suppose you're
right. That's interesting. Yes, it had to be that way. Wait a minute.
It's curious but you're right. Because if he'd muddled thing s any
other way, the players could have learned how to deal with the
muddling details. I mean, suppose the croquet lawn was bumpy,
or the balls were a funny shape, or the heads of the mallets just
wobbly instead of being alive, then the people could still learn
and the game would only be more difficult—it wouldn't be
impossible. But once you bring live things into it, it becomes
impossible. I wouldn't have expected that.
D: Wouldn't you, Daddy? I would have. That seems natural to me.
F: Natural? Sure—natural enough. But I would not have expected
it to work that way.
D: Why not? That's what I would have expected.
40

F: Yes. But this is the thing that I would not have expected. That
animals, which are themselves able to see things ahead and act
on what they think is going to happen—a cat can catch a mouse
by jumping to land where the mouse will probably be when she
has completed her jump—but it's just the fact that animals are
capable of seeing ahead and learning that makes them the only
really unpredictable things in the world. To think that we try to
make laws as though people were quite regular and predictable.
D:Or do they make the laws just because people are not
predictable, and the people who make the laws wish the other
people were predictable?
F: Yes, I suppose so.
* * *
D: What were we talking about ?
F:I don't quite know—not yet. But you started a new line by asking
if the game of croque t could be made into a real muddle only by
having all the things in it alive. And I went chasing after that
question , and I don't think I've caught up with it yet. There is
some-thing funny abou t that point.
D: What?
F: I don't quite know—not yet. Something abou t living things and
the differen ce between them and the things that are not alive—
machines, stones, so on. Hors es don't fit in a world of
automobiles . And that's part of the same point. They're
unpredic table , like flamingos i n the game of croque t.
D:What a bout people , Dadd y?
F:What a bout them?
D:Well, they're alive. Do th ey fit? I mean on the streets?
F: No, I suppose they don't really fit—o r only by work ing pretty
hard to protect themselves and make themselves fit. Yes, they
have to make themselves predictable, be-cause otherwise the
machines ge t angry and kill them.
D: Don't b e silly. If the machines can get angry, then they would ,not
be predic table. They'd be like you, Dadd y. You can't predict
when y ou're a ngry, can you?
F:No, I suppose not.
41

D: But , Dadd y, I'd rather have you unpr edictable— sometimes.
* * *
D: What did you mean by a conve rsation having an out-l ine? Has
this conv ersation had an outline?
F: Oh, surely, yes. But we canno t see it yet because the conve rsation
isn't finished . You cannot ev er see i t whi le you're in the m iddle of
it. Becaus e if you could see it, you would be predictable—l ike
the machine. And I would be predic table—and the two of us
together would be predictable
D: But I don't unders tand. You say it is important to be clear about
things. And you get angry abou t peop le who blur the outlines .
And yet we think it's better to be unpredi ctable and not to be like
a machine. And you say that we cannot see the outlines of our
conver sation till it's over. Then it doesn't matter wheth er we're
clear or not. Because we canno t do anything abou t it then.
F: Yes, I know—and I don't understand it myself. . . . But anyway,
who wants to do anything about it?
42

Metalogue: Why a Swan?*
Daught er: Why a swan?
Father: Yes—and why a puppe t in Petroushka ?
D: No—tha t's different. After all a puppet is sort of human—and
that particular puppet is very human. F: More human than the
peop le?
D: Yes.
F:But still only sort of human? And after all the swan is also sort
of human.
D: Yes.
* * *
D: But what abou t the danc er? Is she human? Of course she really
is, but, on the stage, she seems inhuman or impersonal—pe rhaps
superhuman. I don 't know .
F: You mean—th at while the swan is only a sort of swan and has
no webbing between her toes, the dancer seems only sort of
human.
D: I don't know—perhaps it's something l ike that.
* * *
F: No—I get confu sed when I speak of the "swan" and the dance r as
two different things . I would rather say that the thing I see on the
stage— the swan figure—is both "sort of" human and "sort of"
swan .
D: But t hen y ou would be using the word " sort of" in two s enses .
F: Yes, that's so. But anyhow, when I say that the swan figure is
"sort of" human, I don't mean that it (or she) is a member of that
species or sort which we ca ll human. D: No, of cou rse not.
F: Rathe r that she (or it) is a member of another subdiv ision of a
larger group which would include Petroushka pupp ets and ballet
swans and p eople .
* This metalogue appeared in Impulse 1954 and is re-printed by permission of
Impulse Publications, Inc.
43

D: No, it's not like genera and specie s. Does your larger group
include g eese?
F: All right. Then I eviden tly do not know what the word "sort of"
means. But I do know that the whole of fantasy, poetry, ballet,
and art in general owes its meaning and importance to the
relationship which I refer to when I say that the swan figure is a
"sort of" swan—or a "pretend" swan .
D: Then we shall never know why the dance r is a swan or a puppet
or whateve r, and shall neve r be able to say what art or poetry is
until someone says what i s really meant b y "sort of."
F:Yes.
F:But we don't have to avoid puns. In French the phrase espece de
(literally "sort of") carries a special sort of punch . If one man
calls anothe r "a ca mel" the in sult m ay be a friend ly one. But if he
calls him an espece de chameau—a sort of camel—th at's bad. It's
still worse to call a man an espece d'espece—a sort of a sort. D:
A sort of a sort of what?
F: No—just a sort of a sort. On the other hand, if you say of a man
that he is a true camel, the insult carries a flavor of grudg ing
admiration.
D: But when a French man calls a man a sort of camel, is he using
the phras e sort of in anything like the same way as I, when I say
the swan i s sort of hu man?
* * *
F:It's like—the re's a passage in Macbe th. Macbeth is talking to the
murdere rs whom he is sending out to kill Banquo . They claim to
be m en, and he tells them they are sort of men.
Ay—in the catalogue y e go fo r men.
as hounds and greyhounds, m ongrels, spaniels, curs,
shoughs , wate r-rugs and demi-wolve s are clept
all by the name of dogs.
(Macbeth, Act III, Scene 1)
D: No—that' s what you said just now. What was it? "Anoth er
subdiv ision of a larger group ?" I don' t think t hat's it at all.
44

F: No, it's not only that. Macb eth, after all, uses dogs in his simile.
And "dog s" means either noble hounds or scavenger s. It would
not be the same if he had used the domestic varieties of cats—or
the sub speci es of wild roses.
D: All right, all right. But what is the answer to my quest ion? When
a French man calls a man a "sort of" camel, and I say that the
swan is "sort of" human, do we both mean the same thing by
"sort of"?
* * *
F: All right, let's try to analyze what "sort of" means. Let's take a
single sentenc e and examine it. If I say "the puppe t Petroushka is
sort of human," I state a relation-ship .
D:Between what and what ?
F:Between ideas, I think .
D:Not between a pupp et and peop le?
F:No. Between some ideas that I have about a puppe t and some
ideas that I have about peopl e.
D:Oh.
* * *
D: Well then, what s ort of a relationsh ip?
F:I don' t know . A metaphoric relationship ?
* * *
F: And then there is that other relationsh ip which is emphatic ally
not "sort of." Many men have gone to the stake for the
proposi tion that the bread and wine are not "sort of" the body
and blood .
D:But is that the same thing? I mean—is the swan ballet a
sacrament?
F:Yes—I th ink so— at leas t for some peopl e. In Protes tant langu age
we might say that the swanl ike costume and movements of the
dancer are "outward and visible signs of some inward and
spiritual grace" of woman. But in Catholic language that would
make the ballet into a mere metaphor and not a sacrament.
45

D: But you said that for some people it is a sacrament. You mean
for Protestants?
F: No, no. I mean that if for some people the bread and wine are
only a metaphor, while for others—Catholics —the bread and
wine are a sacrament; then, if there be some for whom the
ballet is a metaphor, there may be others for whom it is
emphatically more than a metaphor—but rather a sacrament.
D: In the Catholic sense?
F:Yes.
* * *
F: I mean that if we could say clearly what is meant by the
propos ition "the bread and wine is not `sort of' the body and
blood" ; then we shou ld know more abou t what we mean when
we say either that the swan is "sort of" human or that the ballet is
a sacrament.
D: Well—how do you tell the difference?
F:Which difference?
D: Between a sacrament and a metaphor .
* * *
F: Wait a minute. We are, after all, talking abou t the per-former or
the artist or the poet, or a given member of the audi ence. You ask
me how I tell the differen ce be tween a sacrament a nd a metaphor .
But my answe r must deal with the person and not the message .
You ask me how I would decide whether a certain dance on a
certain day is or is not sacramental for the particular dance r.
D: All right—bu t get on with it.
F: Well— I think it's a sort of a secret.
D: You m ean you won't tell me?
F: No—it 's not that sort of secret. It's not something that one must
not tell. It's something that one cannot tell.
D: What do you mean? Why not?
F: Let us suppos e I asked the dance r, "Miss X, tell me, that dance
which you perfo rm—is it for you a sacrament or a mere
metaphor ?" And let us imagine that I can make this quest ion
intelligible. She will perhap s put me off by saying, "You saw it
46

—it is for you to decide, if you want to, whether or not it is
sacramental for you." Or she might say, "Sometimes it is and
sometimes it isn't." Or "How was I, last night?" But in any case
she can have no direct control over the matter.
* * *
D: Do y ou mean that a nybody who knew thi s secr et would have i t in
their power to be a great dancer or a great poet?
F: No, no, no. It isn't like that at all. I mean first that great art and
religion and all the rest of it is about this secret; but knowing the
secret in an ordin ary conscious way would not give the knower
control.
* * *
D: Dadd y, wha t has happened ? We were trying to find ou t what "sort
of" means when we say that the swan is "sort of" human. I said
that there must be two senses of "sort of." One in the phrase "the
swan figure is a `sort of' swan, and another in the phrase "the
swan figure is `sort of' human." And now you are talking abou t
mysterious secrets and c ontro l.
F: All right. I'll start again. The swan figure is not a real swan but a
pretend swan. It is also a pretend-no t human being. It is also
"really" a young lady wear ing a white dress. And a real swan
would resemble a young lady in certain wa ys.
D: But which of these is sacramental?
F: Oh Lord , here we go again. I can only say this: that it is not one
of these statements but their combination which constitutes a
sacrament. The "pretend" and the "pretend-no t" and the "really"
someho w ge t fused togethe r into a single meaning .
D: But we ought t o keep t hem separate.
F: Yes. That is what the logicians and the scientists try to do. But
they do not create ballets that way—nor sacraments.
47

Metaloque: What Is an Ins tinct?*
Daughter : Dadd y, what is an instinct?
Fathe r: An instinct, my dear, is a explanato ry principle. D: But what
does it explain?
F:Anything—a lmost anything at all. Anything you want it to
expla in.
D: Don't be silly. It doesn't expla in gravity.
F:No. But that is because nobod y wants "instinct" to expla in
gravity. f they did, it would explain it. We could simply say that
the moon has an instinct whose streng th varies inversely as the
squar e of the distance .. .
D: But that's nonsense , Dadd y.
F:Yes, surely. But i t was y ou who m entioned "instinct," not I.
D:All right—but then what do es exp lain gravity?
F:Nothing , my dear, because g ravity is an explanator y princip le.
D:Oh.
D: Do you mean that you canno t use one explana tory princ iple to
expla in another ? Never ?
F:Hmm . . . hardly ever. That is what Newton meant when he said,
"hypotheses non fingo."
D: And wha t does that mean? Please.
F:Well, you know what "hypothes es" are. Any statement linking
together two descrip tive statements is an hypothes is. If you say
that there was a full moon on February 1st and anoth er on March
1st; and then you link these two observa tions together in any
way, the statement whi ch links them is an hypothesis.
D:Yes—and I know what non means. But what's fingo?
F:Well—fingo is a late Latin word for "make." It forms a verbal
noun fictio from which we get the word "fiction."
D:Daddy, do you mean that Sir Isaac Newton thought that all
hypotheses were just made up like stories?
F: Yes—precisely that.
* This metalo gue is reprinted by permission of Mouton & Co. from Approaches
to Animal Communication, edited b y Thomas A. Sebeok, 1969
48

D: But didn't he discover gravity? With the apple? F: No, dear. He
invented it.
D: Oh…. Daddy, who invented instinct?
F:I don't know. Probably biblical.
D:But if the idea of gravity links together two descriptive
statements, it must be an hypothesis.
F: That's right.
D: Then Newton did fingo an hypothesis after all.
F: Yes—indeed he did. He was a very great scientist. D : Oh.
D:Daddy, is an explanatory principle the same thing as an
hypothesis?
F: Nearly, but not quite. You see, an hypothesis tries to explain
some particular something but an explanatory principle—like
"gravity" or "instinct"—really explains nothing. It's a sort of
conventional agreement between scientists to stop trying to
explain things at a certain point.
D: Then is that what Newton meant? If "gravity" explains nothing
but is only a sort of full stop at the end of a line of explanation,
then inventing gravity was not the same as inventing an
hypothesis, and he could say he did not fingo any hypotheses.
F: That's right. There's no explanation of an explanatory principle.
It's like a black box.
D: Oh.
D: Dadd y, wha t's a black box ?
F: A "black box" is a conven tiona l agreement between scien tists to
stop trying to explain things at a certain point. I guess it's usually
a temporary agreement.
D: But t hat doesn' t sound like a black box .
F:No—but that's what i t's called. Things often don 't sound l ike their
names.
D: No.
F: It's a word that comes from the engineers. When they draw a
diagra m of a complicated machine , they use a sort of shorthand.
Instead of drawing all the details, they put a box to stand for a
49

whole bunch of parts and label the box with what that bunch of
parts is supposed t o do.
D:So a "black box" is a label for what a bunch of things are
supposed to do… .
F:That's right. But it's not an explanat ion of how the bunch work s.
D:And gravity?
F:Is a label for what gravity is supposed to do. It's not an
explan ation of how it does it.
D: Oh.
D: Dadd y, what is an instinct?
F:It's a label for what a certain black box is suppos ed to do.
D: But what's it suppos ed to do?
F:Hm. That is a very difficult ques tion .. .
D:Go on.
F:Well. It's suppo sed to control—part ly control—what an organism
does.
D: Do p lants have instincts?
F: No. If a botan ist used the word "instinct," when talking abou t
plants, he would be accus ed of zoomorphis m. D: Is that bad?
F:Yes. Very bad for botanists. For a botanist to be guilty of
zoomorphis m is as bad as for a zoologis t to be guilty of
anthropomorphis m. Very bad, indeed .
D: Oh. I see.
D: What d id you mean by "partly control"?
F: Well. If an animal falls down a cliff, its falling is con-trolled by
gravity. But if it wigg les while falling, that might be due to
instinct.
D: Se lf-preserv ative instinct?
F:I suppose so.
D: What i s a self, Dadd y? Does a dog know i t has a self?
F:I don't know . But if the dog does know it has a self, and it
wiggles in order to preserve that self, then its wiggling is
rational, not instinctive.
D:Oh. Then a "self-pre serva tive ins tinct" is a con tradiction. F: Well,
it's a sort of halfway house on the road to anthropo morphis m.
D:Oh. That's bad.
50

F: But the dog might know it had a self and not know that that self
shou ld be preserved . It would then be rational to not wiggle . So
if the dog s till wigg les, this would be ins tinctive. But if it learned
to wigg le, then it would not be instinctive.
D: Oh.
D:What would not be instinctive, Dadd y? The learning or the
wigg ling?
F:No—jus t the wigg ling.
D:And t he learning would be instinctive?
F:Well . . . yes. Unless the dog h ad to learn to learn. D : Oh.
D:But, Dadd y, what is instinct suppo sed to exp lain?
F: I keep trying to avoid that quest ion. You see, instincts were
invented before an ybody knew anything abou t gen etics, and m ost
of modern genetics was discovered before anybody knew
anything abou t communication theory. So it is doubl y difficult to
translate "instinct" into modern t erms and i deas.
D:Yes, go on.
F:Well, you know that in the chromosomes, there are genes; and
that the genes are some sort of messages which have to do with
how t he organis m develops and with how it behave s.
D:Is develop ing differen t from behav ing, Dadd y? What's the
difference? And which is learning ? Is it "develop ing" or
"behaving?"
F:No! No! Not so fast. Let's avoid those questions by putting
developing -learning -behavio r all together in one basket. A single
spectru m of pheno mena. Now let's try to say how instinct
contributes to explaining this spectrum.
D: But is it a spectrum?
F:No—that's only a loose way of talking.
D: Oh.
D: But isn't instinct all on the behavior end of that "spectrum"?
And isn't learning all determined by environment and not
chromosomes?
F: Let's get this clear—that there is no behavior and no anatomy
and no learning in the chromosomes them-selves.
51

D: Don't they have their own anatomy?
F: Yes, of course. And their own physiology. But the anatomy and
physiology of the genes and chromosomes is not the anatomy
and physiology of the whole animal.
D: Of course not.
F:But it is about the anatomy and physiology of the whole
animal.
D: Anatomy about anatomy?
F: Yes, just as letters and words have their own forms and shapes
and those shapes are parts of words or sentences and so on—
which may be about anything.
D: Oh.
D: Daddy, is the anatomy of the genes and chromosomes about the
anatomy of the whole animal? And the physiology of the genes
and chromosomes about the physiology of the whole animal?
F: No, no. There is no reason to expect that. It's not like that.
Anatomy and physiology are not separate in that way.
D: Daddy, are you going to put anatomy and physiology together
in one basket, like you did developing-learning-behavior?
F:Yes. Certainly.
D : Oh.
D: The same basket?
F: Why not? I think developing is right in the middle of that
basket. Right smack in the middle.
D: Oh.
D:If chromosomes and genes have anatomy and physiology, they
must hav e deve lopment.
F: Yes. That follows .
D: Do you think their development could be about the deve lopment
of the whole o rganis m?
F:I don't even know what t hat quest ion would mean.
D:I do. It means that the chromoso mes and gene s would be
changing or deve loping somehow while the baby is deve loping ,
and the c hanges i n the ch romosomes would be about the changes
in the baby. Contro lling th em or partly controlling them.
F:No. I don' t think so.
52

D:Oh.
D: Do chromosomes learn?
F: I don't know .
D: They do sound r ather like black boxes .
F: Yes, but if chromosomes or genes can learn, then they are much
more complicated black boxes than anybody at present believes .
Scientists are always assuming or hoping that thing s are simple,
and then discover ing that they are not.
D: Yes, Dadd y.
D: Dadd y, is that an instinct?
F:Is what an instinct?
D:Assu ming that things are simple.
F:No. Of cou rse no t. Scien tists have t o be taught t o do that.
D:But I thought no organism could be taught to be wrong every
time.
F: Young lady, you are being disrespect ful and wrong . In the first
place, scientists are not wrong eve ry time they assu me that things
are simple. Quite often they are right or partly right and still
more often, they think they are right and tell each other so. And
that is enough reinforce ment. And, anyhow you are wrong in
saying that no organis m can be taugh t to be wrong e very time.
D: When people say that something is "instinctive," are they trying
to make things simple? F : Yes, indeed.
D: And are they wrong?
F:I don' t know . It depend s on what t hey mean.
D:Oh.
D: When do they do it?
F: Yes, that's a better way of asking the question. They do it when
they see a creature do something, and they are sure: first, that the
creatu re did not learn how to do that something and, second, that
the creature is too stupid to under stand why it shou ld do that.
D: Any other time?
F: Yes. When they see that all members of the species do the same
things unde r the same circumstances ; and when they see the
53

animal repeating the same action even when the circumstances
are changed so that the action fails.
D: So there are four ways of knowing that it's instinctive. F: No.
Four c ondit ions und er whi ch sc ientists talk about instinct.
D:But what if one condi tion isn't there? An instinct sounds rather
like a habit or a custom.
F:But hab its are learned .
D: Yes.
D: Are habits alwa ys twice learned ?
F: What do you mean?
D: I mean—when I learn a set of chords on the guitar, first I learn
them or find them; and then later when I practice, I get the habit
of playing them that way. And sometimes I get bad h abits.
F:Learn ing to be wrong every time?
D: Oh—all right. But what abou t that twice-over business? Would
both parts of learning be not there if guitar playing were
instinctive?
F:Yes. If both parts of learning were clearly not there, scientists
might say that guitar playing is instinctive.
D: But what i f only one part of learning was missing ?
F:Then, logically, the missing part could b e exp lained by "instinct."
D:Could either part be m issing?
F:I don't know . I don' t think anybody knows.
D: Oh.
D:Do birds practice their songs?
F:Yes. Some birds are said to practice.
D:I guess instinct gives them the first part of singing , but they have
to work on the second part.
F: Pe rhaps .
D:Could practicing be instinctive?
F:I suppose it could be—bu t I am not sure what the word "instinct"
is coming to mean in this conve rsation.
D:It's an explanato ry princip le, Dadd y, just like you said… There's
one t hing I don't unders tand.
F: Yes?
D: Is there a whol e lot of instinct? Or are there lots of instincts?
54

F: Yes. That's a good qu estion, and s cientists hav e talked a great deal
abou t it, making lists of separ ate instincts and then lumping them
togethe r again.
D:But what's the answer ?
F:Well. It's not quite clear. But one thing is certain: That
explanatory principles must be not multiplied beyond necess ity.
D: And that means? Pl ease?
F: It's the idea behind monothei sm—that the idea of one big God is
to be preferred to the idea of two little gods .
D: Is God an explana tory principle?
F: Oh, yes—a very big one. You shouldn' t use two black boxes—or
two instincts—to explain wha t one black box would exp lain .. .
D: If it were big enough.
F:No. It means .. .
D:Are there big instincts and little instincts?
F:Well—a s a matter of fact, scientists do talk as if there were . But
they call the little instincts by other names —"ref lexes ," "innate
releasing mechanis ms," "fixed a ction pa tterns," and s o on.
D: I see—l ike having one big God to explain the universe and lots
of little "imps" or "gob lins" to explain the small things that
happ en.
F: Well, yes. Rath er like that.
D: But, Dadd y, how do they lump things together to make the big
instincts ?
F: Well, for example, they don't say that the dog has one instinct
which makes it wigg le when it falls down the cliff and anothe r
which makes it run away from fire.
D:You mean those would both be explained by a self-preserv ative
instinct?
F:Something l ike that. Yes.
D: But if you put those different acts together under one instinct,
then you cannot get away from saying that the dog has the use of
the no tion of "self."
F:No, perhap s not.
D: What would you do abou t the instinct for the song and the
instinct for practicing the song?
F: Well—d epending on what the song is used for. Both song and
practice might b e unde r a territorial instinct or a sexua l instinct.
55

D: I wouldn't put them togethe r.
F:No?
D:Because what if the bird also practiced picking up seed or
something ? You'd have to multipl y the instincts —wha t is it?—
beyond ne cessi ty.
F:What do you mean?
D: I mean a food -getting instinct to explain the practicing p icking up
seed, and a territory instinct for practicing song. Why not have a
practicing instinct for both? That saves one black box .
F: But then you would throw away the idea of lumping together
under the same instinct actions which hav e the same purpose .
D: Yes—becaus e if the practicing is for a purpose—I mean, if the
bird has a purpos e—then the practicing is rational and not
instinctive. Didn' t you say something like that?
F:Yes, I did say something l ike that.
D: Could we do without the idea of "instinct"?
F:How would you explain things t hen?
D:Well. I'd just look at the little things : When some-thing goes
"pop," the dog jumps. When the ground is not unde r his feet, he
wiggles . And so on .
F:You mean—al l the imps but no gods ?
D: Yes, something l ike that.
F: Well. There are scientists who try to talk that way, and it's
beco ming qu ite fashionab le. They say it is more objec tive.
D: And i s it?
F: Oh , yes.
D: What do es "ob jective" mean?
F:Well. It means that you look very hard at those things which you
choose to look at.
D: That sounds ri ght. But how do the objective people choose which
things t hey will be obj ective abou t?
F: Well. They choose those things about which it is easy to be
objec tive.
D: You m ean easy for them?
F: Yes.
56

D: But how do they know that those are the easy things?
F:I suppos e they try different thing s and find out by expe rience.
D: So it's a subjec tive choic e?
F:Oh, yes. All expe rienc e is sub jective.
D: But it's human and subjective. They decide which bits of animal
behavior to be objective about by consu lting human subjective
experien ce. Didn' t you say that anthropomorphis m is a bad
thing?
F:Yes—bu t they do try to be not human.
D: Which things do they leave ou t?
F:What do you mean?
D:I mean—subjec tive experien ce shows them which things it is
easy to be objec tive about . So, they go and study those things .
But which things does their exper ience show are difficult? So
that they avoid those things. Which a re the things they avoid ?
F: Well, you mentioned earlier something called "practice." That's a
difficult thing to be objective abou t. And there are other things
that are difficult in the same sort of way. Play, for example. And
exploration. It's difficult to be objec tive abou t whethe r a rat is
really explor ing or really playing. So they don't inves tigate those
things . And t hen t here's love. And, of course , hate.
D: I see. Those are the sorts of things that I wanted to invent
separa te instincts for. F: Yes—those thing s. And don't forget
humor.
D: Dadd y—are a nimals objec tive?
F: I don't know—probabl y not. I don't think they are subjec tive
either. I don't think th ey are split that way.
D:Isn't it true that peop le have a special difficulty about being
objective a bout t he more an imal pa rts of their nature ?
F:I guess so. Anyhow Freud said so, and I think he was right. Why
do you ask?
D: Becau se, oh dear, those poor peopl e. They try to study animals.
And they specialize in those things that they can study
objectively. And they can only be objec tive abou t those thing s in
57

which they themselves are least like animals. It must be difficult
for them.
F:No—that does not necess arily follow. It is still possib le for
people to be objective about some things in their animal nature.
You haven't shown that the whole of animal behavio r is within
the set of things t hat peopl e canno t be objective abou t.
D: No?
D: What a re the really big d ifference s between peop le and animals?
F:Well—in tellect, language , tools. Things like that.
D:And it is easy for peop le to be intellectually objective in
languag e and about tools?
F: That's right.
D: But that must mean that in people there is a whole set of ideas or
whatnot which are all tied togethe r. A sort of second creature
within the whole person, and that second creatu re must have a
quite differen t way of think ing about everything. An objective
way.
F: Yes. The royal road to consciousne ss and objectivity is through
languag e and tools.
D: But what happens when this creatu re looks at all those parts of
the person abou t which it is difficult for peop le to be objective?
Does i t just look? Or do es it meddle?
F: It meddles .
D: And wha t happen s?
F: That's a very terrible question.
D: Go on. If we are going to study animals, we must face that
question.
F: Well . . . The poets and artists know the answer better than the
scientists. Let me read you a piece:
Thoug ht chang 'd the infinite to a serpent, that which pitieth
To a devourin g flame; and man fled from its face and hid
In forests of night: then all the et ernal forests were' divided
Into earths rollin g in circles of space , that like an ocean rush'd
And over whelmed all except this finite w all of flesh.
Then w as the serpent temple form'd, ima ge of infinite
Shut up in finite revolutions; and man bec ame an
58

Angel , Heaven a mighty circle turning, God a tyrant c rown'd.*
D: I don't understand it. It sounds terrible, but what does it mean?
F: Well. It's not an objective statement, because it is talking about
the effect of objectivity—what the poet calls here "thought"
upon the whole person or the whole of life. "Thought" should
remain a part of the whole but instead spreads itself and
meddles with the rest.
D: Go on.
F:Well. It slices everything to bits.
D: I don't understand.
F: Well, the first slice is between the objective thing and the rest.
And then inside the creature that's made in the model of
intellect, language, and tools, it is natural that purpose will
evolve. Tools are for purposes and anything which blocks
purpose is a hindrance. The world of the objective creature gets
split into "helpful" things and "hindering" things.
D: Yes. I see that.
F: All right. Then the creature applies that split to the world of the
whole person, and "helpful" and "hinde ring" beco me Good and
Evil, and the world is then split between God and the Serpen t.
And after that, more and more splits follow because the intellect
is always classifying and d ividing th ings up.
D: Multiplying explanatory principles beyond necessity? F: That's
right.
D: So, inevitably, when the objective creature looks at animals, it
splits things up and makes the animals look like human beings
after their intellects have invaded their souls.
F:Exactly. It's a sort of inhuman anthropomorphism.
D:And that is why the objective people study all the little imps
instead of the larger things?
F: Yes. It's called S-R psychology. It's easy to be objective about
sex but not about love.
* Blake, W., 1794, Europ e a Prophec y, printed and published by the author .
(Italics added .)
59

D: Daddy, we've talked about two ways of studying animals—the
big instinct way and the S-R way, and neither way seemed very
sound. What do we do now?
F:I don't know.
D:Didn't you say that the royal road to objectivity and
consciousness is language and tools? What's the royal road to
the other half?
F:Freud said dreams.
D: Oh.
D: What are dreams? How are they put together?
F: Well—dreams are bits and pieces of the stuff of which we are
made. The non-objective stuff.
D: But how are they put together?
F:Look. Aren't we getting rather far from the question of
explaining animal behavior?
D: I don't know, but I don't think so. It looks as if we are going to
be anthropomorphic in one way or another, whatever we do.
And it is obviously wrong to build our anthropomorphism on
that side of man's nature in which he is most unlike the animals.
So let's try the other side. You say dreams are the royal road to
the other side. So . . .
F:I didn't. Freud said it. Or something like it.
D: All right. But how a re dreams put together ?
F:Do you mean how are two d reams related to each other?
D: No. Because, as you said, they are only bits and pieces. What I
mean is: How is a dream put together inside itself? Could
animal behavior be put together in the same sort of way?
F:I don't know where to begin.
D: Well. Do dreams go by opposites?
F: Oh Lord! The old folk idea. No. They don't predict the future.
Dreams are sort of suspended in time. They don't have any
tenses.
D:But if a person is afraid of something which he knows will
happen tomorrow, he might dream about it to-night?
60

F:Certainly. Or about something in his past. Or about both past
and present. But the dream contains no label to tell him what it
is "about" in this sense. It just is.
D: Do you mean it's as if the dream had no title page?
F:Yes. It's like an old manuscript or a letter that has lost its
beginning and end, and the historian has to guess what it's all
about and who wrote it and when—from what's inside it.
D:Then we're going to have to be objective, too?
F:Yes indeed. But we know that we have to be careful about it.
We have to watch that we don't force the concepts of the
creature that deals in language and tools upon the dream
material.
D: How do you mean?
F: Well. For example: if dreams somehow have not tenses and are
somehow suspended in time, then it would be forcing the wrong
sort of objectivity to say that a dream "predicts" something.
And equally wrong to say it is a statement about the past. It's
not history.
D: Only propaganda?
F:What do you mean?
D: I mean—is it like the sort of stories that propagandists write
which they say are history but which are really only fables?
F:All right. Yes. Dreams are in many ways like myths and fables.
But not consc iousl y made up by a propagandis t. Not p lanned .
D: Does a dream always have a moral?
F: I don't know about always. But often, yes. But the moral is not
stated in the dream. The psychoanalyst tries to get the patient to
find the moral. Really the whole dream is the moral.
D: What does that mean?
F:I don't quite know.
D:Well. Do dreams go by opposites? Is the moral the opposite of
what the dream seems to say?
F:Oh yes. Often. Dreams often have an ironic or sarcastic twist. A
sort of reductio ad absurdum.
D: For example?
F:All right. A friend of mine was a fighter pilot in World War II.
After the war he became a psychologist and had to sit for his
61

Ph. D. oral exam. He began to be terrified of the oral, but, the
night before the exam, he had a nightmare in which he
experienced again being in a plane which had been shot down.
Next day he went into the examination without fear.
D: Why?
F: Because it was silly for a fighter pilot to be afraid of a bunch of
university professors who couldn't really shoot him down.
D: But how did he know that? The dream could have been telling
him that the professors would shoot him down. How did he
know it was ironic?
F: Hmm. The answer is he didn't know. The dream doesn't have a
label on it to say it is ironic. And when people are being ironic
in waking conversation, they often don't tell you they are being
ironic.
D: No. That's true. I always think it's sort of cruel. F: Yes. It often
is.
D: Daddy, are animals ever ironic or sarcastic?
F: No. I guess not. But I am not sure that those are quite the words
we should use. "Ironic" and "sarcastic" are words for the
analysis of message material in language. And animals don't
have language. It's perhaps part of the wrong sort of objectivity.
D: All right. Then do animals deal in opposi tes?
F: Well, yes. As a matter of fact, they do. But I'm not sure it's the
same thing .. .
D: Go on. How do they? And when?
F:Well. You know how a puppy lies on his back and presents his
belly to a bigger dog. That's sort of inviting the bigger dog to
attack. But it works in the opposite way. It stops the bigger dog
from attacking.
D: Yes. I see. It is a sort of use of opposites. But do they know
that?
F:You mean does the big dog know that the little dog is saying the
opposite of what he means? And does the little dog know that
that is the way to stop the big dog?
D: Yes.
F:I don't know. I sometimes think the little dog knows a little
more about it than the big dog. Anyhow, the little dog does not
62

give any signals to show that he knows. He obviously couldn't
do that.
D: Then it's like the dreams. There's no label to say that the dream
is dealing in opposites.
F:That's right.
D: I think we're getting somewhere. Dreams deal in opposites, and
animals deal in opposites, and neither carries labels to say when
they are dealing in opposites.
F:Hmm.
D: Why do animals fight?
F: Oh, for many reasons. Territory, sex, food .. .
D: Daddy, you're talking like instinct theory. I thought we agreed
not to do that.
F:All right. But what sort of an answer do you want to the
question, why animals fight?
D: Well. Do they deal in opposites?
F:Oh. Yes. A lot of fighting ends up in some sort of peace-
making. And certainly playful fighting is partly a way of affir-
ming friendship. Or discovering or rediscovering friendship.
D: I thought so. . . .
D: But why are the labels missing? Is it for the same reason in both
animals and dreams?
F: I don't know. But, you know, dreams do not always deal in
opposites.
D: Does a dream always have a moral?
F: I don't know abou t always. But often, yes. But the moral is not
stated in the dream. The psychoana lyst tries to get the patient to
find the m oral. Really the whole dream is the moral.
D: What does that mean?
F:I don't quite know .
D:Well. Do dreams go by oppo sites? Is the moral the oppos ite of
what the dream seems to say?
F:Oh yes. Often. Drea ms often have an ironic or sarcastic twist. A
sort of reductio ad absurdum.
D: For e xample?
63

F: All right. A friend of mine was a fighter pilot in World War II.
After the war he became a psycholog ist and had to sit for his Ph.
D. oral exam. He began to be terrified of the oral, but, the night
befor e the exam, he had a nightmare in which he exper ienced
again being in a plane which had been shot down. Next day he
went i nto the examination without fear.
D: Why?
F:Because it was silly for a fighter pilot to be afraid of a bunch of
unive rsity professor s who cou ldn't really shoot him down.
D: But how did he know that? The dream could have been telling
him that the professo rs would shoot him down. How did he
know it was ironic?
F: Hmm. The answer is he didn't know . The dream doesn' t have a
label on it to say it is ironic. And when peop le are being ironic in
waking conve rsation, they often don't tell you they are being
ironic.
D: No. That's true. I alwa ys think it's sort of cruel.
F:Yes. It often is.
D: Dadd y, are animals ever ironic or sarcas tic?
F:No. I guess not. But I am not sure that those are quite the words
we should use. "Ironic" and "sarcastic" are word s for the anal ysis
of message materia l in language . And animals don't have
languag e. It's perhaps part of the wrong so rt of ob jectivity.
D: All right. Then do animals deal in opposi tes?
F: Well, yes. As a matter of fact, they do. But I'm not sure it's the
same thing .. .
D: Go on . How do they? And when?
F:Well. You know how a pupp y lies on his back and presents his
belly to a bigge r dog. That's sort of inviting the bigge r dog to
attack. But it works in the opposi te way. It stops the bigger dog
from attacking .
D: Yes. I see. It is a sort of use of oppos ites. But do they know that?
F:You mean does the big dog know that the little dog is saying the
opposi te of what he means? And does the little dog know that
that is the wa y to stop the big dog ?
D: Yes.
64

F:I don't know . I sometimes think the little dog knows a little more
abou t it than the big dog. Anyhow, the little dog does not give
any signa ls to show that he knows. He obvious ly couldn't do that.
D: Then it's like the dreams. There's no labe l to say that the dream is
dealing in opposi tes.
F:That's right.
D: I think we're getting somewhere. Drea ms deal in oppo sites, and
animals deal in oppos ites, and neither carries labels to say when
they are de aling in opposi tes.
F: Hmm.
D: Why do animals f ight?
F:Oh, for many reasons. Territory, sex, food . . .
D: Dadd y, you're t alking like instinct theory. I thought we agreed not
to do that.
F:All right. But what sort of an answer do you want to the question,
why animals fight?
D: Well. Do they deal in opposi tes?
F:Oh. Yes. A lot of fighting ends up in some sort of peace-making.
And certainly playful fighting is partly a way of affirming
friend ship. Or discove ring or rediscover ing friendsh ip.
D: I thought so. . . .
D: But why are the labels missing? Is it for the same reason in both
animals and d reams?
F: I don't know . But, you know , dreams do not always deal in
oppos ites.
D: No—of cou rse no t—nor do an imals.
F:All right then.
D: Let's go back to that dream. Its total effect on the man was the
same as if somebod y had said to him, " `you in a fighter plane' is
not equa l to `you in an oral exam.' "
F:Yes. But the dream didn't spell that out. It only says, "you in a
fighte r plane. It leaves out the "not," and it leaves out the
instruction to compare the dream with something else and it
doesn't say what he should compare it with.
D: All right. Let's take the "not" first. Is there any "not" in animal
behavior ?
65

F:How could t here be?
D: I mean can an animal sa y by its actions, "I will not bite you"?
F: Well, to begin with. Communication by actions canno t possibly
have tense s. They are only possible in languag e.
D: Didn' t you say that dreams have no tenses?
F:Hmm. Yes, I did.
D:Okay. But what abou t "not" . Can the an imal say, "I am not biting
you"?
F: That still has a tense in it. But never mind. If the animal is not
biting t he other, he's not biting it, and th at's it.
D:But he might be not doing all sorts of other thing s, sleeping ,
eating, running, and so on. How can he say, "It's biting that I'm
not doing"?
F: He can only do that if biting ha s somehow been mentioned .
D: Do you mean that he could say, "I am not biting you" by first
showing h is fangs and then not biting?
F:Yes. Something l ike that.
D:But what abou t two animals? They'd both have to show their
fangs .
F:Yes.
D: And, it seems to me, they might misunderst and each other, and
get into a fight.
F:Yes. There is alwa ys that dange r when you deal in oppos ites and
do not or cannot say what you are doing , especially when you do
not know what you are doing . D: But the animals would know
that they bared their fangs i n ord er to say, "I won't b ite you."
F:I doubt whether they would know . Certainly neither animal
knows it abou t the other. The dreamer doesn' t know at the
beginn ing of the dream how the dream is going to end.
D:Then it's a sort of exp eriment… .
F: Yes.
D: So they might get into a fight in order to find out wheth er
fighting was what they had t o do.
F: Yes—but I'd rather put it less purpos ively—that the fight shows
them what sort of relationsh ip they have , after it. It's not planned .
D: Then "not" is really not there when t he animals show t heir fangs?
F: I guess not. Or often not. Perhaps old friends might engag e in
playful fighting and know a t the beg inning what t hey are do ing.
66

D: All right. Then the "not" is absen t in animal behav ior becaus e
"not" is part of verba l language , and there can-not be any action
signal for, "not." And because there is no "not," the only way to
agree on a negative is to act out the whole reductio ad
absurdum. You have to a ct ou t the ba ttle to p rove it isn't one, a nd
then you have to act out the submission to prove that the other
won't eat you.
F: Yes.
D: Did t he an imals have t o think that out?
F:No. Because it's all necessarily true. And that which is
necessa rily true will govern what you do regard less of whether
you know that it is necessar ily true. If you put two apples with
three apples you will get five apples—ev en though you cannot
coun t. It's another way of "explaining" t hings .
D : Oh.
D: But, then, why does the dream leave ou t the "no t"?
F:I think r eally for a r ather similar r eason . Drea ms are m ostly made
of images and feelings , and if you are going to communica te in
images and feelings and such, you again are governed by the fact
that there is no image for "not."
D:But you could dream of a "Stop" sign with a line through it,
which would m ean "No Stopp ing."
F: Yes. But that's halfwa y toward language . And the deleting line
isn't the word "not." It's the word "don't ." "Don't" can be
conv eyed in action languag e—if the other person makes a move
to mention what you want to forbid. You can even dream in
words , and the word "not" might be among them. But I doub t if
you can dream a "not" which is about the dream. I mean a "not"
which means "This dream is not to be taken literally."
Sometimes, in very light sleep, one knows that one is dreaming.
D:But, Dadd y, you still haven' t answered the question abou t how
dreams are pu t togethe r.
F:I think really I have answered it. But let me try again . A dream is
a metaphor or a tangle of metaphor s. Do you know what a
metaphor i s?
67

D: Yes. If I sa y you are like a pig t hat is a simile. But if I say you are
a pig, that is a metaphor .
F:Approxi matel y, yes. When a metaphor is labe led as a metaphor it
beco mes a simile.
D: And i t's that labeling that a dream leaves out .
F:That's ri ght. A metaphor compares things withou t spelling out the
comparison . It takes what is true of one group of things and
applies it to another. When we say a nation "decays," we are
using a metaphor , suggest ing that some chang es in a nation are
like changes which bacteria produce in fruit. But we don't stop to
mention the fruit or the bacteria.
D: And a dream is like that?
F:No. It's the other way around . The dream would mention the fruit
and possibly the bacteria but would not mention the nation. The
dream elabor ates on the relationsh ip but does not identify the
things t hat are related.
D: Dadd y, could you make a dream for me?
F:You mean, on this rec ipe? No. Let's take the pie ce of ver se which
I read you just now and turn it into a dream. It's almost dream
material the way it stands . For most of it, you have only to
subst itute images for the word s. And the words are vivid enough .
But the whole string of metaphors or images is pegged down,
which would not be so in a dream.
D:What do you mean by "pegged down"?
F: I mean by the first word: "Thought ." That word the writer is using
literally, and that one word tells you what all the rest is abou t.
D: And i n a dream?
F: That word , too, would have been metaphori c. Then the whole
poem would h ave be en m uch more difficult.
D:All right—change it then.
F:What abou t "Barbara changed the infinite . . ." and so on.
D: But why? Who is she?
F:Well, she's barbarous , and she's female, and she is the mnemonic
name of a s yllogis tic m ood. I though t she would do ra ther wel l as
a monstrous symbol for "Thought ." I can see her now with a pair
of calipers , pinching her own brain to change he r universe .
D: Stop it.
68

F: All right. But you see what I mean by saying that in dreams the
metaphors are not pegged down.
D: Do animals peg down their metaphors?
F: No. They don't have to. You see, when a grown-up bird makes
like a baby bird in approaching a member of the opposite sex,
he's using a metaphor taken from the relationship between child
and parent. But he doesn't have to peg down whose relationship
he is talking about. It's obviously the relationship between him-
self and the other bird. They're both of them present.
D:But don't they ever use metaphors—act out metaphors —about
something other than their own relationships?
F:I don't think so. No—not mammals. And I don't think birds do
either. Bees—perhaps. And, of course, people.
D: There's one thing I don't understand.
F:Yes?
D:We've found a whole lot of things in common between dreams
and animal behav ior. They both deal in opposites, and they both
have no tense s, and they both have no "not," and they both work
by metaphor , and neither of them pegs the metaphors down. But
what I don't unde rstand is—wh y, when the animals do these
things , it makes sense . I mean for t hem to work in oppos ites. And
they don't have to peg down their metaphors—but I don't see
why dreams shou ld be like that, too.
F:Nor do I.
D: And there's anoth er thing.
F:Yes?
D:You talked abou t gene s and chromosomes carrying messages
abou t development. Do they talk like animals and dreams? I
mean in metaphors and wi th no "nots "? Or do they talk like us?
F:I don't know . But I am sure their message system contains no
simple transform of Instinct Theory.
69

Part II: Form and
Pattern in
Anthropolog y

Culture Contact and Schismogen esis*
The Memorandu m written by a Committee of the Soc ial Sciences
Rese arch Council (Man, 1935, 162) has stimulated me to put
forward a point of view which differs considerab ly from theirs; and,
though the beginning of this article may appear to be critical of their
Memorandu m, I wish to make it clear from the outset that I regard
as a rea l contribution an y serious attempt to devis e categor ies for the
study of culture conta ct. Moreove r, since there are severa l passage s
in the Memorandu m (among them the Defin ition) which I do not
perfectly understand, m y criticisms are offered with so me hesitation ,
and are directed not so much against the Committee as against
certain errors prevalen t among anthropo logists.
(1) The uses of such systems of categories. In genera l it
is unwis e to construct systems of this sort until the problems which
they are designed to elucidate have been clearly formulated; and so
far as I can see, the catego ries drawn up by the Committee have
been const ructed not in reference to any specifically defined
proble ms, but to throw a gene ral light on "the proble m" of
acculturation, while the prob lem itself remains vague.
(2) From this it follows that our immediate need is not so much
the construction of a set of catego ries which will throw a light on all
the problems, but rather the schematic formulation of the problems
in such a way that they may be separately invest igable .
(3) Although the Committee leave their proble ms undefined, we
may from a careful reading of the catego ries gather roughl y what
question s they are asking of the materia l. It seems that the
Committee have , as a matter of fact, been influenced by the sort of
question s which administrators ask of anthropolog ists—"Is it a good
thing to use force in culture conta cts?" "How can we make a given
* The whole contro versy of which this article was a part has been reprinted in
Beyond the Frontier, edited by Paul Bohannon and Fred Plog. But the
ripples of this controvers y have long since died down, and the article is included
here only for its positive contributions. It is reprinted , unchanged, from Man,
Article 199, Vol. XXXV , 1935, by permission of the Royal Anthropological
Institute of Great Britain a nd Ireland.
71

people accept a certain sort of trait?" and so on. In response to this
type of question we find in the definition of accu lturation an empha-
sis upon difference in culture between the groups in contact and
upon the resulting changes; and such dicho tomies as that between
"elements forced upon a peopl e or received voluntarily by them"1
may likewis e be regarded as symptomatic of this thinking in terms
of administrative proble ms. The same may be said of the categor ies
V, A, B, and C , "acceptance," "adap tation" and "reaction."
(4) We may agree that answers are badly needed to these
quest ions of administ ration and, further, that a study of culture
conta cts is likely to give these answer s, But it is almost certain that
the scientific formulation of the probl ems of con-t act will not follow
these lines. It is as if in the const ruction of categor ies for the study
of criminolog y we started with a dichoto my of indiv idual s into
criminal and nonc riminal —and , indeed , that curious science was
hampered for a long while by this very attempt to define a "criminal
type."
(5) The Memorandu m is based upon a fallacy: that we can
classify the traits of a culture under such headings as econo mic,
religiou s, etc. We are asked, for example, to classify traits into three
classes, present ed respect ively becaus e of: (a) econo mic profit or
political dominance; (b) desirability of bringing about confo rmity to
value s of donor group; and (c) ethical and religious considera tions .
This idea, that each trait has either a single function or at least some
one function which overtops the rest, leads by extension to the
idea that a culture can be subdivided into "institutions" where the
bundle of traits which make up one institution are alike in their
major functions. The weakness of this method of subdividing a
culture has been conclusively demonstrated by Malinowski and
his pupils, who have shown that almost the whole of a culture
may be seen variously as a mechanism for modifying and
satisfying the sexual needs of the individuals, or for the
enforcement of the norms of behavior, or for supplying the
individuals with food.2 From this exhaustive demonstration we
must expect that any single trait of a culture will prove on
1 In any case it is clear that in a scientific stud y of processes and natural laws this
invocation of free will can have no place.
72

examination to be not simply economic or religious or structural,
but to partake of all these qualities according to the point of view
from which we look at it. If this be true of a culture seen in
synchronic section, then it must also apply to the diachronic
processe s of cul ture cont act and change; and we m ust expect tha t for
the offering, accep tance or refusal of every trait that are
simultaneou s causes of an econo mic, structu ral, sexua l, and
religious nature.
(6) From this it follows that our catego ries "religiou s,"
"econo mic," etc., are not real subd ivisions which are present in the
cultures which we study, but are merely abstractions which we
make for our own convenien ce when we set out to describe cultures
in words. They are not pheno mena present in culture, but are labels
for various points of view which we adop t in our studies. In
hand ling such abstractions we must be careful to avoid Whitehead 's
"fallacy of misplaced concre teness ," a fallacy into which, for
2 Cf. Malino wski, Sexual Life and Crime and Custom; A. I. Richards , Hunger
and Work. This question of the subdi vision of a culture into "institutions " is not
quite as simple as I have indicated; and, in spite of their own works, I believe that
the L ondon School still adher es to a theor y that some such division is practi cable. I t
is likely that confusion arises from the fact that certain native peoples—perhaps all,
but in any case those of Western Europ e—actuall y think that their culture is so
subdivided. Various cultural phenomena also contribute somethin g toward such a
subdivision, e.g., (a) the division of labor and differentiation of norms of behavior
between different groups of individuals in the same community, and (b) an
emphasis, present in certain cultures, upon the subdivisions of place and time upon
which beha vior is ordered . These pheno mena lead to the possibilit y, in such
cultures , of dubbin g all behavior which, f or example, takes place in church bet ween
11.30 and 12.30 on Sunda ys as "religious." But even in the study of such cultures
the anthropologist must look with some suspicion upon his classification of traits
into institutions and must expect to find a great deal of over-lapping between
various institutions.
An analogous fallacy occurs in psychology, and consists in regarding
behavior as classifiable according to the impulses which inspire it, e.g., into
such categories as self-protective, assertive, sexual, acquisitive, etc. Here, too,
confusion results from the fact that not only the psychologist, but also the
individual studied, is prone to think in terms of these categories. The
psychologists would do well to accept the probability that every bit of behavior
is—at least in a well-integrated individual —simultaneously relevant to all
these abstractions
73

example, the Marxian historians fall when they mainta in that
econo mic "phenomena" a re "p rimary."
With this preamble, we may now consider an alterna tive scheme
for the study of contac t phenomena.
(7) Scop e of the inqui ry I sugg est that we should consider under
the head of "culture contac t" not only those cases in which the
conta ct occu rs between two communities with different cultures and
results in profound disturban ce of the culture of one or both groups;
but also cases of contact within a single community. In these cases
the contact is between differen tiated groups of indiv idual s, e.g.,
between the sexes , between old and young, between aristocracy and
plebs , between clans , etc., groups which live together in
approx imate equilibrium. I would even extend the idea of "conta ct"
so widel y as to include those processes wher eby a child is molded
and trained to fit the culture into which he was born,3 but for the
present we may confine ourselve s to contacts between groups of
individua ls, with differen t cultural norms of behav ior in each g roup.
(8) If we consider the possible end of the drastic disturbance s
which follow contac ts between profound ly different communities,
we see that the changes must theore tically result in one or other of
the following p atterns:
(a)the complete fusion of the originally different groups
(b)the elimination of one or both groups
(c)the persistence of both groups in dynamic equilibrium
within one major community
(9) My purpose in extend ing the idea of contact to cove r the
condi tions of differentiation inside a single culture is to use our
knowledge of these quiescen t states to throw light upon the factors
which are at work in states of disequi librium. It may be easy to
obtain a knowledge of the factors from their quiet working , but
impossib le to isolate them when they are violent. The laws of
3 The present scheme is oriented toward the study of social rather than
psycholog ical processes, but a closel y analogous scheme might be constructed for
the study of psychopatholog y. Here the idea of "contact " would be studied ,
especiall y in the conte xts of the moldin g of the individual, and the processes of
schis mogenesis w ould be seen to play an important part not onl y in acc entuating the
maladjustm ents of the deviant, but also in assimilating the normal individual to his
group.
74

gravity canno t conven iently be studied by observation of houses
collapsing i n an earth-qu ake.
(10) Complete fusion Since this is one of the possible ends of the
process we must know what factors are presen t in a group of
individu als with consi stent homogeneous pat-terns of behav ior in all
members of the group. An appro ach to such cond itions may be
found in any community which is in a state of approx imate
equilibriu m but, unfor tunat ely, our own communities in Europe are
in a state of such flux that these cond itions scarcely occur.
Moreove r, even in primitive communities. the cond itions are usually
complicat ed by different iation, so that we must be conten t with
studie s of such homogeneous group s as can be observed within the
major di fferentiated communities.
Our first task will be to ascertain what sorts of unity obtain
within such groups, or rather—bea ring in mind that we are
concerned with aspects and not classes of phenomena—what
aspects o f the un ity of the bod y of tr aits we m ust descr ibe in o rder t o
get a whole view of the situation. I submit that the materia l, to be
fully unde rstood , must be examined in, at least, the following five
separab le aspect s:
(a) A structural aspect of unity The behavior of any one
individual in any one context is, in some sense, cognitively
consistent with the behavior of all the other individuals in all other
contexts. Here we must be prepared to find that the inherent logic
of one culture differs profoundly from that of others. From this
point of view we shall see, for example, that when individual A
gives a drink to individual B, that behavior is consistent with other
norms of behavior obtaining within the group which contains A
and B.
This aspect of the unity of the body of behavio r patterns may be
restated in terms of a standard ization of the cogn itive aspects of the
personal ities of the individuals . We may say that the patterns of
thought of the individua ls are so standardized that their behav ior
appears to them logica l.
(b) Affective aspec ts of unity In studying the culture from this
point of view, we are conc erned to show the emotional setting of all
the details of behav ior. We shall see the whole body of behav ior as a
75

conce rted mechanis m oriented toward affective satisfaction and
dissatisfaction of the individuals .
This aspec t of a culture may also be descr ibed in terms of a
stand ardization of affective aspec ts of the personal ities of the
individua ls, which are so modified by their culture that their
behavio r is to them emotion ally consi stent.
(c) Economic uni ty Here we shall see the whole bod y of behavio r
as a mechanis m oriented toward the produc tion and distribution of
material objec ts.
(d) Chronologica l and spatial unity Here we shall see the
behavio r patterns as schematica lly ordered according to time and
place. We shall see A as giving the drink to B "because it is Satu rday
evening in the Blue Boar ."
(e) Socio logica l unity Here we shall see the behavio r of the
individua ls as oriented towa rd the integr ation and disintegration of
the major unit, the Group as a whol e. We shall see the giving of
drink s as a factor whi ch pro motes t he so lidarity of the group .
(11) In addition to studying the behav ior of members of the
homogeneous group from all these point s of view, we must examine
a number of such groups to discove r the effects of standard ization of
these various points of view in the people we are studying. We have
stated above that every bit of behav ior must be rega rded as probabl y
relevant to all these viewpoin ts, but the fact remains that some
people s are more inclined than others to see and phrase their own
behavio r as "logica l" or "for the good of the State."
(12) With this knowledg e of the cond itions which obtain in
homogeneous groups , we shall be in a position to examine the
proce sses of fusion of two diverse groups into one. We may even be
able to prescrib e measures which will either promote or retard such
fusion, and predic t that a trait which fits the five aspects of uni ty can
be added to a culture with-out other change s. If it does not fit, then
we can search for appropr iate modifica tions either of the culture or
of the trait.
(13) The elimina tion of one or both groups This end result is
perhap s scarcely worth studying, but we should at least examine any
material that is available, to determine what sort of effects such
hostile activity has upon the culture of the survivo rs. It is possib le,
for example, that the patterns of behavio r associ ated with
76

elimination of other groups may be assimilated into their culture so
that they are impelled to eliminate m ore and more.
(14) Persistence of both groups in dynami c equilibrium This is
probabl y the most instru ctive of the possible end results of contac t,
since the factors active in the dynamic equilibriu m are likely to be
identical or analogou s with those which , in disequ ilibrium, are
active in cultural change . Our first task is to study the relationsh ips
obtaining between group s of ind ividua ls with differentiated behav ior
patterns, and later to conside r what light these relationsh ips throw
upon what are more usually called "contac ts." Ever y anthropolog ist
who has been in the field has had oppo rtuni ty of studying such
differentiated groups .
(15) The possibilities of differentiation of groups are by no
means infinite, but fall clearly into two categori es (a) cases in which
the relationsh ip is chiefly symmetrical, e.g., in the differentiation of
moieties , clans, villages and the nations of Europe ; and (b) cases in
which the relationship is complementar y, e.g., in the differentiation
of social strata, classes , castes, age grades, and, in some cases, the
cultural differentiation between the sexes .4 Both these types of
differentiation contain dynamic elements, such that when certain
restraining factors are removed the differentiation or split between
the groups increases progressively toward either breakdown or a
new equilibrium.
(16) Symmetr ical differentia tion To this category may be referred
all those cases in which the individua ls in two groups A and B have
the same aspirations and the same behavior patterns, but are
differentiated in the orienta tion of these patterns. Thus members of
group A exhibit behavio r patterns A,B,C in their dealing s with each
other, but a dopt the patterns X,Y,Z in their dealing s with members o f
group B. Similarly, group B adop t the patterns A,B,C among them-
4 Cf. Margaret Mead, Sex and Temperament, 1935. Of the communities
described in this book, the Arapesh and the Mundugumor have a preponderantly
symmetrical relationship between the sexes, while the Chambuli have a
comple mentar y relationship . Among the Iatmul, a tribe in the same area, which I
have studied , the relationship between the sexes is complem entar y, but on rather
different lines from that of the Chambuli. I hope shortl y to publish a book on the
Iatmul with sketches of their culture from the points of view a, b, and e out-lined in
paragraph 10. (See Bibliography, items 1936 and 1958 B.)
77

selves, but exhib it X,Y,Z in dealing with group A. Thus a position is
set up in which the behavio r X,Y,Z is the standard reply to X,Y,Z.
This position conta ins elements which may lead to progr essive
differen tiation or schismogenes is along the same lines. If, for
example, the patterns X,Y,Z include boasting, we shall see that there
is a Iikelihood , if boasting is the reply to boasting, that each group
will drive the other into exces sive emphasis of the pattern, a proce ss
which if not re-strained can only lead to more and more extreme
rivalry and ultimately to hostility and the breakdown of the whol e
system.
(17) Complementary differentiation To this categor y we may
refer all those cases in which the behavior and aspirations of the
members of the two groups are funda mentally differen t. Thus
members of group A treat each other with patterns L,M,N, and
exhib it the patterns O,P,Q in dealings with group B. In reply to
O,P,Q, the members of group B exhibit the patterns U,V,W, but
among themselve s they adop t patterns R,S,T. Thus it comes abou t
that O,P,Q is the reply to U,V,W, and vice versa. This differentiation
may be-co me progre ssive . If, for exa mple, the series, O,P ,Q inc ludes
patterns culturally regarded as assertive, while U,V,W includes
cultural sub missiveness , it is likely that sub missivene ss wil l pro mote
further assertiveness which in turn will promote further
submissiveness . This schismogenesi s, unless it is re-strained, leads
to a progress ive unilateral distortion of the personal ities of the
members of both groups , which results in mutual hostility between
them and must end i n the break-down o f the system.
(18) Reciprocity Though relation ships between groups can
broad ly be classified into two categori es, symmetrical and
complementary, this subd ivision is to some extent blurred by
anothe r type of different iation which we may describe as reciprocal.
In this type the behav ior patterns X and Y are adop ted by members
of each group in their dealings with the other group , but instead of
the symmetrica l system whereb y X is the reply to X and Y is the
reply to Y, we find here that X is the reply to Y. Thus in every single
in-stance the behav ior is asymmetrical, but symmetry is regain ed
over a large number of instances since sometimes group A exhib it X
to which group B reply with Y, and sometimes group A exhib it Y
and group B reply with X. Cases in which group A sometimes sell
78

sago to group B and the latter some-times sell the same commodity
to A, may be regard ed as recipro cal; but if group A habitually sell
sago to B while the latter habitually sell fish to A, we must, I think,
regard the pattern as complementar y. The recipro cal pattern, it may
be noted , is compensated and balanc ed within itself and therefore
does not tend towa rd sch ismogenesi s.
(19) Points for inves tigation:
(a) We need a prope r surve y of the types of behavio r which can
lead to schismogenese s of the symmetrica l type. At presen t it is only
possible to point to boasting and commercial rivalry, but no doubt
there are many other patterns which will be found to be
accompanied b y the same type of effect.
(b) We need a surve y of the types of behavio r which ar e mutually
comple mentar y and l ead to s chismogeneses o f the s econd t ype. Here
we can at presen t only cite assertiveness versus submissiveness ,
exhibitionism versus admiration , fostering versus express ions of
feeblen ess and, in addition, the variou s possib le combinations of
these pairs.
(c) We need verification of the gene ral law assumed abov e, that
when two groups exhibit complementary behav ior to each other, the
internal behavio r between members of group A must necessarily
differ from the interna l beh avior between members of group B.
(d) We need a systematic examination of schismogenes es of both
types from the various point s of view outlined in paragraph 10. At
present I have only looked at the matter from the ethologic al and
structural points of view (paragr aph 10, aspects a and b) . In
addition to this, the Ma rxian historians h ave g iven u s a picture of the
econo mic aspec t of comple mentary schismogenesis in Western
Europ e. It is likely, howeve r, that they themselves have been
influen ced unduly by the schismogenes is which they studied and
have been thereby prompted i nto exagger ation .
(e) We need to know something about the occur rence of
reciprocal behavior in relationship s which are preponder antly either
symmetrical or complementar y.
(20) Restrain ing factor s But, more important than any of the
proble ms in the previous paragraph , we need a study of the factors
which re strain both types of schismogenesi s. At the presen t moment,
the nations of Europe are far advanced in symmetrical schismo-
79

genes is and ar e read y to fly at each o ther's t hroats; whi le with in each
nation are to be observed growing hostilities between the various
social strata, symptoms of comple mentary schismogenes is. Equal ly,
in the count ries ruled by new dictatorship s we may observe early
stages of comple mentary schismogenes is, the behav ior of his
assoc iates pushing the dictator into ever greater pride and
assertiveness .
The purpose of the present article is to suggest problems and
lines of invest igation rather than to state the answer s, but,
tentatively, sugg estions may be offered as to the factors contro lling
schismogenesis :
(a) It is possible that, actually, no healthy equilibrated
relationship between groups is either purely symmetrical or purely
complementary, but that every such relation ship contains elements
of the other t ype. It is tru e that it is ea sy to cl assify relationship s into
one or the other category according to their predo minant emphases,
but it is possible that a very small admixture of comple mentar y
behavio r in a symmetrical relation ship, or a very small admixture of
symmetrica l behavior in a complementar y relationship , may go a
long way toward stabilizing the position. Examples of this type of
stabilization are perhaps common. The squire is in a predominant ly
complementary and not always comfortable relationship with his
village rs, but if he participate in village cricket (a symmetrical
rivalry) but once a year, this may have a curious ly dispropor tiona te
effect upon h is relationsh ip wi th them.
(b) It is certain that, as. in the case quoted above in which group
A sell sago to B while the latter sell fish to A, complementar y
patterns may sometimes have a real stabilizing ef fect by promoting a
mutual dependenc e between th e group s.
(c) It is possible that the presence of a number of truly reciproca l
elements in a relationsh ip may tend to stabilize it, preventing the
schismogenesis which otherwis e might result either from
symmetrica l or comple mentar y elements. But this would seem to be
at best a very weak defense : on the one hand, if we consid er the
effects of symmetrica l schismogenes is upon the reciproc al behavio r
patterns , we see that the latter tend to be less and less exhib ited.
Thus, as the indiv idual s composing the nations of Europe become
more and more involved in their symmetrica l international rivalries,
80

they gradual ly leave off behaving in a reciprocal manner, de-
liberately reducing to a minimum their former recipro cal com-
mercial behavio r.5 On the other hand , if we conside r the effects of
comple mentar y schismogenesis upon the recipro cal behav ior
patterns, we see that one-ha lf of the reciproc al pat-tern is liable to
lapse. Where formerly both groups exhibited both X and Y, a system
gradual ly evolves in which one of the groups exhib its only X, while
the other exhib its only Y. In fact, behavior which was formerly
reciprocal is reduced to a typical complementary pattern and is
likely after that to con tribute to the complementar y schismogenesis .
(d) It is certain that either type of schismogenesi s between two
groups can be checked by factors which unite the two groups either
in loyalty or oppos ition to some outside element. Such an outside
element may be either a symbolic individua l, an enemy peop le or
some quite impersonal circumstance— the lion will lie down with
the lamb if only it rain hard enough . But it must be noted that where
the outside element is a person or group of persons , the relation ship
of the combined groups A and B to the outside group will alwa ys be
itself a potentially schismogenic relationship of one or the other
type. Examination of multiple systems of this kind is badly needed
and especially we need to know more about the systems (e.g.,
military hierarchies) in which the distortion of personality is
modified in the middle groups of the hierarchy by permitting the
individuals to exhibit respect and submission in dealings with
higher groups while they exhibit assertiveness and pride in dealing
with the lower.
(e) In the case of the European situation, there is one other
possibility—a special case of control by divers ion of attention to
outside circumstanc es. It is possible that those respons ible for the
policy of classes and nations might become conscious of the
processe s with which they are playing and cooper ate in an attempt
to solve the difficulties. This, how-eve r, is not very likely to occur
5 In this, as in the other exam ples given, no attempt is made to consider the
schismo genesis from all the points of view outlined in paragraph 10. Thus, inasmuch
as the econo mic aspect of the matter is not here being considered, the effects of the
slump upon the schismogenesis are ignored. A complete study would be sub-
divided into separ ate sections , each treating one of the aspe cts of the phenomena.
81

since anthropolog y and social psycholog y lack the prestige
neces sary to advise; and, with-out such advice, gove rnments will
continue t o react to each other's reaction s rather than pa y attention to
circumstance s.
(21) In conclus ion, we may turn to the proble ms of the
administrator faced with a black-whi te culture contac t. His first task
is to decid e which of the end results outlined in paragraph 8 is
desirable and possib le of attainment. This decision he must make
withou t hypocris y. If he chooses fusion, then he must ende avor to
contrive every step so as to promote the condit ions of consistency
which are outlined (as problems for investigation) in paragr aph 10.
If he chooses that both groups shall persist in some form of dynamic
equilibrium, then he m ust cont rive t o establish a s ystem in whi ch the
possib ilities of schismogenesi s are proper ly compensat ed or
balan ced against each other. But at every step in the scheme which I
have outlined there are proble ms which must be studied by trained
studen ts and which when solved will contribut e, not only to applied
socio logy, but to the very basis of our unders tanding of human
beings in society.
82

Experiments in Thinki ng About
Observed
Ethnological Material*
As I unde rstand it, you have asked me for an hones t, intro-
spective—person al—accoun t of how I think about anthropologica l
materia l, and if I am to be honest and persona l about my thinking ,
then I must be impersona l abou t the results of that think ing. Even if
I can bani sh both pride and shame for half an hour , honesty will still
be difficult.
Let me try to build up. a picture of how I think by giving you an
autobiog raphica l account of how I have acquir ed my kit of
conceptua l tools and intellectu al habits. I do not mean an academic
biograph y or a list of what subjects I have studied, but something
more significant than that—a list rather of the motifs of though t in
various scientific subjects which left so deep an impress ion on my
mind that when I came to work on anthropo logic al material, I
naturally used those borrowed motifs to guide my approach to this
new m aterial.
I owe the greatest part of this kit of tools to my father, William
Bateson, who was a gene ticist. In schoo ls and universities they do
very little to give one an idea of the basic principles of scientific
thinking , and what I lea rned of this ca me in very large measure from
my father's conv ersation and perhaps especially from the overtones
of his talk. He himself was inarticulate about philosophy and
mathematics and logic, and he was articulately distrustful of such
subjects, but still, in spite of himself, I think, he passed on to me
something of these matters.
The attitudes which I got from him were espec ially those which
he had denied in himself. In his early—and as I think he knew—his
best work he posed the proble ms of animal symmetry, segmentat ion,
* This paper was given at the Seventh Conference on Methods in Philosoph y
and the Scien ces, held at the New School for Social Research, April 28, 1940. It is
here -reprint ed from Philosoph y of Science , Vol. 8, No. 1, copyright 1941, The
Williams & Wilkins Co. Reproduced by permission.
83

serial repetition of parts, patterns , etc. Later he turned away from
this field into Mendel ism, to which he devoted the remainder of his
life. Bu t he had alwa ys a hankering a fter the p roble ms of pa ttern and
symmetry, and i t was th is hanke ring and t he m ysticis m that in-spired
it that I picked up and which, for better or worse , I called "science ."
I picked up a vague mystical feeling that we must look for the
same sort of processes in all fields of natural pheno mena—tha t we
might expect to fi nd the same sort of laws at work in the structure of
a crystal as in the structure of society, or that the segmentat ion of an
earthwor m might really be comparable to the process by which
basalt pillars are formed.
I shou ld not preach this mystical faith in quite those terms today
but would say rather that I believe that the types of mental operation
which are usefu l in analyzing one field may be equal ly useful in
anothe r—tha t the framework (the eidos) of science , rather than the
framework of Natu re, is the same in a ll fields . But th e more mystical
phras ing of the matter was what I vague ly learnt, and it was of para-
mount importance . It lent a certain dignity to any scientific
inves tigation, implying that when I was analyzing the pat-terns of
partridg es' feathers, I might really get an answer or a bit of an
answer to the whole puzzling busines s of pattern and regularity in
nature. And further, this bit of mysticism was important becaus e it
gave me freedom to use my scientific background , the ways of
though t that I had picked up in biolog y and elementar y physics and
chemistr y; it encou raged me to expect th ese ways of thought to fi t in
with very different fields of observa tion. It enabled me to regard all
my training as potentially useful rather than utterly irrelevant to
anthropolog y.
When I came into an thropo logy there was a conside rable reaction
taking plac e again st the use of l oose ana logies , espe cially against t he
Spencer ian analogy between the Organis m and Society. Thanks to
this mystical belief in the pervading unity of the pheno mena of the
world , I avoid ed a great deal of intellectual waste. I never had any
doubt that this analog y was fundamental ly sound ; since to doub t
would have been emotion ally expens ive. Nowada ys, of course, the
emphasis has shifted. Few would serious ly doubt that the ways of
analysis which have been found useful in analyzing one complex
functioning system are likely to be of use in analyzing any other
84

similar system. But the mystical prop was usefu l then, though its
phrasing was bad .
There is another wa y, too, in which tha t mysticis m has helped—a
way which is espec ially relevant to my thesis. I want to emphasize
that wheneve r we pride ourselv es upon finding a newe r, stricter way
of thought or exposi tion; when-ever we start insisting too hard upon
"operational ism" or symbolic logic or any other of these very
essent ial systems of tramlines, we lose something of the ability to
think new though ts. And equa lly, of course, wheneve r we rebel
against the sterile rigidity of formal though t and expos ition and let
our ideas run wild, we likewi se lose. As I see it, the advances in
scientific though t come from a combina tion of loose and strict
thinking, and t his combination i s the most precious tool of science .
My mystical view of pheno mena cont ribut ed spec ifically to build
up this doub le habit of mind—it led me into wild "hun ches" and, at
the same time, compelled more formal thinking abou t those
hunch es. It encou raged looseness of thought and then immediately
insisted that that looseness be measured up against a rigid
concreteness . The point is that the first hunch from analogy is wild,
and then, the moment I begin to work out the analogy, I am brought
up again st the rigid formulations which have been devised in the
field from which I borrow the analogy.
Perhaps it is worth giving an example of this; it was a matter of
formulating the social organization of a New Guin ea tribe,—the
Iatmul. The Iatmul social system differs from ours in one very
essent ial point. Their society complete ly lacks any sort of
chieftainsh ip, and I phras ed this matter loose ly by saying that the
control of the individua l was achieved by what I called "lateral"
sanctions rather than by "sanc tions from abov e." Going over my
materia l, I found further that in genera l the subdivi sions of the
society—the clans, moietie s, etc.—had virtually no means of
punishing their own members. I had a case in which a ceremonial
hous e owned by a particula r junior age grade had been defiled, and
though the other members of the grade were very angry with the
defiler, they could do nothing abou t it. I asked whethe r they would
kill one of his pigs or take any of his prope rty, and they replied "No,
of course not. He is a member of their own initiatory grade." If the
same thing had happened in the big senio r ceremonial house which
85

belongs to several grades , then the defiler would be punished . His
own grade would d efend him but the others would s tart a brawl .6
I then began looking for more concre te cases which could be
compared with the contr ast between this system and our own. I said,
"It's like the difference between the radially symmetrica l animals
(jellyfish, sea ane mones, etc.) and the animals whi ch have transver se
segmentation (earthwor ms, lobsters, man, etc.)."
Now in the field of animal segmentation we know very little
about the mechanis ms concern ed, but at least the proble ms are more
concr ete than in the social field. When we compare a social proble m
with a proble m of animal differentiation , we are at once provided
with a visual diagram, in terms of which we may be able to talk a
little more precisely. And for the transver sely segmented animals, at
least, we have something more than a merely anatomical diagram.
Thanks to the work that has been done on exper imental embryology
and axial gradien ts, we have some idea of the dynamics of the
system. We know that some sort of asymmetrica l relation obtains
between the succe ssive segments, that each segment would, if it
could (I speak loosely) form a head , but that the next anterior
segment prevents this. Furth er, this dynamic asymmetry in the
relations between success ive segments is reflected morpholog ically;
we find in most such animals a serial differenc e—what is called
metameric differen tiation—be tween t he. succe ssive segments.
Their appendages, though they can be shown to conform to a
single basic structure, differ one from another as we go down the
series. (The legs of the lobster provide a familiar example of the
sort of thing I mean.)
In contrast with this, in the radially symmetrical animals, the
segments, arranged around the center like sectors of a circle, are
usually all alike.
As I say, we do not know much about the segmentation of
animals, but at least here was enough for me to take back to the
problem of Iatmul social organization. My "hunch" had provided
me with a set of stricter words and diagrams, in terms of which I
could try to be more precise in my thinking about the Iatmul
6 For details of this and other similar incidents cf. Naven, pp. 98-107,
Cambridg e, Cambridg e Universit y Press, 1936.
86

problem. I could now look again at the Iatmul material to
determine whether the relationship between the clans was really in
some sense symmetrical and to determine whether there was
anything that could be compared with the lack of metameric
differentiation. I found that the "hunch" worked. I found that so far
as opposition, control, etc. between the clans was concerned, the
relations between them were reasonably symmetrical, and further,
as to the question of differentiation between them, it could be
shown that, though there were considerable differences, these
followed no serial pattern. Additionally, I found that there was a
strong tendency for clans to imitate each other, to steal bits of each
other's mythological history and to incorporate these into their own
past—a sort of fraudulent heraldry, each clan copying the others so
that the whole system tended to diminish the differentiation
between them. (The system perhaps also contained tendencies in
an opposite direction, but this question I need not discuss now.)
I followed up the analogy in another direction . Impressed by the
pheno mena of m etameric d ifferentiation, I made the poin t that in ou r
society with its hierarch ical systems (comparable to the earthwor m
or the lobste r), when a group seced es from the parent society, it is
usual to find that the line of fission, the division between the new
group and the old, marks a differentiation of mores. The Pilgrim
Fathers wander off in order to be different. But among the Iatmul,
when two groups in a village quarr el, and one half goes off and
found s a new community, the mores of the two groups remain
identical. In our society, fission tends to be heretical (a following
after other doctrines or mores), but in Iatmul, fission is rather
schismatic (a following after other leaders withou t change of
dogma).
You will note that . here I overrode my analogy at one point and
that this matter is still not perfectly clear. When a transve rse fission
or a lateral budding occurs in a transversely segmented animal, the
products of that bud or fission are identical, the posterior half which
was held in check by the anterior is relieved of this control and
develops into a normal, complete animal. I am therefo re not in step
with my analogy when I regard the differen tiation which accom-
panies fission in a hierarchica l society as comparable with that
which exists befor e fission in a transvers ely segmented animal. This
87

divergence from the analogy will surely be worth inves tigation; it
will take us into a more precise study of the asymmetrical relations
which obtain between the units in the two cases and raise question s
about the reactions of the subo rdina te member to its position in the
asymmetry. This aspec t of the matter I have no t yet examined.
Having got some sort of conceptua l frame within which to
describe the interrelations between clans, I went on from this to
consid er the interrelations between the variou s age grades in terms
of this same frame. Here , if anywhere, where age might be expec ted
to provid e a basis for serial differentiation, we ought to expec t to
find some analogue of the transve rse segmentation with
asymmetrical relation s between the successive grades—and to a
certain extent the age-g rade system fitted this picture. Each grade
has its ceremonies and its secrets of initiation into that grade ; and in
these ceremonies and secrets it was perfectly easy to trace a met-
americ differentiation . Ceremonies which are fully developed at the
top of the system are still recogn izable in their basic form in the
lower l evels—but m ore rud imentar y at each l evel a s we go down th e
series.
But the initiatory system contains one very interesting element
which was brough t into sharp relief when my point of view was
defined in terms of animal segmentation. The grades alterna te, so
that the whole system consi sts of two oppo sed groups , one group
made up of grade s 3, 5, 7, etc. (the odd numbers), and the other
made up of 2, 4, 6, etc.; and these two group s maintain the type of
relationship which I had already described as "symmetrica l"—each
provid ing sanctions by quarreling with the other when their rights
are infringed .
Thus even where we might expec t the most definite hierarchy,
the Iatmul have substituted for it a headless system in which one
side is symmetrically opposed to the other.
From this conclus ion my enqu iry, influenced by many other
types of material, will go on to look at the matter from other points
of view—especia lly the psycholog ical proble ms of wheth er a
preference for symmetrical rather than asymmetrica l relationships
can be implanted in the individual , and what the mechanis ms of
such chara cter formation may be. But we need no t go into that now.
88

Enough has been said to bring out the methodolog ical theme—
that a vague "hunch " derived from some other science leads into the
precis e formulations of that other science in terms of which it is
possible to think m ore fruit-fully abou t our own materia l.
You will have noticed that the form in which I used the
biologi cal findings was really rather different from that in which a
zoologis t would talk about his materia l. Where the zoologi st might
talk of axial gradien ts, I talked abou t "asymmetrica l relationsh ips
between su ccess ive segments," and in my phrasing I was p repared to
attach to the word "successive" two simultaneous meanings—in
referring to the animal material it meant a morphologic al series in a
three-dimensiona l concre te organis m, while in referring to the
anthropo logica l material the word "succ essive " meant some
abstracted prope rty of a hierarch y.
I think it would be fair to say that I use the analog ies in some
curious ly abstract form—that , as for "axial gradien ts" I substitute
"asymmetrica l relation ships ," so also I endow the word "success ive"
with some abstract meaning which makes it applicable to both sorts
of cases.
This bring s us to another very importan t motif in my thinking—a
habit of constructing abstractions which refer to terms of
comparison between entities; and to illustrate this I can clearly
remember the first occas ion on which I was guilty of such an
abstraction . It was in my Zoolog ical Tripos examination at
Cambridge , and the examiner had tried to compel me to answer at
least one quest ion on each branch of the subject. Comparat ive
anatomy I had alwa ys regarded as a waste of time, but I found
myself face to face with it in the examnation and had not the
necessa ry detailed knowledge . I was asked to compare the
urinogen ital system of the amphibia with that of the mammalia, and
I did not know much abou t it.
Neces sity was the mother of invention . I decided that I ought to
be able to defend the position that comparative anatomy was a
muddled waste of time, and so I set to work to attack the whole
emphasis on homology in zoologica l theory. As you probab ly will
know , zoologi sts conv ention ally deal in two sorts of comparabi lity
between organs—homology and analogy. Organs are said to be
"homologous" when it can be shown that they have similar structure
89

or bear similar structural relation s to other organs, e.g., the trunk of
the elephan t is homologous with the nose and lip of a man be-cause
it has the same formal relation to other parts— eyes, etc.; but the
trunk of an elephant is analogous to the hand of a man becau se both
have the same uses. Fifteen years ago comparativ e anatomy
revolved endlessly around these two sorts of comparabi lity, which
incidenta lly are good examples of what I mean by "abstraction s
which d efine the terms of a comparison between entities."
My attack on the system was to sugges t that there might be other
sorts of comparability and that these would con-fus e the issue to
such a degree that mere morphologica l analysis would not suffice. I
argued that the bilateral fins of a fish would convent ional ly be
regarded as homologous with the bilateral limbs of a mammal, but
that the tail of a fish, a median organ, would conven tiona lly be
regarded a "different from" or at most only "analogou s to" the fins.
But what abou t the doub le-tailed Japanese goldf ish? In this animal
the factors causing an anomaly of the tail also cause the same
anomaly in the bilateral fins; therefo re there was here anothe r sort of
comparabi lity, an equiva lence in terms of processes and laws of
growth . Well, I don't know what mark I got for my answer . I found
out m uch later that, as a m atter o f fact, the lateral fins of the go ldfish
are scarcely, if at all, affected by the factors which cause the
anomaly in the tail, but I doub t if the examiner caugh t me in my
bluff; and I found also that, curious ly, Haekel in 1854 had actually
coined the word "homonomy" for the very type of equivalence that I
was invent ing. The word is, so far as I know , obsolete, and was
obsole te when I wrote my answer .
So far as I was conce rned, howeve r, the idea was new and I had
though t of it myself. I felt that I had discove red how to think. That
was in 1926 , and this same old clue—re cipe, if you like—has
remained with me ever since. I did not realize that I had a recipe;
and it was not until ten years later that I fully grasp ed the
significance of this analogyhomolog y-homonomy busine ss.
Perhaps it will be of interest to recount in some detail my various
brushe s with these concep ts and the recipe which they conta ined.
Soon after the examination to which I have referred, I went into
anthropolog y and for some time stopped thinking—wondering
rather what could be made of this subjec t, but not getting anything
90

clear except a repudi ation of most of the convent ional approache s
which , to me, seemed meaningle ss. I wrote a little skit on the
concept of totemism in 1930 , first proving that the totemism of the
Iatmul is true totemism becaus e it contains a "high percen tage" of
characteristics of totemism listed in "Notes and Quer ies on Anth ro-
pology" issued more or less ex cathedra by the Royal
Anth ropologi cal Institute, and then going on to the question, what
sort of equiva lence we though t we were referring to when we equate
some bits of Iatmul culture with the totem-ism of North America,
and dragg ing in homolog y-homonomy, etc.
In this discus sion of "true" totemism I still had the homonomy-
homolog y abstractions perfec tly clear and was using the concep ts
with a clean (though inarticulate) under stand ing of what sort of
abstraction s they were—bu t it is interesting that I afterward s made
some other comparable abstractions for the analysis of latmul
materia l and muddled the issues through f orgetting this very thing.
I was especially interested in studying what I called the "feel" of
culture, and I was bored with the convent ional study of the more
formal details. I went out to New Guine a with that much vaguel y
clear— and in one of my first letters home I complained of the
hope lessne ss of putting any sort of salt on the tail of such an
imponderable concep t as the "feel" of culture. I had been watching a
casual group of natives chewing betel, spitting, laughing , joking,
etc., and I felt acutely the tantalizing impossib ility of what .I wanted
to do.
A year later, still in New Guin ea, I read Arabia Deserta and
recogniz ed with a thrill that Dought y had in a sense done what I
want ed to do. He had put salt on the tail of the very bird that I was
hunting. But I realized also—sadl ythat he had used the wrong kind
of salt. I was not interested in achieving a literary or artistic
repres entation of the "feel" of the culture; I was interested in a
scientific an alysis of it.
On the whol e I think that Dought y was an encourag ement to me,
and the greatest encourage ment I got from him was due to a
fallacious bit of thinking which he prompted. It appea red to me that
it was impossible to unders tand the behavio r of his Arabs apar t from
the "feel" of their culture, and from this it seemed to follow that the
"feel" of the culture was in some way causative in shaping native
91

behavio r. This encour aged me to go on thinking that I was trying
after something that was importan t—so far so good . But it also
guided me into regard ing the "feel" of the culture as much more
concr ete and c ausal ly active than I had any right to do.
This false concreteness was reinfo rced later by an accident of
languag e. Radc liffe-Brown called to my attention the old word
"ethos" and told me that that was what I was trying to study. Words
are dange rous things , and it so hap-p ens that "etho s" is in some
ways a very bad word . If I had been compelled to make up my own
word for what I wanted to say, I might have done better and saved
myself a great deal of confus ion. I would, I hope, have put forward
something like "ethono my," which would have reminded me that I
was referring to an abstraction of the same order as homolog y or
homonomy. The troubl e with the word "ethos" is just this— that it is
too short. It is a unit word, a single Greek substan tive, and as such
helped me to go on thinking that it referred to a unit something
which I could still regard as causative. I hand led the word as if it
were a categor y of behavio r or a sort of factor which shaped
behavio r.
We are all familia r with this loose use of word s in such phras es
as: "the causes of war are econo mic," "econo mic behavior ," "he was
influenced b y his emotions," "hi s symptoms are the resu lt of conflict
between his super ego and his id." (I am not sure how many of these
fallacies are conta ined in that last example; at a rough count , there
seem to be five with a possib le sixth, but there may be more.
Psychoanal ysis has erred sadly in using words that are too short and
there-fo re appear more concrete than they are.) I was guilty of just
this sort of shodd y thinking in my handl ing of the word "ethos ," and
you must excuse me if I have gathered moral support for this
confes sion by a digression to show that at any rate other s have
committed the same crime.
Let us examine the stages by which I got into the fallacy and the
way in which I got out of it. I think the first step toward an escape
from sin was to multiply offenses—and there is a good deal to be
said for this method. Vice is after all a dull busines s whethe r it be
physical or intellectual, and an effective cure can sometimes be
achieved by indulgence to the point at which the patient realizes
this. It is a way of proving that a given line of thought or conduct
92

will not do, by exper imentally extrapo lating it to infinity, when its
absurdi ties beco me eviden t.
I multiplied my offenses by creating severa l more concepts of
abou t the same degree of abstraction as "ethos"—I had "eido s," "cultural structur e," "sociolog y"—and all these I hand led as though
they were conc rete entities. I pictured the relations between ethos
and cultural structure as being like the relation between a river and
its bank s—"Th e river molds the banks and the banks guid e the river.
Similar ly, the ethos molds the cultural structure and is guided by it."
I was still looking for physical analogies , but now the position was
not quite the same as when I was looking for analogies in order to
get concep ts which I c ould use i n anal yzing observed m ateria l. I was
looking now for physical analog ies which I could use in analyzing
my own conc epts, and that is a very much less satisfac tory business .
I do not mean, of course, that the other sciences can give one no
help in the attempt to straighten out one's thoughts ; they surely can.
For example, the theory of Dimensions in physics may be of
enormous help in this field. What I mean is that when one is seeking
an analogy for the elucida tion of material of one sort, it is good to
look a t the wa y analogous material has been analyzed. Bu t when one
is seeking an elucidation of one's own concep ts, then one must look
for analogi es on an equally abstract level. Howeve r, these similes
abou t rivers and their banks seemed pretty to me and I treated them
quite seriousl y.
Here I must digress for a moment to describe a trick of though t
and speech , which I have found usefu l. When I am faced with a
vagu e conc ept and feel that the time is not yet ripe to bring that
concept into strict express ion, I coin some loose expression for
referring to this concept and do not want to prejudge the issue by
giving the concept too meaningful a term. I therefore dub it hastily
with some brief concrete colloquial term—generally Anglo-Saxon
rather than Latin—I will speak of the "stuff" of culture, or "bits" of
culture, or the "feel" of culture. These brief Anglo-Saxon terms
have for me a definite feeling tone which reminds me all the time
that the concepts behind them are vague and await analysis. It is a
trick like tying a knot in a handkerchief—but has the advantage
that it still permits me, if I may so express it, to go on using the
handkerchief for other purposes. I can go on using the vague
93

concept in the valuable process of loose thinking—still continually
reminded that my thoughts are loose.
But these similes about ethos being the river and the
formulations of culture or "cultural structure" being its banks were
not Anglo-Saxon reminders that I was leaving some-thing for
analysis at a later date. They were, as I thought, the real thing—a
real contribution to our understanding of how culture works. I
thought that there was one sort of phenomenon which I could call
"ethos" and another sort which I could call "cultural structure" and
that these two worked together—had mutual effect one on the
other. All that remained for me to do was to discriminate clearly
between these various sorts of phenomena so that other people
could perform the same sort of analysis that I was doing.
This effort of discrimination I postponed, feeling perhaps that
the problem was not quite ripe—and I went on with the cultural
analysis. And did what I still think was good work. I want to
emphasize this last point—that, as a matter of fact, considerable
contributions to science can be made with very blunt and crooked
concepts. We may joke about the way misplaced concreteness
abounds in every word of psycho-analytic writing—but in spite of
all the muddled thinking that Freud started, psychoanalysis
remains as the outstanding contribution, almost the only
contribution to our understanding of the family—a monument to
the importance and value of loose thinking.
Finally I had completed my book on Iatmul culture, with the
excep tion of t he la st chap ter, the wri ting of which was t o be the f inal
testing and review of my various theoretical concep ts and
contribut ions. I plann ed that this chap ter should contain some
attempt to discriminate between the sort of thing t hat I called "ethos "
and t he sort of thing that I called "e idos, " etc.
I was in a state approxi mating t hat panic in the exa mination room
which formerly produced the concep t of homonomy. I was due to
sail for my next field trip—my book had to be finished before I
sailed—th e book could not stand without some clear statement
about the interrelations of these concep ts of mine.
Here I will quote what finally appeared in the book in this last
chapt er:
94

"I began to doubt the validity of my own categori es, and
performed an expe riment. I chose three bits of culture: (a) a wau
(mother's brother) giving food to a laua (sister's son); a pragmatic
bit, (b) a man scolding his wife; an ethologic al bit, and (c) a man
marrying his father's sister's daugh ter; a structu ral bit. Then I drew a
lattice of nine squar es on a large piece of pape r, three rows of
squares with three squar es in each row. I labeled the horizon tal rows
with my bits of culture and the vertical columns with my catego ries.
Then I forced myself to see each bit as conceivab ly belonging to
each c ategor y. I found th at it could be done.
"I found that I could think of each bit of culture structurally; I
could see it as in accordance with a consisten t set of rules or
formulations. Equa lly, I could see each bit as `pragmatic,' either as
satisfying the needs of individuals or as contr ibuting to the
integration of soc iety. Again, I could s ee each b it etholog ically, as an
expression of emotion .
"This experiment may seem puerile, but to me it was very
important , and I have recoun ted it at length because there may be
some among my reade rs who tend to regard such concep ts as `structure' as concr ete parts which `interac t' in culture, and who find,
as I did, a difficulty in think ing of these concep ts as labels merely
for points of view adop ted either by the scientist or by the natives . It
is instructive to perform the same expe riment with such concepts as
economics, etc."7
In fact, "ethos" and the rest were finally reduced to abstractions
of the same gene ral order as "homology," "homonomy," etc.; they
were labels for points of view volun tarily adopt ed by the
invest igato r. I was, as you may imagine , enormously excited at
getting this tangle straighten ed out—bu t I was also worr ied because
I though t I shou ld be compelled to rewrite the whole book. But I
found that this was not so. I had to tune up the defin itions, check
through to see that each time the technica l term appeared I could
substitute the new definition for it, mark the more egreg ious pieces
of nonsense with footno tes warning the reader that these passages
might be taken as a warning of how not to say things—and so on.
7 Loc. cit., p. 261.
95

But the body of the book was sound enough—a ll that it need ed was
new castors on its legs.
So far I have spoken of my own personal experiences with strict
and loose thinking, but I think actually the story which I have
narrated is typical of the whole fluctuating business of the advance
of science. In my case, which is a small one and comparatively
insignificant in the whole advance of science, you can see both
elements of the alternating process—first the loose thinking and
the building up of a structure on unsound foundations and then the
correction to stricter thinking and the substitution of a new
underpinning beneath the already constructed mass. And that, I
believe, is a pretty fair picture of how science advances, with this
exception, that usually the edifice is larger and the individuals who
finally contribute the new underpinning are different people from
those who did the initial loose thinking. Sometimes, as in physics,
we find centuries between the first building of the edifice and the
later correction of the foundations—but the process is basically the
same.
And if you ask me for a recipe for speeding up this process, I
would say first that we ought to accept and enjoy this dual nature
of scientific thought and be willing to value the way in which the
two processes work together to give us advances in understanding
of the world. We ought not to frown too much on either process, or
at least to frown equally on either process when it is
unsupplemented by the other. There is, I think, a delay in science
when we start to specialize for too long either in strict or in loose
thinking. I suspect, for example, that the Freudian edifice has been
al-lowed to grow too big before the corrective of strict thought is
applied to it—and now when investigators start rephrasing the
Freudian dogmas in new stricter terms there may be a lot of ill
feeling, which is wasteful. (At this point I might perhaps throw out
a word of comfort to the orthodox in psychoanalysis. When the
formulators begin rooting about among the most basic of analytic
premises and questioning the concrete reality of such concepts as
the "ego" or "wishes" or the "id" or the "libido"—as indeed they
are already be-ginning to root—there is no need to get alarmed and
to start having terror dreams of chaos and storms at sea. It is
certain that most of the old fabric of analysis will still be left
96

standing after the new underpinning has been inserted. And when
the concepts, postulates, and premises have been straightened out,
analysts will be able to embark upon a new and still more fruitful
orgy of loose thinking, until they reach a stage at which again the
results of their thinking must be strictly conceptualized. I think that
they ought to enjoy this alternating quality in the progress of
science and not delay the progress of science by a refusal to accept
this dualism.)
Further than this, besides simply not hindering progress, I think
we might do something to hasten matters, and I have suggested
two ways in which this might be done. One is to train scientists to
look among the older sciences for wild analogies to their own
material, so that their wild hunches about their own problems will
land them among the strict formulations. The second method is to
train them to tie knots in their handkerchiefs whenever they leave
some mat-ter unformulated—to be willing to leave the matter so
for years, but still leave a warning sign in the very terminology
they use, such that these terms will forever stand, not as fences
hiding the unknown from future investigators, but rather as
signposts which read: "UNEXPLORED BEYOND THIS POINT."
97

Mora le and National Chara cter*
We shall proceed as follows: (1) We shall examine some of the
criticisms which can be urged agains t our entertaining any concep t
of "nationa l chara cter." (2) This examination will enable us to state
certain concep tual limits within which the phras e "nationa l
character" is likely to be valid. (3) We shall then go on, within these
limits, to outline what order s of differen ce we may expect to find
among Western nations, trying, by way of illustration, to guess more
concr etely at some of these difference s. (4) Last ly, we shall consid er
how the problems of morale and international relations are affected
by differences of this order.
Barriers to Any Concept of "National Character"
Scien tific enquir y has been diverted from questions of this type
by a number of trains of though t which lead scientists to regard all
such quest ions as unprofitable or unsound. Be-fo re we hazard any
const ructive opinion as to the order of differenc es to be expec ted
among European popula tions , therefo re, these diverting trains of
though t must be ex amined.
It is, in the first place, argued that not the peop le but rather the
circumstance s under which they live differ from one community to
anothe r; that we have to deal with differences either in historical
background or in current conditions, and that these factors are
sufficient to account for all differen ces in behavior without our
invok ing any difference s of characte r in the individua ls conce rned.
Essent ially this argument is an appe al to Occa m's Razo r—an
assertion that we ough t not to multiply entities beyond neces sity.
The argument is that, wher e observabl e differences in circumstance
exist, we ough t to invok e those rather than mere inferred differences
in chara cter, which we canno t obse rve.
* This essay appeared in Civilian Morale , edited by Good win Watson , copyright
1942 b y the Societ y for the Psycholog ical Stud y of Social Issues. It is here reprint ed
by permission of the publisher . Some introductor y material has been edited out .
98

The argument may be met in part by quoting exper imental data,
such as Lewin's experiments (unpubl ished material), which showed
that there are great differences in the way in which Germans and
Americans respond to failure in an expe rimental setting. The
Americans treated failure as a challeng e to increas e effort; the
Germans responded to the same failure with discou ragement. But
those who argue for the effectiveness of condi tions rather than
character can still reply that the exper imental condi tions are not, in
fact, the same for both groups; that the stimulus value of any
circumstance depends upon how that circumstance stands out
against the backg round of other circumstances in the life of the
subject, and that this con trast cannot be the same for both groups .
It is possib le, in fact, to argue that since the same circumstances
never occur for individuals of differen t cultural back-g round, it is
therefore unnec essar y to invoke such abstractions as nationa l
character. This argument breaks down, I believe , when it is pointed
out that, in stressing circumstance rather than chara cter, we would
be ignoring the known facts about learning . Perhaps the best
docu mented gene ralization in the field of psycholog y is that, at any
given moment, the behavioral characteristics of any mammal, and
especia lly of man, depend upon the previou s expe rienc e and
behavior of that individua l. Thus in presuming that characte r, as wel l
as circumstance , must be taken into account , we are not multiplying
entities beyond necess ity; we know of the signif icance of learned
character from other types of data, and it is this knowledge which
compels us to consid er the additiona l "en tity."
A second barrier to any accep tance of the notion of "national
character" arises after the first has been nego tiated. Those who grant
that character must be conside red can still doubt whether any
uniformity or regularity is likely to obtain within such. a sample of
human being s as constitutes a nation . Let us grant at once that
uniformity obviousl y does not occur, and let us proceed to conside r
what sorts of regularity may be expected .
The criticism which we are trying to meet is likely to take five
forms. (1) The critic may point to the occu rrence of subcul tural
differentiation, to difference s between the sexes, or between classes,
or between occup ation al groups within the community. (2) He may
point to the extreme heterogenei ty and confus ion of cultural norms
99

which can be observed in "melting -pot" communities. (3) He may
point to the acciden tal devian t, the individua l who has unde rgone
some "accidenta l" traumatic experi ence, not usual among those in
his social environment. (4) He may point to the pheno mena of
cultural chang e, and especi ally to the sort of differen tiation which
results when one part of the community lags behind some other in
rate of chang e. (5) Lastly, he may point to the arbitrary nature of
national boundar ies.
These objections are closely interrelated, and the replies to them
all derive ultimatel y from two postulates: first, that the individu al,
whether from a physiologi cal or a psycho-logi cal point of view, is a
single organized entity, such that all its "parts" or "aspec ts" are
mutually modifiab le and mutually interac ting; and second , that a
community is like-wi se organized in this sen se.
If we look a t social differentiation in a s table co mmunity—say, at
sex differentiation in a New Guin ea tribe8—we find that it is not
enough to say that the habit system or the character structure of one
sex is different from that of anothe r. The signif icant point is that the
habit system of each sex cogs into the habi t system of the othe r; that
the behavior of each promotes the habits of the other.9 We find, for
example, between the sexes , such comple mentar y patterns as
spectatorship-e xhibi tionism, dominance -submission , and succoring –
dependenc e, or mixtures of these. Never do we find mutual
irrelevance between such g roups.
Although it is unfortuna tely true that we know very little about
the terms of habit differentiation between classes , sexes,
occupa tiona l groups , etc., in Western nations , there is, I think, no
danger in applying this genera l conclus ion to all cases of stable
differen tiation between groups which are living in mutual contac t. It
8 Cf. M. Mead (Sex and Temperament in Three Primitive Societies, New York,
Morro w, 1935) , especiall y Part III, for an analysis of sex differentiation amon g the
Chambuli; also G. Bateson (Naven, Cambridg e, Cambrid ge Universit y Press, 1936)
for an analysis of sex differentiation among adults in Iatmul, New Guinea.9 We are considering here only those cases in which ethological differentiation
follows the sex dichotomy. It is also probable that, wher e the ethos of the two sexes is
not sharpl y differentiated, it would still be correct to say that the ethos of each
promotes that of the other , e.g., through such mechanisms as competition and mutual
imitation. Cf. M. Mead (op. cit.).
100

is, to me, incon ceivab le that two differing groups could exist side by
side in a community with-ou t some sort of mutual relevanc e
between the special characteristics of one group and those of the
other. Such an occurrence would be contrary to the postulate that a
community is an organiz ed unit. We shall, therefor e, presu me that
this gener alization app lies to all stable soc ial differentiation .
Now, all that we know of the m echanics of ch aracter fo rmation—
especia lly the proce sses of projec tion, reaction formation,
compensation , and the like— force s us to regard these bipolar
patterns as unitary within the individual . If we know that an
individu al is trained in overt express ion of one-half of one of these
patterns, e.g., in dominance behav ior, we can predict with certainty
(though not in precise language) that the seeds of the other half—
submission—a re simultaneous ly sown. in his persona lity. We have
to think of the individual , in fact, as trained in dominance –
submission , not in either dominance or submission. From this it
follows that where we are dealing with stable differen tiation within
a community, we are justified in ascribing common character to the
members of that community, provid ed we take the precau tion of
describ ing that common chara cter in terms of the motifs of
relationsh ip be tween t he differentiated sections of the community.
The same sort of conside rations will guide us in dealing with our
second criticism—the extremes of heterogenei ty, such as occur in
modern "melting-pot " communities. Suppose we attempted to
analyze out all the motifs of relationsh ip between individu als and
groups in such a community as New York City; if we did not end in
the madhouse long before we had completed our study, we shou ld
arrive at a picture of common character that would be almost
infinitely complex—cer tainly that would conta in more fine
differentiations than the human psyche is capabl e of resolving
within itself. At this point, then, both we and the individua ls whom
we are studying are forced to take a short cut: to treat heterogen eity
as a posi tive char acteristic of the common environment, sui gener is.
When, with such an hypothesis, we begin to look for common
motifs of behavior , we note the very clear tenden cies toward
glorying in heterogenei ty for its own sake (as in the Robinson
Latouche "Ballad for American s") and toward regard ing the world
101

as made up of an infinity of disconn ected quiz-bits (like Riple y's
"Beli eve I t or Not").
The third objection, the case of the individual deviant, falls in
the same frame of reference as that of the differentiation of stable
groups. The boy on whom an English public-school education does
not take, even though the original roots of his deviance were laid
in some "accidental" traumatic incident, is reacting to the public-
school system. The behavioral habits which he acquires may not
follow the norms which the school intends to implant, but they are
acquired in reaction to those very norms. He may (and often does)
acquire patterns the exact opposite of the normal; but he cannot
conceivably acquire irrelevant patterns. He may become a "bad"
public-school Englishman, he may become insane, but still his
deviant characteristics will be systematically related to the norms
which he is resisting. We may describe his character, indeed, by
saying that it is as systematically related to the standard public-
school character as the character of Iatmul natives of one sex is
systematically related to the character of the other sex. His
character is oriented to the motifs and patterns of relationship in
the society in which he lives.
The same frame of reference applies to the fourth consideration,
that of changing communities and the sort of differentiation which
occurs when one section of a community lags behind another in
change. Since the direction in which a change occurs will
necessarily be conditioned by the, status quo ante, the new
patterns, being reactions to the old, will be systematically related
to the old. As long as we confine ourselves to the terms and
themes of this systematic relationship, therefore, we are
entitled to expect regularity of character in the individuals.
Furthermore, the expectation and experience of change may, in
some cases, be so important as to become a common character-
102

determining factor10 sui generis, in the same sort of way that
"heterogeneity" may have positive effects.
Lastly, we may conside r cases of shifting nation al boundaries,
our fifth criticism. Here, of course , we cannot expect that a
diplomat's signature on a treaty will immediate ly modify the
characters of the individu als whose nationa l allegian ce is thereb y
changed. It may even happen—fo r example, in cases where a
preliterate native popu lation is brought for the first time in contact
with Europeans— that, for some time after the shift, the two parties
to such a situation will behave in an exploratory or almost rando m
manner, each retaining its own norms and not yet deve loping any
special adjustments to the situation of contact. During this period,
we shou ld still not expec t any gener alizations to apply to both
groups. Very soon , howeve r, we know that each side does deve lop
special patterns of behav ior to use in its contacts with the other.11 At
this point, it beco mes meaningful to ask what systematic terms of
relationsh ip will descr ibe the common character of the two groups ;
and from this point on, the degree of common character structure
will increase until the two groups become related to each other just
as two classes or two sexes in a stable, differentiated society.12
In sum, to those who argue that human communities show too
great internal differentiation or contain too great a random element
for any notion of common character to apply, our reply would be
that we expect such an approach to be useful (a) provided we
10 For a discussion of the role played by "chan ge" and "heterogeneity" in
melting-pot communities, cf. M. Mead ("Educative effects of social
environment as disclosed by studies of primiti ve societies." Paper read at the
Symposiu m on Environm ent and Education, Universit y of Chicago, September 22,
1941). Also F. Alexander ("Educative influence of personalit y factors in the
environment." Paper read at the Symposium on Environm ent and Education ,
Universit y of Chicago, September 22,1941) .11 In the South Seas, those special modes of behavior which Europeans
adopt toward native peoples , and those other modes of behavior which the native
adopts toward Europeans, are very obvious. Apart from analyses of "pidgin"
langua ges, we have, however, no psycholo gical data on these patterns. For a
description of the analogous patterns in Negro-white relationships, cf. J.
Dollard (Caste and Class in a Southern Toivn, New Haven, Yale University
Press, 1937), especially Chapter XII, Accommodation Attitudes of Negroes.12 Cf. G. Bateson, "Culture Contact and Schismogen esis," Man, 1935, 8: 199.
(Reprinted in this volume.)
103

describe common character in terms of the themes of relationship
between groups and individuals within the community, and (b)
provided that we allow sufficient time to elapse for the community
to reach some degree of equilibrium or to accept either change or
heterogeneity as a characteristic of their human environment.
Differenc es Wh ich W e May Expec t Between Nationa l Groups
The abov e examination of "straw men" in the case against
"national charac ter" has very string ently limited the scope of this
concep t. But the conclusions from this examination are by no means
simply negat ive. To limit the scope of a conc ept is almost
synonymous with defin ing it.
We have added one very important tool to our equip ment —the
techn ique of describing the common character (or the "high est
common factor" of chara cter) of individua ls in a human community
in terms of bipola r adjectives . Instead of despa iring in face of the
fact that nations are highl y differentiated, we shall take the
dimension s of that differentiation as our clues to the nationa l
character. No longe r conten t to say, "Germans are submissive," or
"Engli shmen are aloof," we shall use such phrases as "dominant –
submissive" when relationsh ips of this sort can be shown to occur .
Similarly, we shall not refer to "the paranoid al element in German
character," unless we can show that by "parano idal" we mean some
bipolar chara cteristic of German- German or German-fore ign
relationships . We shall not descr ibe varieties of character by
defining a given character in terms of its position on a continuu m
between extreme dominance and extreme submissiveness , but we
shall, instead, try to use for our descriptions some such continua as
"degre e of interest in, or orientation toward , dominance –
submission."
So far, we have mentioned only a very short list of bipola r
characteristics: dominance-sub mission, succo ring-d ependence , and
exhib itionis m-spectatorship . One criticism will certainly be
upper most in the reader 's mind, that, in short, all three of these
characteristics are clearly present in all West-e rn cultures. Befor e
104

our method beco mes useful, therefore, we must try to expand it to
give us sufficient scope and discriminator y power to differentiate
one Western culture from another.
As this concep tual frame deve lops, no doub t, many further
expansions and discriminations will be introduced . The present
paper wi ll deal with only three su ch types of expans ion.
Alterna tives to Bipolarity
When we invoked bipola rity as a means of handl ing differ-
entiation within society without foregoing some notion of common
character structure, we conside red only the possib ility of simple
bipola r differentiation. Certainly this pattern is very common in
Western cu ltures; take , for instance, Republican-De mocrat, poli tical
Righ t-Lef t, sex differentiation, God and the devil, and so on. These
peop les even try to impose a bina ry pattern upon pheno mena which
are not dual in nature— youth versus age, labor versus capital, mind
versus matter—and , in gene ral, lack the organiza tiona l devices for
hand ling triangul ar systems; the inception of any "third " party is
always regarded, for example, as a threat to our political
organiza tion. This clear tendenc y toward dual systems ought not,
howeve r, to blind us to the occur rence of other patterns.13
There is, for example, a very interes ting tendenc y in English
communities toward the formation of ternary systems, such as
parents -nurse -child, king-m inister s-peop le, officers-N.C .O.'s-
privates.14 While the precise motifs of relationship in these ternary
13 The Balinese social system in the mountain communities is almost entirel y
devoid of such dualisms. The etholo gical di fferentiation o f the se xes is rather sli ght;
political factions are completel y absent. In the plains, there is a dualism which has
resulted from the intrusive Hindoo caste system, those with caste being discrimi-
nated from those without caste. At the symbolic level (partl y as a result of Hindoo
influence) dualisms are much more frequent, however, than they are in the social
structure (e.g., Northeast vs. Southwest, Gods vs. demons, symbolic Left vs. Right,
symbolic Male vs. Female, etc.).14 A fourth instance o f this three fold pattern occurs in some great public s chools
(as in Charterhous e), where the authorit y is divided between the quieter , more
105

systems remain to be invest igated, it is important to note that these
systems, to which I refer as "terna ry," are neither "simple
hierarch ies" nor "triangle s." By a pure hierarchy, I should mean a
serial system in which face-t o-face relations do not occur between
members when they are separa ted by some interv ening member; in
other words, systems in which the only communicat ion between A
and C passes through B. By a triangle I should mean a threefo ld
system with no serial properties. The ternary system, paren t-nurse-
child, on the other hand, is very differen t from either of these other
forms. It contains serial elements, but face-t o-face conta ct does
occur between the first and the third members. Essentially, the
function of the middle member is to instruct and discipline the third
member in the forms of behav ior which he shou ld adopt in his
conta cts with the first. The nurse teaches the child how to behave
toward its parents , just as the N.C.O. teaches and disciplines the
private in how he should behave toward officers. In psychoanal ytic
terminolog y, the process of introjection is done indirectly, not by
direct impact of the paren tal personal ity upon the child.15 The face-
to-face contacts between the first and third members are, however ,
very important. We may refer, in this connection , to the vital daily
ritual in the British Army, in which the officer of the day asks the
assembled private s and N.C.O.'s whethe r there are any complaint s.
Certa inly, any full discussion of English charac ter ough t to allow
for ternary, as wel l as bipolar patterns .
Symm etrical Motifs
So far, we have considered only what we have called "com-
plementary" patterns of relationsh ip, in which the behavior patterns
at one end of the relation ship are differen t from, but fit in with, the
behavio r patterns at the other end (dominance -submission, etc.).
polished, intelle ctual leaders ("monitors ") and the rougher , louder , athletic leaders
(captain of football, head of long room, etc.), who have the duty of seein g to it that
the " fags" run when the monitor c alls.15 For a general discussion of cultural variants of the Oedipus situation and the
related s ystems of cultural sanctions , cf. M . Mead ("Social chan ge and cultural
106

There exists, however , a whole categor y of human interpersona l
behavior which does not conform to this description . In addition to
the contrasting complementary patterns, we have to recogn ize the
existence of a series of symmetrical patterns, in which peop le
respond to what others are doing by themselves doing something
similar. In particular, we have to conside r those competitive16 pat-
terns in which individual or group A is stimulated to more of any
type of behav ior by perceiving more of that same type of behav ior
(or greater success in that type of behav ior) in individua l or group B.
There is a very profound contrast between such competitive
systems of behav ior and comple mentar y dominance -submission
systems—a highly significan t contra st for any discussion of national
character. In comple mentary striving , the stimulus which prompts A
to grea ter efforts is the relative weakness in B; if we want to make A
subside or submit, we ough t to show him that B is strong er than he
is. In fact, the complementary charac ter structure may be
summarized by the phrase "bully-coward ," implying the
combination of these charac teristics in the personal ity. The
symmetrical competitive systems, on the other hand , are an almost
precis e funct ional oppo site of the complementar y. Here the stimulus
which evokes greater striving in A is the vision of greater strength
or greater striving in B; and , inversely, if we demonstrate to A that B
is really weak , A will relax h is efforts.
It is probabl e that these two contr asting patterns are alike
available as potentialities in all human beings; but clearly, any
individu al who behaves in both ways at once will risk internal
confusion and conflict. In the various national groups , consequent ly,
different methods of resolv ing this discrepancy have deve loped. In
Engl and and in America , where children and adults are subjec ted to
an almost continuous barrage of disapprova l wheneve r they exhibit
the complementary patterns , they inevitably come to accep t the
16 The term "cooperation," which is sometimes used as the opposite of
"competition," covers a very wide variety of patterns , some of them symmetrical and
others complementar y, some bipolar and others in which the cooperatin g
individuals are chiefly oriented to some personal or impersonal goal. We may
expect that some careful analysis of these patterns will give us vocabulary for
describing other sorts of national characteristics. Such an anal ysis cannot be attempted
in this paper .
107

ethics of "fair play." Responding to the challeng e of difficulties, they
cannot , without guilt, kick the unde rdog.17 For British morale
Dunkirk was a stimulus, not a dep ressan t.
In Germany, on the other hand, the same cliche s are apparently
lacking, and the community is chiefly organized on the basis of a
complementary hierarchy in terms of dominance -submission. The
dominance behav ior is sharp ly and cl early developed ; yet the pic ture
is not perf ectly clear and needs further investigation. Whether a pure
dominance-s ubmission hierarchy could ever exist as a stable system
is doub tful. It seems that in the case o f Ger many, the sub mission e nd
of the pattern is masked, so that overt submissive behavior is almost
as strong ly tabooed as it is in America or England. In place of
submission, we find a sort of parad e-ground impassivity.
A hint as to the proces s by which the submissive role is modified
and render ed tolerable comes to us out of the inter-views in a
recently begun study of German life histories.18 One German subject
described how differen t was the treatment which he, as a boy,
received in his South German home, from that which his sister
received . He said that much more was demanded of him; that his
sister was allowed to evade discipline; that wherea s he was always
expec ted to click his heels and obey with precision , his sister was
allowed much more freedom. The interviewe r at once began to look
for intersex sibling jealou sy, but the subject declared that it was a
greater hono r for the boy to obey. "One doesn't expect too much of
girls," he said. "What one felt they (boys) should accomplish and do
was very serious , because they had to be prepa red for life." An
interes ting inversion of noblesse oblige.
17 io It is, however, possible that in certain sections of these nations ,
complem entar y patterns occur with some frequency—particularly among groups who
have suffered from prolon ged insecurit y and uncertaint y, e.g., racial minorities,
depressed areas, the stock exchange, political circles, etc.
18 G. Bateson, unpublished research for the Council on Human Relations .
108

Combina tions of Motifs
Among the complementary motifs, we have mention ed only
three—do minance-sub mission, exhibi tionism-spec tatorship, and
succoranc e-dependen ce—but these three will suffice to illustrate the
sort of verifiable hypothese s at which we can arrive by describing
nationa l character in this hyphenated t erminolog y.19
Since, clearly, all three of these motifs occur in all Western
cultures, the possibilities for interna tiona l differenc e are limited to
the propor tions and ways in which the motifs are combined. The
proport ions are likely to be very difficult to detect, excep t where the
differences are very large. We may be sure ourselv es that Germans
are more oriented toward dominance -submission than are
Americans, but to demonstrate this certainty is likely to be difficult.
To estimate differences in the degree of development of
exhibitionismspectatorsh ip or succo rance -dependence in the various
nations will, indeed , probabl y be qu ite impossib le.
If, howeve r, we consider the possible ways in which these motifs
may be combined togethe r, we find sharp qualitative differences
which are susceptible of easy verification. Let us assume that all
three of these m otifs are developed in al l relationsh ips in all Western
cultures, and from this assumption go on to conside r which
individual plays which role.
It is logically possible that in one cultural environment A will be
dominant and exhib itionis t, while B is submissive and spectator;
while in another culture X may be dominant and spectator, while Y
is submissive and exhibitionist.
Examples of this sort of contrast rather easily come to mind.
Thus we may note that whereas the dominant Nazis preen
themselves before the people, the czar of Russia kept his private
ballet, and Stalin emerges from seclusion only to review his troops.
We might perhaps present the relationship between the Nazi Party
and the people thus:
19 "For a fuller study, we ought to consider such other motifs as aggression-
passivity, possessive-possessed, agent-tool, etc. And all of these motifs will require
some what more critic al definition than can be att empted in this paper .
109

Party People
DominanceSubmission
ExhibitionismSpectatorship
While the czar and his ballet would be represented:
Czar Ballet
DominanceSubmission
SpectatorshipExhibitionism
Since these Europ ean examples are comparatively unproved, it is
worthwhil e at this point to demonstra te the occurrence of such
differenc es by descr ibing a rather striking ethnog raphic difference
which has be en docu mented m ore fully. In Europe , where we tend t o
assoc iate succoring behavior with socia l superio rity, we construct
our paren t symbols accordingl y. Our God, or our king, is the
"father" of his peop le. In Bali, on the other hand , the gods are the
"children" of the people , and when a god speaks through the mouth
of a person in trance , he addresses anyone who will listen as
"father." Similarly, the rajah is sajanganga ("spoi lt" like a child) by
his peopl e. The Bal inese , further, are ve ry fond of putting c hildren in
the combined roles of god and danc er; in mytholog y, the perfect
prince is polished and narcissistic. Thus the Balin ese pattern might
be summarized t hus:
High Status Low Status
Dependence Succoring
ExhibitionismSpectatorship
And this diagra m would imply, not only that the Balinese feel
dependenc e and exhib itionis m and super ior status to go naturally
together, but also that a Balinese will not readily combine succor ing
with exhibitionism (that is, Bali completely lacks the ostenta tious
gift-giving characteristic of many primitive peopl es) or will be
embarra ssed i f forced by the contex t to attempt such a combination .
110

Although the analogous diagra ms for our Western cultures
cannot be drawn with the same certainty, it is worthwhil e to attempt
them for the parent-child relationship s in Engl ish, American, and
German cultures. One extra complicat ion must, howeve r, be faced ;
when we look a t paren t-child relationsh ips ins tead of a t relationship s
between princes and people, we have to make specific allowance for
the changes i n the p attern which occu r as t he chi ld grows olde r. Suc-
corance -dependence is undoubt edly a dominant motif in early
childhood , but various mechanis ms later modify this extreme
dependence , to bring abou t some degree of psycholog ical in-
dependence .
The Engl ish uppe r- and middle-c lass system would be rep-
resent ed diagra mmatically thus:
ParentsChildren
DominanceSubmission
(modified b y "ternary" nurse system)
Succor ingDependence
(dependence habits broken by separation
—children sent to school)
Exhibi tionismSpectatorship
(children listen silently at meals)
In contrast with this, the analogous American pattern seems to
be:
Parents Children
Dominance (slight)Submission (slight)
Succoring Dependence
Spectatorship Exhibitionism
And this pattern differs from the English not only in the reversal
of the spectatorship -exhib itionis m roles, but also in the conten t of
what is exhibited. The American child is encourag ed by his parents
to show off his independen ce. Usua lly the process of psychologic al
wean ing is not accomplished by sending the child away to a
111

board ing schoo l; instead, the child's exhibitionis m is played off
again st his independ ence, until the latter is neutralized . Later, from
this beginn ing in the exhib ition of indep endence , the individu al may
sometimes go on in adult life to show off succor ance, his wife and
family beco ming in some degre e his "exh ibits."
Though the analogous German pattern probabl y resembles the
American in the arrange ment of the paired complementar y roles,
certainly it differs from the American in that the father's dominance
is much stronger and much more consistent, and especially in that
the content of the boy's exhibitionism is quite different. He is, in
fact, dominated into a sort of heel-clicking exhib itionis m which
takes the place of overt submissive behavior . Thus, while in the
American character exhibitionis m is encouraged by the parent as a
method of psychologica l wean ing, both its function and its content
are for the Ger man entirely different.
Differences of this order, which may be expected in all European
nations, are probably the basis of many of our naive and often
unkind internationa l comments. They may, indeed, be of
consid erable importance in the mechanic s of interna tiona l relations ,
in as much as an unde rstand ing of them might dispel some of our
misunders tanding s. To an American eye, the English too often
appea r "arrogan t," whereas to an English eye the American appears
to be "boastful." If we could show precisely how much of truth and
how much of distortion is presen t in these impress ions, it might be a
real con tribution to interallied coop eration.
In terms of the diagrams abov e, the "arrogance" of the
Engli shman would be due to the combinat ion of dominance and
exhib itionis m. The Engl ishman in a performing role (the parent at
break fast, the newspaper editor, the political spok esman, the
lecturer, or what not) assumes that he is also in a dominant role—
that he can decid e in accordance with vague , abstract standa rds wha t
sort of performance to give —and the audien ce can "take it or leave
it." His own arrogance he sees either as "natural" or as mitigated by
his humility in face of the abstract standards. Quite unaware that
his behavior could conceivably be regarded as a comment upon his
audience, he is, on the contrary, aware only of be-having in the
performer's role, as he understands that role. But the American
does not see it thus. To him, the "arrogant" behavior of the
112

Englishman appears to be directed against the audience, in which
case the implicit invocation of some abstract standard appears only
to add insult to injury.
Similarly, the behavior which an Englishman interprets as
"boastful" in an American is not aggressive, although the
Englishman may feel that he is being subjected to some sort of
invidious comparison. He does not know that, as a matter of fact,
Americans will only behave like this to people whom they rather
like and respect. According to the hypothesis above, the "boasting"
pattern results from the curious linkage whereby exhibition of self-
sufficiency and independence is played off against
overdependence. The American, when he boasts, is looking for
approval of his upstanding independence; but the naive
Englishman interprets this behavior as a bid for some sort of
dominance or superiority.
In this sort of way, we may suppose that the whole flavor of one
national culture may differ from that of another, and that such
differences may be considerable enough to lead to serious
misunderstandings. It is probable, however, that these differences
are not so complex in their nature as to be beyond the reach of
investigation. Hypotheses of the type which we have advanced
could be easily tested, and research on these lines is urgently
needed.
National Characte r and Ameri can Mo rale
Using the motifs of interpe rsona l and intergroup relation-ship as
our clues to national charac ter, we have been able to indicate certain
orders of regular difference which we may expec t to find among the
peop les who share our Western civilization. Of necess ity, our
statements have been t heoretical rather t han e mpirical; still, from the
theore tical structure which we have built up, it is possible to extract
certain formulas whi ch m ay be usefu l to the bu ilder of morale.
All of these formulas are based upon the general assumption
that people will respond most energetically when the context is
structured to appeal to their habitual patterns of reaction. It is not
113

sensible to encourage a donkey to go up hill by offering him raw
meat, nor will a lion respond to grass.
(1) Sinnce all Western nations tend to think and behave in
bipolar terms, we shall do well, in building American morale, to
think of our various enemies as a single hostile entity. The
distinctions and gradations which intellectuals might prefer are
likely to be disturbing.
(2) Since both Americans and English respond most ener-
getically to symmetrical stimuli, we shall be very unwise if we
soft-pedal the disasters of war. If our enemies defeat us at any
point, that fact ought to be used to the maximum as a challenge
and a spur to further effort. When our forces have suffered some
reverse, our newspapers ought to be in no hurry to tell us that
"enemy advances have been checked." Military progress is always
intermittent, and the moment to strike, the moment when
maximum morale is needed, occurs when the enemy is solidifying
his position and preparing the next blow. At such a moment, it is
not sensible to reduce the aggressive energy of our leaders and
people by smug re-assurance.
(3) There is, however, a superficial discrepancy between the
habit of symmetrical motivation and the need for showing self-
sufficiency. We have suggested that the American boy learns to
stand upon his own feet through those occasions in childhood
when his parents are approving spectators of his self-sufficiency. If
this diagnosis is correct, it would follow that a certain bubbling up
of self-appreciation is normal and healthy in Americans and is
perhaps an essential ingredient of American independence and
strength.
A too literal following of the formula above , therefore , a too
great insistence upon disasters and difficulties, might lead to some
loss of energy through the damming up of this spontaneous
exuber ance. A rather conc entra ted diet of "blood , swea t, and tears"
may be good for the English ; but Americans, while no less
dependen t upon symmetrica l motivat ion, cannot feel their oats when
fed on nothing but disaster. Our public spok esmen and newspaper
editors should never softpeda l the fact that we have a man-sized job
on our hands , but they will do well to insist also that America is a
man-siz ed nation. Any sort of attempt to reassure Americans by
114

minimizing the strength of the enemy must be avoid ed, but frank
boasts of real success a re good.
(4) Because our vision of the peace is a factor in our war-making
morale, it is worthwhil e to ask at once what light the study of
nationa l differen ces may throw upon the problems of the peace
table.
We have to devise a peace treaty (a) such that Americans and
British will fight to achieve it, and (b) such that it will bring out the
best rather than the worst characteristics of our enemies. If we
approach it sc ientifically, such a problem is by no means beyond our
skill.
The most conspicuous psychologica l hurdle to be nego tiated, in
imagining such a peace treaty, is the contrast between British and
American symmetrical patterns and the German complementary
pattern, with its taboo on overt sub-missive behav ior. The allied
nations are not psychologica lly equipped to enfor ce a harsh treaty;
they might draw up such a treaty, but in six months they would tire
of keeping the underdog down. The Germans, on the other hand , if
they see their role as "sub missive," will not stay down withou t harsh
treatment. We have seen that these consider ation s applied even to
such a mildly punitive treaty as was devised at Versailles; the allies
omitted to enforc e it, and the Germans refused to accept it. It is,
therefore, usele ss to dream of such a treaty, and wors e than useless
to repea t such dreams as a way of raising our morale now, when we
are angr y with Ger many. To do t hat would on ly obscure t he issues i n
the final settlement.
This incompatib ility between comple mentar y and symmetrical
motivat ion means, in fa ct, th at the tre aty cannot be organiz ed around
simple dominance -submissive motifs; hence we are forced to look
for alternative solutions. We must ex-amine, for example, the motif
of exhibitionis m-spectatorsh ip —what dignified role is each of the
various nations best fitted to play?—and that of succor ing-
dependence—in the starving postwar world, what motivat ional
patterns shall we evok e between those who give and those who
receiv e food? And, alterna tive to these solutions, we have the
possibility of some threefo ld structu re, within which both the allies
and Germany would submit, not to each other, but to some abstract
principle.
115

Bali: The Value Sys tem of a Steady
State*
"Ethos" and "Sch ismog enesis"
It would be an overs implification—i t would even be false —to
say that science necess arily advan ces by the construction and
empirical testing of succes sive working hypotheses . Among the
physicists and chemists there may be some who really proceed in
this oithoclox manner, but among the social scientists there is
perhap s not one. Our concept s are loosel y defined—a haze of
chiaroscuro prefigur ing sharpe r lines still undrawn—and our
hypothese s are still so vagu e that rarely can we imagine any crucia l
instance whose i nvest igation wi ll test them.
The present pape r is an attempt to make more precis e an idea
which I published in 193620 and which has lain fallow since that
time. The not ion of ethos had p roved a use ful conceptua l tool f or me,
and with it I had been able to get a sharper unders tanding of Iatmul
culture. But this experience by no means proved that this tool would
neces sarily be usefu l in other hand s or for the analysis of other
cultures . The most gene ral conclusion I could draw was of this
order: that my own mental processes had certain characteristics;
that the sayings, actions, and organization of the Iatmul had certain
characteristics; and that the abstraction, "ethos," performed some
role—catalytic, perhaps—in easing the relation between these two
specificities, my mind and the data which I myself had collected.
Immediate ly after completing the manuscrip t of Naven , I went to
Bali with the intention of trying upon Balinese data this tool which
had been evolved for the analysis of Iatmul. For one reason or
anothe r, however , I did not do this, partly becaus e in Bali Margaret
* This essay appeared in Social Structure: Studies Presented to A. R. Radcliffe-
Brown, edited by Meyer Fortes, 1949. It is reprinted by permission of the Clarendon
Press. Preparation of the essay was aided by a Guggenheim Fello wship.20 G. Bateson , Naven , Cambrid ge, Cambrid ge University Press, 1936 .
116

Mead and I were engaged in devising other tools—photog raphi c
methods of record and description—and partly because I was
learning the techniqu es of applying genetic psycholog y to cultural
data, but more especially becaus e at some inarticulate level I felt
that the tool was unsui table for this new task.
It was not that ethos was in any sense disproved— indeed, a tool
or a method can scarcely be proved false. It can only be shown to be
not usefu l, and in this case there was not even a clear demonstra tion
of uselessne ss. The method remained almost untried, and the most I
could say was that, after that surrender to the data which is the first
step in all anthropo logica l study, ethologi cal analysis did not seem
to be the next thing to do.
It is now possible to show with Balinese data what peculiarities
of that culture may have influenced me away from ethologic al
analysis, and this demonstra tion will lead to a greater gene ralization
of the abstraction; ethos. We shall in the process make certain
heuristic advances which may guide us to more rigorous descrip tive
procedure s in dealing with other cultures.
(1) The analysis of Iatmul data led to the definition of ethos as
"The expression of a culturally standa rdized system of organization
of the instincts and e motions of the individuals."21
(2) Anal ysis of Iatmul ethos—con sisting in the ordering of data
so as to make evident certain recu rrent "emphases" or "the mes"—led
to recogni tion of schismogenes is. It appeared that the working of
latmul society involved inter alia two classes of regenerative22 or
"vicious" circles. Both of these were sequences of social interaction
21 Naven, p. 118.22 The terms "regenerati ve" and "degenerati ve" are borrowed from communi-
cations engineering. A regenerative or "vicious " circle is a chain of variables of the
general type: increase in A causes increase in B; increase in B causes increase in C;
.. increase in N causes incre ase in A. Such a system, if provided with the necessar y
energy sources and if external factors permit, will clearly operate at a greater and
greater rate or intensit y. A "degenerative" or "self-corrective" circle differs from a re-
generati ve circle in containing at least one link of the type: "increase in N causes
decrease in M." The house thermostat or the steam engine with a governor are ex-
amples of such self-corre cting systems. It will be noted that in many instances the
same material circuit may be either regenerative or degenerati ve accordin g to the
amount of loading, frequenc y of impulses trans mitted around the path, and time
chara cteristics of the total path.
117

such that A's acts were stimuli for B's acts, which in turn became
stimuli for more intense action on the part of A, and so on, A and B
being persons acting either as individuals or as group members.
(3) These schismogenic sequences could be classified into two
classes: (a) symmetrica l schismogenesis , where the mutually
promoting actions of A and B were essent ially similar, e.g., in cases
of competition, rivalry, and the like; and ( b) complementar y
schismogenesis , where the mutually promoting actions are
essen tially dissimilar but mutually appropriate, e.g., in cases of
dominance-s ubmission , succo ring-d ependence , exhib itionis m-
spectatorship , and the like.
(4) In 1939 a considerab le advanc e was made in defining the
formal relations between the concept s of symmetrical and
complementary schismogenesis . This came from an attempt to state
schismogenic theory in terms of Richa rdson's equations for
interna tiona l armaments races23. The equation s for rivalry evident ly
gave a first approximation to what I had called "symmetrical
schismogenesis ." These equa tions assume that the intens ity of A's
actions (the rate of his arming, in Richard son's case) is simply
propor tiona l to the amount by which B is ahead of A. The stimulus
term in fact is (B —A) , and when this term is positive it is expect ed
that A will en-gag e in efforts to arm. Richa rdson's second equa tion
makes the same assumption mutatis mutandis abou t B's actions .
These equat ions s uggested that other simply rivalrou s or competitive
pheno mena—e.g ., boasting—though not subjec t to such simple
measure ment as expendi ture on armament, might yet when
ultimately measured be reduc ible to a simply analogous set of
relations.
The matter was, however , not so clear in the case of com-
plementary schismogenesis . Richard son's equations for "sub-
mission" eviden tly define a pheno menon somewhat differen t from a
progr essive complementar y relationsh ip, and the form of his
equat ions descr ibes the action of a factor "submissiveness" which
slows down and ultimately reverses the sign of warlike effort. What
was, however , required to describe complementar y schismogenesis
23 L. F. Richardson, "Generalized Foreign Politics ," British Journal of
Psycholog y, Monograph Supple ment x xiii, 1939 .
118

was an equ ation al form giving. a sharp and d iscon tinuous r evers al of
sign. Such an equa tiona l form is achieved by suppos ing A's actions
in a complementary relation ship to be propor tiona l to a stimulus
term of the type (A —B) . Such a form has also the advantage of
automatically defining the actions of one of the participants as
negative, and thus gives some mathematical analogue for the
apparent psychologica l relatedness of domination to submission,
exhibitionism to spect atorship, succo ring to dependence , etc.
Notab ly this formulation is itself a negat ive of the formulation
for rivalry, the stimulus term being the opposi te. It had been
observed that symmetrical sequence s of actions tend sharp ly to
reduce the strain of excess ively complementar y relationsh ips
between persons or groups.24 It is tempting to ascribe this effect to
some hypothes is which would make the two types of
schismogenes is in some degree psycholog ically incompatible , as is
done by the abov e formulation.
(5) It is of interest to note that all the modes associated with the
erogenous zone s,25 though no t clea rly quan tifiable , define th emes for
complementary relationship .
(6) The link with erogenous zones suggest ed in 5, above,
indica tes that we ought , perhap s, not to think of simple rising
expon ential curves of intensity limited only by factors analogous to
fatigu e, such as Richardson' s equa tions would imply; but rather that
we shou ld expect our curve s to be bounded by pheno mena
comparable to orgasm—that the achieve ment of a certain degree of
bodily or neural involve ment or intensity may be followed by a
release of schismogenic tension. Indeed, all that we know about
human being s in various sorts of simple contests would seem to
24 Naven, p. 173.25 E. H. Hombur ger, "Configurations in Play: Psycho-logical Notes ,"
Psychoanal ytical Quart erly, 1937 , vi: 138-214. This paper , one of the most
important in the literatur e seeking to state psychoanal ytic hypotheses in more
rigorous terms, deals with the "modes" appropriate to the various erogenous zones
—intrusion , incorporation, retention, and the like—and shows how these modes may
be trans ferred from one zone to another . This leads the writer to a chart of the
possible permutations and combinations of such transferred modalities . This chart
provides precis e means of describin g the course of the development of a large
variety of different types of character structure (e.g., as met with in different
cultures) .
119

indicate that this is the case, and that the consciou s or unconsc ious
wish for release of this kind is an importan t factor which draws the
participants on and prevents them from simply withd rawing from
conte sts which would otherwise not commend themselve s to "common sense." If there be any basic human chara cteristic which
makes man prone to strugg le, it would seem to be this hope of
release from tension through total involve ment. In the case of war
this factor is undoubted ly often potent. (The real truth— that in
modern warfa re only a very few of the participan ts achieve this
climactic releas e—see ms hardly to stand again st the insidious myth
of "total" war.)
(7) In 1936 it was suggest ed that the pheno menon of "falling in
love" might be comparable to a schismogenes is with the signs
reversed, and even that "if the course of true love ever ran smooth it
would follow an exponen tial curve."26 Rich ardson27 has since,
independ ently, made the same point in more formal terms.
Parag raph 6, abov e, clearly indicates that the "exponent ial curves"
must give place to some type of curve which will not rise
indefinitely but will reach a climax and then fall. For the rest,
however , the obvious relation-sh ip of these interactive pheno mena
to climax and orgasm very much streng thens the case for regarding
schismogenesis and those cumulative sequences of interaction
which lead to love as often psycholog ically equivalent. (Witness the
curious confus ions between fighting and lovemaking, the symbolic
identifications of orgasm with death, the recurrent use by mammals
of organs of offense as ornaments of sexua l attracttion, etc.)
(8) Schismogenic sequences were not found in Bali. This
negat ive statement is of such importan ce and conflicts with so many
theories of social oppos ition and Marx ian determinism that, in order
to achieve credib ility, I must here de-sc ribe schematica lly the
proce ss of character formation , the resulting Balinese chara cter
structu re, the excep tiona l in-stances in which some sort of
cumulative interaction can be recognized , and the methods by which
quarrels and status differen tiation are handl ed. (Detailed analysis of
the various points and the suppor ting data cannot here be
26 Naven , p. 197.27 Op. cit., 1939.
120

reproduced , but references will be given to published sources where
the da ta can b e exa mined.)28
Balinese Char acter
(a) The most importan t except ion to the above generalization
occurs in the relationsh ip between adults (espec ially parents) and
children. Typically, the mother will start a small flirtation with the
child, pulling its penis or otherwi se stimulating it to interperson al
activity. This will excite the child, and for a few moments
cumulative interaction will occur. Then just as the child,
approaching some small climax, flings its arms around the mother's
neck, her attention wanders . At this point the child will typical ly
start an alternative cumulative interac tion, building up towa rd
temper tantrum. The mother will either play a spectator's role,
enjoying the child's tantrum, or, if the child actually attacks her, will
brush off his attack with no show of anger on her part. These se-
quen ces can be seen either as an express ion of the mother's distaste
for this type of persona l involve ment or as contex t in which the
child acqui res a deep distrust of such involvement. The perhaps
basically human tenden cy toward s cumulative personal interaction
is thus muted.29 It is possible that some sort of continuing plateau
of intensity is substituted for climax as the child becomes more
fully adjusted to Balinese life. This cannot at present be clearly
documented for sexual relations, but there are indications that a
plateau type of sequence is characteristic for trance and for
quarrels (see d, below).
(b)Similar sequences have the effect of diminishing the child's
tendencies toward competitive and rivalrous behavior. The mother
will, for example, tease the child by suckling the baby of some
28 See especially G. Bateson and M. Mead , Balinese Character: A Photographic
Anal ysis. Since this photo- graphic record is available, no photo graphs are included
in the present paper .29 Balinese Ch aracter: A Photographic Anal ysis, pl . 47, and pp. 32-6.
121

other woman and will enjoy her own child's efforts to push the
intruder from the breast.30
(c)In general the lack of climax is characteristic for Balinese
music, drama, and other art forms. The music typically has a
progression, derived from the logic of its formal structure, and
modifications of intensity determined by the duration and
progress of the working out of these formal relations. It does not
have the sort of rising intensity and climax structure characteristic
of modern Occidental music, but rather a formal progression.31
(d) Balines e culture includes definite techniqu es for dealing with
quarrels. Two men who have quarrelled will go formally to the
office of the local represen tative of the Rajah and will there register
their quarrel, agreeing that which ever speaks to the other shall pay a
fine or make an offering to the gods. Later, if the quarr el terminates ,
this contract may be formally nullified. Smaller—bu t similar—
avoidan ces (pwik) are practiced, even by small children in their
quarrels. It is significant, perhaps, that this procedure is not an
attempt to influen ce the protagon ists away from hostility and toward
friendship . Rathe r, it is a formal recognit ion of the state of their
mutual relationsh ip, and possibly, in some sort, a pegging of the
relationship at that state. If this interpret ation is correct, this method
of dealing with quarrels would correspond to the substitution of a
plateau for a climax.
(e) In regard to warfare, contemporar y comment on the old wars
between the Rajahs indica tes that in the period when the comments
were collected (1936–39 ) war was thought of as containing large
elements of mutual avoidanc e. The village of Bajoeng Gede was
surrounded by an old vallum and foss, and the people explained the
functions of these fortifications in the following terms: "If you and I
had a quarrel, then you would go and dig a ditch around your hous e.
Later I would come to fight with you, but I would find the ditch and
then there would be no fight"—a sort of mutual Maginot Line
psycholog y. Similarl y the boundar ies between neighbor ing
kingdo ms were, in genera l, a deserted no-man's land inhabi ted only
by vagrants and exiles. (A very different psycholog y of warfare was
30 lbid., pls. 49, 52, 53, and 69-72.31 See Colin McP hee, "The Absolute Music of Bali," Modern Music, 1935; and
A House in Bali, London, Gollancz, 1947.
122

no doub t develop ed when the kingdo m of Karanga sem embarked on
the conques t of the neighbor ing island of Lombok in the beginn ing
of the eighteen th century. The psycholog y of this militarism has not
been investigated , but there is reason to believe that the time
perspec tive of the Balinese colonis ts in Lombok is today sig-
nificantly differen t from that of Bal inese in Bali.)32
(f) The formal techniques of socia l influence—o ratory and the
like—a re almost totally lacking in Balinese culture. To demand the
continued attention of an individua l or to exert emotiona l influence
upon a group are alike distasteful and virtually impossibl e; because
in such circumstances the attention of the victim rapidly wande rs.
Even such continued speech as would , in most cultures , be used for
the telling of stories does not occur in Bali. The narrator will,
typically, pause after a senten ce or two, and wait for some member
of the audience to ask him a concre te question abou t some detail of
the plot. He will then answer the question and so resume his
narration. This procedure appa rently breaks the cumulative tension
by irrelev ant interaction.
(g) The principa l hierarch ical structures in the society—the caste
system and the hierarchy of full citizens who are the village coun cil
—are rigid. There are no contex ts in which one individua l could
conceivab ly compete with anoth er for position in either of these
systems. An individua l may lose h is membership in the hierarchy for
various acts, but his place in it cannot be altered. Should he later
return to orthodoxy and be accept ed back, he will return to his
origin al po sition in relation to the other members.33
The forego ing descriptive gene ralizations are all partial answers
to a nega tive quest ion—"Why is Balinese society non-
schismogenic ?"—and from the combination of these gener alization s
we arrive at a picture of a socie ty differing very markedl y from our
own, from that of the Iatmul, from those systems of social
oppos ition which Radcl iffe-Brown has analyzed, and from any
social structure postu lated by Marxian analysis.
32 See G. Bateson, "An Old Temple and a New Myth," Djawa, xvii, Batavia, 1937.33 See M. Mead, "Public Opinion Mechanis ms among Primitive Peoples,"
Public Opinion Quarterly, 1937, is 5-16.
123

We started with the hypothes is that human being s have a
tenden cy to involve themselves in sequences of cumulative
interac tion, and this hypothesi s is still left virtually intact. Among
the Balin ese the babies , at least, evid ently have such tendenc ies. But
for sociolog ical validity this hypothesis must now be guarded with a
paren thetical claus e stipula ting that these tendencie s are operative in
the dynamics of society only if the childhood training is not such as
to preven t their expression in adult life.
We have made an advance in our knowledge of the scope of
human characte r formation in demonstr ating that these tendencies
toward cumulative interaction are subjec t to some sort of
modification , decond itioning, or inhibition.34 And this is an
important advanc e. We know how it is that the Balines e are
nonschi smogenic and we know how their distaste for schismogenic
patterns is expressed in variou s details of the social organiza tion—
the rigid h ierarchi es, the institutions fo r the handling of quarr els, etc.
—but we still know nothing of the positive dynamics of the society.
We have an swered on ly the nega tive question.
Balinese Ethos
The next step, therefore, is to ask about Balinese ethos. What
actually are the motives and the values which accompany the
complex and rich cultural activities of the Balinese? What, if not
competitive and other types of cumulative interrelationship, causes
the Balinese to carry out the elaborate patterns of their lives?
(1) It is immediate ly clear to any visitor to Bali that the driving
force for cultural activity is not either acqui sitivenes s or crude
physiologica l need. The Balinese , especia lly in the plains, are not
hungr y or pove rty-stricken. They are wasteful of food, and a very
consid erable part of their activity goes into entirely nonproduc tive
activities of an artistic or ritual nature in which food and wealth are
34 As is usual in anthropology, the data are not sufficiently precise to give us any clue
as to the natur e of the learning processes involved. Anthropology, at best, is only
able to raise problems of this order. The next step must be left for laborator y
experim entation.
124

lavish ly expended . Essentially, we are dealing with an econo my of
plenty rather than an econo my of scarcity. Some, indeed , are rated
"poo r" by their fellows, but none of these poor are threatened by
starva tion, and th e suggest ion tha t human beings m ay actually starve
in great Occiden tal cities was, to the Balinese , unutterably shocking .
(2) In their econo mic transac tions the Balinese show a great
deal of carefu lness in their small dealings. They are "penn y wise."
On the other hand , this carefuln ess is counte r-acted by occasiona l
"pound foolishnes s" when they will expend large sums of money
upon ceremonials and other forms of lavish consu mption . There are
very few Baline se who have the idea of steadily maximizing their
wealth or proper ty; these few are partly disliked and partly regard ed
as oddities. For the vast majority the "saving of penn ies" is done
with a limited time perspe ctive and a limited level of aspiration.
They are saving until they have enough to spend largely on some
ceremonial. We shou ld not describ e Balinese econo mics in terms of
the individual's attempt to maximize value, but rather compare it
with the relaxa tion oscillation s of physiology and engineer ing,
realizing that not only is this analog y descriptive of their sequen ces
of transac tions , but that they themselves see these sequenc es as
naturally having some such form.
(3) The Balinese are markedl y dependen t upon spatial
orientation. In order to be able to behave they must know their
cardina l points, and if a Balinese is taken by motor car over twisting
roads so that he loses his sense of direction, he may beco me
severe ly disorien tated and unabl e to act (e.g., a dancer may beco me
unab le to dance) until he has got back his orienta tion by seeing s ome
important landmark, such as the central mountain of the island
around which the cardinal points are structu red. There is a
comparable dependence upon social orientation, but with this
difference : that where the spatial orientation is in a horizont al plane ,
social orientation is felt to be, in the main, vertical. When two
strange rs are brought togeth er, it is necessa ry, before they can
conv erse with any freedo m, that their relative caste positions be
stated. One will ask the other, "Where do you sit?" and this is a
metaphor for caste. It is asking , essen tially, "Do you sit high or
low?" When each knows the caste of the other, each will then know
125

what etiquette and what linguistic forms he shou ld adop t, and
conver sation can then proceed. Lacking such orientation, a Baline se
is tongue -tied.
(4) It is common to find that activity (other than the "penn y
wisdo m" mention ed above ) rather than being purposive , i.e., aimed
at some deferred goal, is valued for itself. The artist, the dancer , the
musician, and the priest may receive a pecun iary reward for their
profe ssiona l activity, but only in rare cases is this reward a dequate to
recompense the artist even for his time and materia ls. The reward is
a token of appreciation , it is a defin ition of the context in which the
theatrical company performs, but it is not the econo mic main-stay of
the troupe. The earnings of the troupe may be saved up to enable
them to buy new costumes, but when finally the costu mes are
bought it is usual ly necess ary for every member to make a
consid erable contribut ion to the common fund in order to pay for
them. Similarl y, in regard to the offerings which are taken to every
temple feast, there is no purpose in this enormous expendi ture of
artistic work and real wealth. The god will not bring any benef it
becaus e you made a beautiful structu re of flowers and fruit for the
calendric feast in his temple, nor will he avenge your abstention.
Instead of deferred purpo se there is an immediate and immanent
satisfaction in performing beau tifully, with everybody else, that
which i t is correct to perform in each particular contex t.
(5) In gener al there is eviden t enjoyment to be had from doing
things busily with large crowds of other peop le.35 Conversely there
is such misfortune inheren t in the loss of group membership that the
threat of this loss is one of the most seriou s san ction s in the culture.
(6) It is of great interest to note that many Baline se actions are
articulately accoun ted for in socio logic al terms rather than in terms
of individual goals or values.36
This is most conspicuous in regard to all actions related to the
village counc il, the hierarchy which includes all full citizens . This
body, in its secular aspects, is referred to as I Desa (literally, "Mr.
Village "), and numerous rules and procedures are rationa lized by
reference to this abstract personage. Similar ly, in its sacred aspec ts,
35 Bateson and M ead, op. cit., p1. 5.36 Cf. Naven, pp. 250 if., where it was suggested that we must expect to find that some
peoples of the world would relate their actions to the sociological fra me.
126

the village is deified as Betara Desa (God Village) , to whom shrines
are erected and offerings brough t. (We may guess that a
Durkhe imian analysis would seem to the Balinese to be an obviou s
and appropr iate approach to the unde rstand ing of much of their
public culture.)
In particular all money transactions which involve the village
treasu ry are gove rned by the genera lization, "The village does not
lose" (Desanne sing dadi potjol). This genera lization applies, for
example, in all cases in which a beast is sold from the village herd.
Under no circumstances can the village accep t a price less than that
which it actually or nominally paid. (It is important to note that the
rule take s the form of fixing a lower limit and is not an injun ction to
maximize the village treasu ry.)
A pecu liar awareness of the nature of social processes is evident
in such in ciden ts as the following : A poor man was about to under go
one of the important and expens ive rites de passage which are
necessa ry for persons as they approach the top of the counc il
hierarchy. We asked what would hap-pen if he refused to unde rtake
this expenditu re. The first answer was that, if he were too poor, I
Desa would lend him the money. In respon se to further press ing as
to what would happ en if he really refused, we were told that nobod y
ever had refused , but that if somebod y did, nobod y would go
through the ceremony again . Implicit in this answer and in the fact
that nobod y ever does refuse is the assumption that the ongoing
cultural proces s is itself to be valued.
(7) Actions which are culturally correct (patoet) are acceptable
and aesthetically valued. Actions which are permissible (dadi) are
of more or less neutral value; while actions which are not
permissible (sing dadi) are to be depreca ted and avoided. These
generalizations , in their translated form, are no doub t true in many
cultures, but it is important to get a clear under stand ing of what the
Balinese mean by dadi. The notion is not to be equa ted with our
"etique tte" or "law," since each of these invokes the value judgment
of some other person or sociolog ical entity. In Bali there is no
feeling that actions h ave be en or are categor ized as dadi or sing dadi
by some human or supernatur al authori ty. Rather, the statement that
such-and-s uch an action is dadi is an absolu te generalization to the
127

effect that unde r the given circumstances this action is regu lar.37 It is
wrong for a casteless person to address a prince in other than the
"polished language ," and it is wrong for a menstrua ting woman to
enter a temple. The prince or the deity may express annoyance, but
there is no feeling that either the prince, the deity, or the casteless
per-son made the rules. The offense is felt to be against the order
and natural structure of the universe rather than agains t the actual
person o ffended . The offender, even i n such se rious m atters a s ince st
(for which he may be extruded from the society)38 is not blamed for
anything worse than stupid ity and clumsiness . Rathe r, he is "an
unfor tunate person" (anak latfoer), and misfortune may come to
any of us "when it is our turn." Further, it m ust be stressed that thes e
patterns which defin e correct and permissibl e behav ior are
exceed ingly complex (espec ially the rules of language ) and that the
individua l Baline se (even to some degree inside his own family) has
continual anxiety lest he make an error. Moreove r, the rules are not
of such a kind that they can be summarized either in a simple recipe
or an emotional attitude . Etiqu ette cannot be deduced from some
comprehens ive statement abou t the other person's feelings or from
respect for super iors. The details are too complex and too variou s
for this, and so the individu al Baline se is foreve r picking his way,
like a tightrop e walker , afraid at any moment lest he make some
misstep.
(8) The metaphor from postural balance used in the last
parag raph is demonstrab ly applicab le in many contexts of Balinese
culture:
(a)The fear of loss of suppor t is an importan t theme in
Balin ese c hildhood .39
(b)Elevation (with its attendant problems of physical and
metaphor ical balance) is the pas sive complement o f respect.40
(c)The Balinese child is elevated like a super ior per-son or a
god.41
37 The word dadi is also used as a copula referring to changes in social status . I
Anoe dadi Koebajan means "So-and-so has become a villa ge of ficial. "38 Mead, "Public Opinion Mechanisms among Primitive Peoples ," loc. cit., 1937.39 Bateson and M ead, op. cit., pls. 17, 67, and 79.40 Ibid., pls. 10-14.41 Ibid., p1. 45.
128

(d)In cases of actual physical elevat ion42 the duty of
balancing the system falls on the support ing lower person, but
control of the direction in which the system will move is in the
hand s of the elevated . The little girl in the figure standing in
trance on a man's shoulde rs can cause her bearer to go wherever
she desires by merely leaning in that direction. He must then
move in that direction in order to maintain the balance of the
system.
(e)A large proport ion of our collection of 1200 Balinese
carvings shows preoccupa tion on the part of the artist with
proble ms of balance .43
(f)The Witch, the personif ication of fear, frequent ly uses a
gesture called kapar, which is described as that of a man falling
from a coconut palm on suddenly seeing a snake. In this
gesture the arms are raised sideways to a position some-what
above the head.
(g)The ordin ary Balin ese term for the period before the
coming of the white man is "when the world was steady"
(doegas go emine en teg).
Applica tions of the Von N eumannian Game
Even this very brief listing of some of the elements in Balinese
ethos suffices to indicate theoretical proble ms of prime importan ce.
Let us consider the matter in abstract terms. One of the hypotheses
unde rlying most sociology is that the dynamics of the social
mechanis m can be described by assuming that the individua ls
constituting that mechanis m are motivated to maximize certain
variables. In conven tiona l econo mic theory it is assumed that the
individu als will maximize value, while in schismogenic theory it
was tacitly assumed that the individu als would maximize intangib le
but still simple variables such as prestige, self-esteem, or even
42 Ibid., p1. 10, fig. 3.43 At present it is not possible to make such a statement in sharpl y defined
quantitative terms, the available judgments being subjective and Occidental.
129

submissiveness . The Balin ese, howeve r, do not maximize any such
simple variables.
In order to define the sort of contrast which exists between the
Balinese system and any competitive system, let us start by
considering the premisses of a strictly competitive Von
Neumannian game and proceed by considering what changes we
must make in these premisses in order to approximate more closely
to the Balinese system.
(1) The players in a Von Neumannian game are, by hypothesis,
motivated only in terms of a single linear (sc. monetary) scale of
value. Their strategies are determined: (a) by the rules of the
hypothetical game; and (b) by their intelligence, which is, by
hypothesis, sufficient to solve all problems presented by the game.
Von Neumann shows that, under certain definable circumstances
depending upon the number of players and upon the rules,
coalitions of various sorts will be formed by the players, and in
fact Von Neumann's analysis concentrates mainly upon the
structure of these coalitions and the distribution of value among
the members. In comparing these games with human societies we
shall regard social organizations as analogous to coalition systems.44
(2) Von Neumannian systems differ from human societies in the
following respects:
(a)His "players" are from the start completely intelligent, whereas
human beings learn. For human beings we must expect that the rules
of the game and the conventions associated with any particular set of
coalitions will become incorporated into the character structures of
the individual players.
44 Alternatively, we might handle the analogy in another way. A social system is, as
Von Neumann and Morgenstern point out, comparable to a non-zero sum game in
which one or more coalitions of people play against each other and against nature. The
non-zero sum characteristic is based on the fact that value is continually extracted from
the natural environment. Inasmuch as Balinese society exploits nature, the total entity,
including both environment and people, is clearly comparable to a game requiring
coalition between people. It is possible, however, that that subdivision of the total
game comprising the people only might be such that the formation of coalitions
within it wou ld not be essential—that is, Balinese society may differ from most other
societies in that the "rules" of the relationship between people de-fine a "game" of the
type Von Neumann would call "non-essential." This possibility is not here examined.
(See Von Neumann and Morgenstern, op. cit.)
130

(b)The mammalian value scale is not simple and mono-tone, but
may be exceedingly complex. We know, even at a physiological
level, that calcium will not replace vitamins, nor will an amino acid
replace oxygen. Further, we know that the animal does not strive to
maximize its supply of any of these discrepant commodities, but
rather is required to maintain the supply of each within tolerable
limits. Too much may be as harmful as too little. It is also doubtful
whether mam malian preference is always transitive.
(c) In the Von Neumannian system the number of moves in a
given "play" of a game is assumed to be finite. The strategic
proble ms of the individu als are solub le because the individua l can
operate within a limited time persp ective. He need only look
forward a finite distance to the end of the play when the gains and
losses wil l be paid up and every-thing will sta rt again fro m a tabula
rasa. In human society life is not punctuated in this way, and each
individual faces a vista of unknowable factors whose number
increases (probably exponentially) into the future.
(d) The Von Neumannian players are, by hypothesis, not
susceptible either to economic death or to boredom. The losers can
go on losing forever, and no player can withdraw from the game,
even though the outcome of every play is definitely predictable in
probability terms.
(3) Of these differences between Von Neumannian and human
systems, only the difference s in value scales and the possib ility of
"death" conc ern us here. For the sake of simplicity we shall assume
that the othe r differences , though very profound, can for the m oment
be ignored.
(4) Curious ly, we may note that, although men are mammals and
therefore have a primary value system which is multidimensiona l
and nonmaxi mizing, it is yet possible for these creatu res to be put
into contexts in which they will strive to maximize one or a few
simple v ariab les (money, prestige, powe r, etc.) .
(5) Since the multidi mension al value system is apparen tly
primary, the problem presen ted by, for example, Iatmul social
organiza tion is not so much to account for the behavio r of Iatmul
individu als by invok ing (or abstracting) their value system; we
shou ld also ask how t hat value s ystem is imposed on the m ammalian
individu als b y the social or ganization in which the y find th emselves .
131

Convention ally in anthropolog y this question is attacked through
genet ic psycholog y. We endeavo r to collect data to show how the
value system implicit in the social prganization is built into the
character structure of the individuals in their childhood. There is,
however , an alternative approa ch which would momentarily ignore ,
as Von Neumann does, the pheno mena of learning and conside r
merely the st rateg ic implic ation s of thos e contexts which m ust occu r
in ac cordance wi th the g iven "ru les" and t he coa lition s ystem. In this
connec tion it is importan t to note that competitive contex ts—
provid ed the individu als can be made to recogn ize the contex ts as
competitive— inevi tably reduc e the complex gamut of values to ver y
simple and even linear and monotone te rms.45 Considerations of this
sort, plus descriptions of the regularities in the process of character
formation, probably suffice to de-scribe how simple value scales
are imposed upon mammalian individuals in competitive societies
such as that of the Iatmul or twentieth-century America.
(6) In Balin ese society, on the other hand , we find an. entirely
differen t state of affairs. Neithe r the individu al nor the village is
conce rned to maximize any simple variable . Rather, they would
seem to be conce rned to maximize some-thing which we may call
stability, using this term perhaps in a highly metaphor ical way.
(There is, in fact, one simple quantitative variable which does
appea r to be maximized. This variable is the amount of any fine
imposed by the village . When first imposed the fines are mostly
very small, but if payment is delayed the amount of the fine is
increased very steeply, and if there be any sign that the offender is
refusing to pay—"opposing the village "—the fine is at once raised
to an enormous sum and the offender is deprived of membership in
the community until he is willing to give up his oppos ition. Then a
part of the fine m ay be excus ed.)
(7) Let us now conside r an hypothetical system consisting of a
number of identical players, plus an umpire who is concerned with
the mainten ance of stability among the players. Let us further
suppose that the players are liable to econo mic death, that our
umpire is conce rned to see that this shall not occu r, and that the
umpire has powe r to make certain alterations in the rules of the
45 L. K. Frank, "The Cost of Com petition," Plan Age, 1940, vi : 314-24.
132

game or in the probabi lities assoc iated with chan ce moves. Clearly
this umpire will be in more or less continua l conflict with the
players. He is striving to maintain a dynamic equilibrium or steady
state, and this we may rephra se as the attempt to maximize the
chances against the m aximization of any single simple va riable.
(8.) Ashb y has pointed out in rigorous terms that the steady state
and continu ed existence of complex interac tive systems depend
upon prevent ing the maximization of any variable , and that any
continued increase in any variab le will inevitably result in, and be
limited by, irreve rsible change s in the system. He has also pointed
out that in such systems it is very importan t to permit certain
variables to alter.46 The steady state of an engine with a governor is
unlikely to be maintained if the position of the balls of the
governor is clamped. Similarly a tightrope walker with a balancing
pole will not be able to maintain his balance except by varying the
forces which he exerts upon the pole.
(9) Return ing now to the conceptu al model sugge sted in
paragraph 7, let us take one further step toward making this model
comparable with Baline se society. Let us substitute for the umpire a
village counc il composed of all the players. We now have a system
which presents a number of analog ies to our balanc ing acroba t.
When they speak as members of the village counc il, the players by
hypothesis are interested in maintaining the steady state of the
system—that is, in preventing the maximization of any simple
variable the excessive increas e of which would produc e irreversib le
change. In their daily life, however , they are still engag ed in simple
competit ive strategies.
(10) The next step toward making our model resemble Baline se
society more closely is clearly to postulate in the character structu re
of the individua ls and/or in the contexts of their daily life those
factors which will motivate them toward maintenanc e of the steady
state not only when they speak in coun cil, but'also in their other
interpersona l relations . These factors are in fact recogn izable in Bali
and have been enumerated above . In our analysis of why Balines e
society is nonsch ismogenic , we noted that the Balinese child learns
46 W. R. Ashby, "Effect of Controls on Stability," Nature, clv, no. 3930,
February 24, 1945, 242-43.
133

to avoid cumulative interaction , i.e., the maximization of certain
simple variab les, and that the social organization and contexts of
daily life are so const ructed as to preclude competitive interaction.
Furthe r, in our analysis of the Baline se ethos, we noted recurrent
valua tion: (a) of the clear and static definition of status and spatial
orientation, and (b) of balan ce and such movement as will conduc e
to balan ce.
In sum it seems that the Balinese extend to human relationships
attitudes based upon bodily balance, and that they generalize the
idea that motion is essential to balance. This last point gives us, I
believe, a partial answer to the question of why the society not
only continues to function but functions rapidly and busily,
continually undertaking ceremonial and artistic tasks which are not
economically or competitively determined. This steady state is
maintained by continual nonprogressive change.
Schismog enic Sys tem versus the Steady State
I have discuss ed two types of social system in such schematic
outline that it is possib le to state clearly a contrast between them.
Both types of system, so far as they are capab le of maintain ing
themselves without progr essive or irrevers ible change , achieve the
steady state. There are, how-eve r, profound differences between
them in the m anner in which the steady state is regulated .
The Iatmul system, which is here used as a proto type of
schismogenic systems, includes a number of regene rative causal
circuits or vicious circles. Each such circuit consis ts of two or more
individua ls (or groups of individu als) who participate in potentially
cumulative interaction. Each human individua l is an energy source
or "relay," such that the energy used in his respons es is not derived
from the stimuli but from his own metabolic proces ses. It therefore
follows that such a schismogenic system is—unl ess contro lled—
liable to exces sive increase of those acts which characterize the
schismogeneses. The anthropo logis t who attempts even a qualitative
description of. such a system must therefo re identify: (1) the
individua ls and groups involved in schismogenes is and the routes of
134

communica tion between them; (2) the categor ies of acts and
contexts chara cteristic of the schismogeneses ; (3) the proce sses
whereb y the individua ls become psychologica lly apt to perfor m
these acts and/or the natur e of the contexts which force these acts
upon them; and lastly, (4) he must identify the mechanis ms or
factors which control the schismogeneses . These controlling factors
may be of at least three distinct types: (a) degenera tive causa l loops
may be superposed upon the schismogeneses so that when the latter
reach a certain intensi ty some form of restraint is applied as
occurs in Occident al systems when a gove rnment intervenes to limit
econo mic competition; (b) there may be, in addition to the
schismogeneses already considered , other cumulative interactions
acting in an oppos ite sense and so promoting social integr ation
rather than fission; (c) the increas e in schismogenesis may be
limited by factors which are internally or externally environmental
to the parts of the schismogenic circuit. Such factors which have
only small restraining effect at low intensities of schismogenesis
may increase with increase of intensity. Friction, fatigue, and
limitation of energy supply would be examples of such factors.
In contr ast with these schis mogenic systems, Baline se society is
an entirely different type of mechanis m, and in de-scribing it the
anthropo logis t must follow entirely different procedures , for which
rules cannot as yet be laid down. Since the class of
"nons chismogenic " social systems is defined only in nega tive terms,
we cannot assume that members of the class will have common
characteristics. In the analysts of the Balines e system, howeve r, the
following steps occurred, and it is possible that some at least of
these may be applicable in the analysis of other cultures of this
class: (1) it was observed that schismogenic sequences are rare in
Bali; (2) the exceptional cases in which such sequen ces occu r were
invest igated; (3) from this inves tigation it appeared, (a) that in
general the contexts which recur in Balinese social life preclude
cumulative interaction and ( b) that childhood exper ience trains the
child away from seeking climax in personal interaction; (4) it was
shown that certain positive values—r elated to balance— recur in the
culture and are incorpo rated into the character structu re during
child-hood, and, further, that these values may be specifically re-
lated to the steady state; (5) a more deta iled study is now required to
135

arrive at a systematic statement abou t the self-correcting
characteristics of the system. It is evident that the ethos alone is
insufficient to maintain the steady state. From time to time the
village or some other entity does step in to correct infrac tions . The
nature of these instances of the working of the corrective mechanis m
must be studied ; but it is clear that this intermittent mechanis m is
very different from the continual ly acting restraints which must be
present in all schismogenic systems.
136

Style, Grac e, and Information in
Primitive Ar t*
Introduction
This pape r consists of several still-sep arate attempts to map a
theory associated with culture and the nonv erbal arts. Since no one
of these attempts is completel y succe ssful , and since the attempts do
not as yet meet in the middle of the territory to be mapped, it. may
be usefu l to state, in non-te chnica l language , what it is I am after.
Aldous Huxle y used to say that the central proble m for humanity
is the qu est fo r grace. This word he us ed in what he t hought was th e
sense in, which it is used in the New Testament. He explained the
word , however , in his own terms. He argued—l ike Walt Whit man—
that the communication and behavio r of animals has a naivete, a
simplicity, which man has lost. Man's behav ior is corrup ted by
deceit—even s elf-dece it—b y purpose , and by self-consciousn ess. As
Aldous saw the matter, man has lost the "grace" which animals still
have.
In terms of this contra st, Aldous argued that God resembles the
animals rather than man: He is ideally unabl e to deceive and
incapab le of interna l confusion s.
In the total scale of being s, therefo re, man is as if displaced
sideways and lacks that grace which the animals have and which
God has.
I argue that art is a part of man's quest for grace; some-times his
ecstasy in partial success, sometimes his rage and agony at failure.
I argue also that there are many species of grace within the
major genus; and also that there are many kinds of failure and
* This essay was a position paper for the Wenner-Gren Conference on Primitive
Art, 1967. It is here reprinted from A Study of Primitive Art, edited by Dr.
Anthon y Forge, to be published by Oxford University Press, by permission of the
publisher .
137

frustration and departure from grace. No doubt each culture has its
characteristic species of grace toward which its artists strive, and
its own species of failure.
Some cultures may foster a negative approach to this difficult
integration, an avoidance of complexity by crass preference either
for total consciousness or total unconsciousness. Their art is
unlikely to be "great."
I shall argue that the problem of grace is fundamentally a
problem of integration and that what is to be integrated is the
diverse parts of the mind—especially those multiple levels of
which one extreme is called "consciousness" and the other the
"unconscious." For the attainment of grace, the reasons of the heart
must be integrated with the reasons of the reason.
Edmund Leach has confronted us, in this conference, with the
question: How is it that the art of one culture can have meaning or
validity for critics raised in a different culture? My answer would
be that, if art is somehow expressive of something like grace or
psychic integration, then the success of this expression might well
be recognizable across cultural barriers. The physical grace of cats
is profoundly different from the physical grace of horses, and yet a
man who has the physical grace of neither can evaluate that of
both.
And even when the subject matter of art is the frustration of
integration, cross-cultural recognition of the products of this
frustration is not too surprising.
The central question is: In what form is information about
psychic integration contained or coded in the work of art?
138

Style and Mean ing
They say that "ever y picture tells a story," and this generalization
holds for most of art if we exclude "mere" geometric orna mentation.
But I want precisely to avoid analyzing the "story." That aspec t of
the work of art which can most easily be reduc ed to words— the
mytho logy connec ted with the subje ct matter—i s not what I want to
discuss . I shall not even mention the uncons cious mythology of
phallic symbol-is m, excep t at the end .
I am concerned with what important psychic information is in
the art object quite apart from what it may "represent." "Le style
est l'homme meme" ("The style is the man him-self") (Buffon).
What is implicit in style, materials, composition, rhythm, skill, and
so on?
Clearly this subject matter will include geometrical orna-
mentation along with the composition and stylistic aspects of more
representational works.
The lions in Trafalgar Square could have been eagles or
bulldogs and still have carried the same (or similar) messages
about empire and about the cultural premises of nineteenth-century
England. And yet, how different might their message have been
had they been made of wood!
But representationalism as such is relevant. The extremely
realistic horses and stags of Altamira are surely not about the same
cultural premises as the highly conventionalized black outlines of a
later period. The code whereby perceived objects or persons (or
supernaturals) are transformed into wood or paint is a source of
information about the artist and his culture.
It is the very rules of transformation that are of interest to me—
not the message, but the code.
My goal is not instrumental. I do not want to use the
transformation rules when discovered to undo the transformation
or to "decode" the message. To translate the art object into
mythology and then examine the mythology would be only a neat
way of dodging or negating the problem of "what is art?"
139

I ask, then, not about the meaning of the encoded message but
rather about the meaning of the code chosen. But still that most
slippery word "meaning" must be defined.
It will be convenient to define meaning in the most general
possible way in the first instance.
"Meaning" may be regarded as an approximate synonym of
pattern, redundancy, information, and "restraint," within a
paradigm of the following sort:
Any aggregate of events or objects (e.g., a sequence of
phonemes, a painting, or a frog, or a culture) shall be said to
contain "redundancy" or "pattern" if the aggregate can be divided
in any way by a "slash mark," such that an observer perceiving
only what is on one side of the slash mark can guess , with better
than random success, what is on the other side of the slash mark.
We may say that what is on one side of the slash contains
information or has meaning about what is on the other side. Or, in
engineer's language, the aggregate contains "redundancy." Or,
again, from the point of view of a cybernetic observer, the
information available on one side of the slash will restrain (i.e.,
reduce the probability of) wrong guessing. Examples:
The letter T in a given location in a piece of written English
prose proposes that the next letter is likely to be an H or an R or a
vowel. It is possible to make a better than random guess across a
slash which immediately follows the T. English spelling contains
redundancy.
From a part of an English sentence, delimited by a slash, it is
possible to guess at the syntactic structure of the remainder of the
sentence.
From a tree visible above ground, it is possible to guess at the
existence of roots below ground. The top provides information
about the bottom.
From an arc of a drawn circle, it is possible to guess at the
position of other parts of the circumference. (From the diameter of
an ideal circle, it is possible to assert the length of the
circumference. But this is a matter of truth within a tautological
system. )
From how the boss acted yesterday, it may be possible to guess
how he will act today.
140

From what I say, it may be possible to make predictions about
how you will answer. My words contain meaning or information
about your reply.
Telegraphist A has a written message on his pad and sends this
message over wire to B, so that B now gets the same sequence of
letters on his message pad. This transaction (or "language game" in
Wittgenstein's phrase) has created a redundant universe for an
observer O. If 0 knows what was on A's pad, he can make a better
than random guess at what is on B's pad.
The essence and raison d'etre of communication is the creation
of redundancy, meaning, pattern, predictability, information, and/or
the reduction of the random by "restraint."
It is, I believe, of prime importance to have a conceptual system
which will force us to see the "message" (e.g., the art object) as
both itself internally patterned and itself a part of a larger patterned
universe—the culture or some part of it.
The characteristics of objects of art are believed to be abou t, or
to be partly derived from, or determined by, other characteristics of
cultural and psychological systems. Our problem might therefore
he oversimply represented by the diagram:
[Characteristic, of art object/Characteristics of rest of culture]
where square bracke ts enclo se the univers e of relevance , and
where the oblique stroke represen ts a slash across which some
guessing is possib le, in one direction or in both. The probl em, then,
is to spell out what sorts of relationsh ips, correspondence s, etc.,
cross or transcend t his oblique s troke .
Consid er the case in which I say to you, "It's raining," and you
guess that if you look out the window you will see raindrops . A
similar diagra m will serve :
[Characteristics of "It's raining"/Perception of raindrops]
Notice, however , that this case is by no means simple. Only if
you know the language and have some trust in my veracity will you
be able to make a guess abou t the rain-d rops. In fact, few peop le in
this situation restrain them-selves from seemingly duplicating their
information by looking out of the window . We like to prove that our
141

guesse s are right, and that our friends are honest. Still more
important, we like to test or verify the correctness of our view of our
relation ship t o others .
This last point is nontrivial. It illustrates the necessarily
hierarchic structure of all communicational systems: the fact of
conformity or nonconformity (or indeed any other relationship)
between parts of a patterned whole may itself be informative as
part of some still larger whole. The matter may he diagrammed
thus:
[("It's raining"/raindrop.)/Yom—me relationship)
where redund ancy acros s the slash mark within the smaller
unive rse enclosed in round brack ets proposes (is a message about) a
redundan cy in the larger unive rse enclos ed in squa re bracke ts.
But the message "It's raining" is itself convent ional ly coded and
interna lly patterned, so that severa l slash marks could be drawn
across the m essage i ndica ting patterning with in the m essage i tself.
And the same is true of the rain. It, too, is patterned and
structu red. From the direction of one drop, I could predi ct the
direction of others. And so on.
But the slash marks across the verba l message "It's raining" will
not correspond in any simple way to the slash marks across the
raind rops.
If, instead of a verba l message, I had given you a picture of the
rain, some of the slashes on the picture would have corresponded
with slashes on the perceived rain.
This differenc e provides a neat formal criterion to separa te the
"arbitrary" and digital coding charac teristic of the verbal part of
languag e from the iconic coding o f dep iction.
But verba l description is often iconic in its larger structure. A
scientist descr ibing an earthwor m might start at the head end and
work down its length—thus produc ing a description iconic in its
sequenc e and elonga tion. Here again we observe a hierarch ic
structu ring, digital or verbal at one l evel and iconic at anothe r.
Levels and Logical Types
142

"Leve ls" have been mentioned : (a) It was noted that the
combina tion of the message "It's raining" with the perception of
raindrop s can itself constitute a message about a universe of
personal relation s; and (b) that when we change our focus of
attention from smaller to larger units of message material, we may
discover that a larger unit contains iconic coding though the smaller
parts of which it was made are verbal: the verba l description of an
earthwor m may, as a whol e, be elongat ed.
The matter of levels now crops up in anothe r form which is
crucia l for any epistemolog y of art:
The word "know" is not merely ambiguous in cove ring both
conna itre (to know through the senses , to recogniz e or perceive ) and
savoir (to know in the mind), but varies —ac tively shifts— in
meaning for basic systemic reasons. That which we know through
the senses can become knowledge in the mind.
"I know the way to Cambridge" might mean that I have studied
the map and can give you directions . It might mean that I can recall
details all along the route. It might mean that when driving that
route I recogn ize many deta ils even though I could reca ll onl y a few.
It might mean that when driving to Cambridge I can trust to "habit"
to make me turn at the right points, withou t having to think where I
am going. And so on.
In all cases , we deal with a redundanc y or patterning of a quite
complex sort:
[("I know . . ."/my mind)//the road]
and the difficulty is to determine the nature of the patterning
within the round brackets , or, to put the matter anothe r way: what
parts of the mind are redundant with the particular message abou t
"knowing ."
Last, there is a special form of "knowing" which is usually
regarded as adapt ation rather than information . A shark is beau tifully
shaped for locomotion in water , but the geno me of the shark surely
does not contain direct information abou t hydrodynamics. Rathe r,
the geno me must be supposed to contain information or instructions
which are the complement of hydrodynamics. Not hydrodynamics,
but what hydrodynamics requir es, has been built up in the shark's
geno me. Similarly, a migratory bird perhaps does not know the way
143

to its destination in any of the senses outlined above , but the bird
may contain the comple mentary instructions necessar y td cause it to
fly right.
"Le coeur a ses raison s que la raison ne conna it point" ("The
heart has its reason s which the reason does not at all perce ive") . It is
this—the co mplex la yering of cons ciousne ss and unconsc iousness—
that creates difficulty when we try to discuss art or ritual or
mytholog y. The matter of levels of the mind has been discu ssed
from many points of view , at least four of which must be mentioned
and woven into any scientific approach to art:
(1) Samuel Butler's insistence that the better an organism
"knows" something, the less conscious it becomes of its knowledge,
i.e., there is a proces s whereb y knowledge (or "habi t" —whethe r of
action, perception, or thought ) sinks to deeper and deeper levels of
the mind. This pheno menon, which is centra l to Zen discipline (cf.
Herrigel, Zen in the Art of Archery), is also relevan t to all art and all
skill.
(2) Adalbert Ames' demonstrations that the conscious, three-
dimensional visual images, which we make of that which we see,
are made by processes involving mathematical premises of
perspective, etc., of the use of which we are totally unconscious.
Over these processes, we have no voluntary control. A drawing of
a chair with the perspective of van Gogh affronts the conscious
expectations and, dimly, reminds the consciousness of what had
been (unconsciously) taken for granted.
(3) The Freudian (especially Fenichel's) theory of dreams as
metaphors coded according to primary process. I shall consider
style—neatness, boldness of contrast, etc.—as metaphoric and
therefore as linked to those levels of the mind where primary
process holds sway.
(4) The Freudian view of the unconscious as the cellar or
cupboard to which fearful and painful memories are con-signed by
a process of repression.
Class ical Freudian theory assumed that dreams were a secondary
produc t, created by "dream work." Material unacceptab le to
consc ious thought was suppos edly translated into the metaphori c
idiom of primary process to avoid waking the dreamer. And this may
be true of those items of information which are held in the
144

uncon scious by the proce ss of repression. As we have seen,
howeve r, many other sorts of information are inaccessible to
conscious inspection, including most of the premises of mammalian
interaction. It would seem to me sensible to think of these items as
existing primari ly in the idiom of primary process , only with
difficulty to be translated into "rationa l" terms. In other words , I
believe that much of early Freudian theory was upside down. At that
time many thinke rs regarded consc ious reason as normal and self-
explanatory while the unconscious was regard ed as mysterious ,
needing proof, and needing explanat ion. Repr ession was the
explanation, and the uncon scious was filled with thoughts which
could have been consciou s but which repression and dream work
had distorted. Today we think of consciou sness as the mysteriou s,
and of the computat ional methods of the unconsciou s, e.g., primary
process , as con tinua lly active, neces sary, and all-embracing.
These considera tions are especially relevant in any at-tempt to
derive a theory of art or poetry. Poetry is not a sort of distorted and
decorat ed prose , but rather prose is poetry which has been stripped
down and pinned to a Procrus tean bed of logic. The computer men
who would progr am the transla tion of language s sometimes forget
this fact abou t the primary nature of langu age. To try to const ruct a
machine to translate the art of one culture into the art of another
would be equally silly.
Allegor y, at best a distasteful sort of art, is an inversion of the
normal creative process . Typical ly an abstract relation, e.g., between
truth and justice, is first conce ived in rational terms. The
relationsh ip is then metaphor ized and dolled up to look like a
product of primary process. The abstractions are person ified and
made to participate in a pseudo myth, and so on. Much advertising
art is allegor ical in this sense, that the creative process is inverted.
In the cliche system of Anglo-Saxon s, it is commonly assumed
that it would be someho w better if what is unconscious were made
conscious . Freud, even, is said to have said, "Where id was, there
ego shall be," as though such an increase in consciou s knowledge
and control would be both possible and, of course , an improve ment.
This view is the produc t of an almost totally distorted epistemology
and a totally distorted view of what sort of thing a man, or any other
organis m, is.
145

Of the four sorts of unconsciou sness listed above , it is very clear
that the first three are necess ary. Consc iousnes s, for obvious
mechanica l reasons ,47 must always be limited to a rather small
fraction of mental process . If useful at all, it must therefore be
husbanded . The unconsc iousness assoc iated with habit is an
econo my both of thought and of consciousnes s; and the same is true
of the inacc essab ility of the process es of percept ion. The conscious
organism does not requi re (for pragmatic purposes ) to know how it
perceives —onl y to know what it perceive s. (To sugges t that we
might operate without a foundat ion in primary proces s would be to
sugges t that the human brain ough t to be differen tly structured.) Of
the four types, only the Freud ian cupboard for skeletons is perhaps
undesi rable and could be obviated. But there m ay still be advan tages
in keeping the skeleton off the dining room table.
In truth, our life is such that its uncon scious components are
continuousl y presen t in all their multiple forms. It follows t hat in our
relationships we continuous ly exch ange messages abou t these
unconsc ious materials, and it beco mes important also to exchange
metamessages b y which we tell each other what order and species of
unconsc iousness (or consc iousnes s) attaches to ou r messages.
In a merely pragmatic way, this is importan t becau se the orders
of truth are differen t for different sorts of messages . Insof ar as a
message is consc ious and volun tary, it could be deceitful. I can tell
you that the cat is on the mat when in fact she is not there. I can tell
you "I love you" when in fact I do not. But discou rse about
relationship is commonly accompanied by a mass of semivolunta ry
kines ic and autonomic signal s which provide a more trustworth y
comment on the verba l message.
Similarly with skill, the fact of skill indica tes the presen ce of
large uncon scious components in the performance.
It thus beco mes relevan t to look at any work of art with the
quest ion: What components of this message materia l had what
orders of unconsc iousnes s (or consciousne ss) for the artist? And this
47 Consider the impossibility of constructing a television set which would report upon
its screen all the workings of its component parts, including especially those parts
concerned in this reportin g.
146

question , I believe, the sensitive critic usually asks, though perhaps
not cons cious ly.
Art beco mes, in this sense , an exercise in communicat ing about
the species of unconsc iousness . Or, if you prefer it, a sort of play
behavior whose function is, amongst other things, to practice and
make more pe rfect communicat ion o f this kind.
I am indebted to Dr. Anthon y Forge for a quotation from Isado ra
Duncan : "If I could tell you what it meant, there would be no point
in danc ing it."
Her statement is ambiguous. In terms of the rather vulgar
premises of our culture, we would translate the statement to mean:
"There would then be no point in dancing it, be-caus e I could tell it
to you, quicker and with less ambiguity, in words ." This
interpret ation goes along with the silly idea that it would be a good
thing to be cons cious of everything o f which we are uncons cious .
But there is anothe r possible meaning of Isador a Duncan's
remark: If the message were the sort of message that could be
communica ted in words , there would be no point in dancing it , but it
is not that sort of message . It is, in fact, precisely the so rt of message
which would be falsified if communicated in words, becau se the use
of word s (othe r than poetry) would imply that this is a fully
conscious and vo lunta ry message, a nd this would be simply untru e.
I believe that what Isadora Duncan or any artist is trying to
communica te is more like: "This is a particular sort of partly
uncon scious message. Let us engage in this particular sort of partly
uncon scious communica tion." Or perhaps : "This is a message about
the interface between consc ious a nd unconscious .
The message of skill of any sort must always be of this kind.
The sensa tions and qualities of skill can never be put in words , and
yet the fact of skill is consciou s.
The artist's dilemma is of a pecul iar sort. He must practice in
order to perform the craft components of his job. But to practice has
always a double effect. It makes him, on the one hand , more able to
do what ever it is he is attempting; and, on the other hand , by the
pheno menon of habit formation, it makes him less awar e of how he
does it.
If his attempt is to communicate abou t the uncons cious
component s of his performance, then it follows that he is on a sort of
147

moving stairway (or escalator) about whose position he is trying to
communicate but whose movement is itself a function of his efforts
to communicate.
Clear ly, his task is impossibl e, but, as has been remarked, some
people do it very prettily.
Primary Proce ss
"The heart has its reasons which the reason does not at all
perceive." Among Anglo-Saxons , it is rather usual to think of the
"reasons" of the heart or of the unconsc ious as inchoate forces or
pushes or heavings—wha t Freud called Trieben. To Pasca l, a
French man, the matter was rather different, and he no doub t though t
of the reason s of the heart as a body of logic or computation as
precise and co mplex as the reasons o f con sciousne ss.
(I have noticed that Anglo-Saxon anthropologists some-times
misunderstand the writings of Claude Levi-Strauss for precisely
this reason. They say he emphasizes too much the intellect and
ignores the "feelings." The truth is that he assumes that the heart
has precise algorithms.)
These algorithms of the heart, or, as they say, of the un-
conscious, are, however, coded and organized in a manner totally
different from the algorithms of language. And since a great deal
of conscious thought is structured in terms of the logics of
language, the algorithms of the unconscious are doubly
inaccessible. It is not only that the conscious mind has poor access
to this material, but also the fact that when such access is achieved,
e.g., in dreams, art, poetry, religion, intoxication, and the like,
there is still a formidable problem of translation.
This is usually expressed in Freudian language by saying that
the operations of the unconscious are structured in terms of
primary process, while the thoughts of consciousness (especially
verbalized thoughts) are expressed in secondary process.
Nobody, to my knowledge, knows anything about secondary
process. But it is ordinarily assumed that everybody knows all
148

about it, so I shall not attempt to describe secondary process in any
detail, assuming that you know as much about it as I.
Primary process is characterized (e.g., by Fenichel) as lacking
negatives, lacking tense, lacking in any identification of linguistic
mood (i.e., no identification of indicative, subjunctive, optative,
etc.) and metaphoric. These characterizations are based upon the
experience of psychoanalysts, who must interpret dreams and the
patterns of free association.
It is also tr ue that the subje ct matter of pr imary-proce ss discours e
is differen t from the subjec t matter of language and consciousnes s.
Consc iousness talks about things or persons , and attache s predi cates
to the specific things or person s which have been mentioned . In
primary proce ss the things or person s are usually not identified, and
the focus of the discours e is upon the relationships which are
asserted to obtain between them. This is really only anoth er way of
saying that the discourse of primary proces s is metaphoric . A
metaphor retains unchanged the relationship which it "illustrates"
while substituting other things or persons for the relata. In a simile,
the fact that a metaphor is being used is marked by the insertion of
the words "as if" or "like." In primary process (as in art) there are
no markers to indicate to the conscious mind that the message
material is metaphoric.
(For a schizophrenic, it is a major step towards a more
conventional sanity when he can frame his schizophrenic ut-
terances or the comments of his voices in an "as if" terminology.)
The focus of "relationship" is, however, somewhat more narrow
than would be indicated merely by saying that primary-process
material is metaphoric and does not identify the specific relata.
The subject matter of dream and other primary-process material is,
in fact, relationship in the more narrow sense of relationship
between self and other persons or between self and the
environment.
Anglo-Saxons who are uncomfortable with the idea that
feelings and emotions are the outward signs of precise and
complex algorithms usually have to be told that these matters, the
relationship between self and others, and the relationship between
self and environment, are, in fact, the subject matter of what are
called "feelings"—love, hate, fear, confidence, anxiety, hostility,
149

etc. It is unfortunate that these abstractions referring to pattern s of
relationship have received names, which are usually handled in
ways that assume that the "feelings" are mainly characterized by
quantity rather than by precise pattern. This is one of the
nonsensical contributions of psychology to a distorted
epistemology.
Be all that as it may, for our present purposes it is important to
note that the characteristics of primary process as described above
are the inevitable characteristics of any communicational system
between organisms who must use only iconic communication. This
same limitation is characteristic of the artist and of the dreamer
and of the prehuman mammal or bird. (The communication of
insects is, perhaps, an-other matter.)
In iconic communication, there is no tense, no simple negative,
no modal marker.
The absence of simple negatives is of especial interest be-cause
it often forces organisms into saying the oppo site of what they
mean in order to get across the proposition. that they mean the
opposite of what they say.
Two dogs approach each other and need to exchange the
message: "We are not going to fight." But the only way in which
fight can be mentioned in iconic communication is by the showing
of fangs. It is then necessary for the dogs to discover that this
mention of fight was, in fact, only exploratory. They must,
therefore, explore what the showing of fangs means. They
therefore engage in a brawl; discover that neither ultimately
intends to kill the other; and, after that, they can be friends.
(Consider the peace-making ceremonials of the Andaman
Islanders. Consider also the functions of inverted statement or
sarcasm, and other sorts of humor in dream, art, and mythology.)
In general, the discourse of animals is concerned with rela-
tionship either between self and other or self and environment. In
neither case is it necessary to identify the relata. Animal A tells B
about his relationship with B and he tells C about his relationship
with C. Animal A does not have to tell animal C about his
relationship with B. Always the relata are perceptibly present to
illustrate the discourse, and always the discourse is iconic in the
sense of being composed of part actions ("intention movements")
150

which mention the whole action which is being mentioned. Even
when the cat asks you for milk, she cannot mention the object
which she wants (unless it be perceptibly present). She says,
"Mama, mama," and you are supposed from this invocation of de-
pendency to guess that it is milk that she requires.
All this indicates that primary-process though ts and the
communica tion of such thought s to others are, in an evolutionary
sense, more archaic than the more consc ious operations of language,
etc. This has implica tions for the whol e econo mics and dynamic
structure of the mind. Samuel Butle r was perhaps first to point out
that that which we know be st is that of which we are least consc ious,
i.e., that the process of habit formation is a sinking of knowledg e
down to less consc ious and more archaic levels. The uncons cious
contains not only the painful matters which consc iousness prefers to
not inspect , but also many matters which are so familiar that we do
not need to inspect them. Habit , therefore, is a major econo my of
conscious thought . We can do things without consc iousl y thinking
abou t them. The skill of an artist, or rather his demonstration of a
skill, beco mes a message about these parts of his unconsciou sness .
(But not perhaps a message from the un conscious .)
But the matter is not quite so simple. Some types of knowledge
can conveniently be sunk to unconscious levels, but other types
must be kept on the surface. Broadly, we can afford to sink those
sorts of knowledge which continue to be true regardless of changes
in the environment, but we must maintain in an accessible place all
those controls of behavior which must be modified for every
instance. The lion can sink into his unconscious the proposition
that zebras are his natural prey, but in dealing with any particular
zebra he must be able to modify the movements of his attack to fit
with the particular terrain and the particular evasive tactics of the
particular zebra.
The economics of the system, in fact, pushes organisms toward
sinking into the unconscious those generalities of relationship
which remain permanently true and toward keeping within the
conscious the pragmatics of particular instances.
The premises may, economically, be sunk, but particular
conclusions must be conscious. But the "sinking," though eco-
nomical, is still done at a price—the price of inaccessibility. Since
151

the level to which things are sunk is characterized by iconic
algorithms and metaphor, it becomes difficult for the organism to
examine the matrix out or which his conscious conclusions spring.
Conversely, we may note that what is common to a particular
statement and a corresponding metaphor is of a generality
appropriate for sinking.
Quant itative Limits of Consciousne ss
A very brief conside ration of the problem shows that it is not
conce ivabl y possib le for any system to be totally consciou s.
Suppose that on the screen of consc iousness there are reports from
many parts of the total mind, and consid er the addition to
consc iousness of those reports neces sary to cover what is, at a given
stage of evolu tion, not alread y cover ed. This addition will involve a
very great increa se in the circuit structur e of the brain but still will
not achieve total coverag e. The next step will be to cove r the
proce sses and events occurring in the circuit structure which we
have just added. And so on.
Clearly, the problem is insoluble, and every next step in the
approach to total consciousness will involve a great in-crease in
the circuitry required.
It follows that all organisms must be content with rather little
consciousness, and that if consciousness has any useful functions
whatever (which has never been demonstrated but is probably
true), then economy in consciousness will be of the first
importance. No organism can afford to be conscious of matters
with which it could deal at unconscious levels.
This is the economy achieved by habit formation.
Qualitative Limits of Consciousness
It is, of cours e, true for the TV set that a satisfac tory picture on
the screen is an indication that many parts of the machine are
152

work ing as they shou ld; and similar considerations apply to the
"screen" of consc iousness . But what is provided is only a very
indirect report of the work ing of all those parts. If the TV suffers
from a blown tube, or the man from a stroke , effects of this
patholog y may be evident enough on the screen or to consciousn ess,
but diagnosis must still be done b y an exp ert.
This matter has bearings upon the nature of art. The TV which
gives a distorted or otherwis e imperfec t picture is, in a sense,
communica ting about its uncons cious patho logies— exhibi ting its
symptoms; and one may ask whether some artists are not doing
something s imilar. But t his still won't do.
It is sometimes said that the distortions of art (say, van Gogh's
"Cha ir") are directly repres entative of what the artist "sees ." If such
statements refer to "seeing " in the simplest physical sense (e.g.,
remediabl e with spectacles), I presume that they are nonsense . If van
Gogh could only see the chair in that wild way, his eyes would not
serve properly to guide him in the very accura te placing of paint on
canvas. And, convers ely, a photogr aphica lly accura te representa tion
of the chair on the canvas would also be seen by van Gogh in the
wild wa y. Re would see no ne ed to distort the pa inting.
But suppose we say that the artist is painting today what he saw
yesterda y—or that he is painting what he somehow knows that he
migh t see. "I see as well as you do—but do you realize that this
other way of seeing a chair exists as a human poten tiality? And that
that poten tiality is alwa ys in you and in me?" Is he exhib iting
symptoms which he might have , because the whol e spectrum of
psychopatholog y is possible for us all?
Intoxi cation by alcoho l or drugs may help us to see a distorted
world, and these distortions may be fascina ting in that we recogn ize
the distortions as ours. In vino pars veritatis. We can be humbled or
aggrandiz ed by realizing that this, too, is a part of the human self, a
part of Truth. But intoxic ation does not increase skill—at best it
may release sk ill previous ly acqu ired.
Without skill is no art.
Consid er the case of the man who goes to the blackbo ard —or to
the side of his cave—and draws, freehand, a perfe ct reind eer in its
posture of threat. He cannot tell you about the drawing of the
reindee r ("If he could, there would be no point in drawing it"). "Do
153

you know that his perfect way of seeing— and drawing—a reinde er
exists as a human potentiality?" The consu mmate skill of the
draftsman validate s the artist's message about his relationship to the
animal—his e mpath y.
(They say the Altamira things were made for sympatheti c
hunting magic. But magic only needs the crudest sort of rep-
resentations . The scrawl ed arrows which deface the beau tiful
reindeer may have been magical—pe rhaps a vulgar attempt to
murder the artist, like moustach es scrawled on the M ona Li sa.)
The Corr ectiv e Nature of Art
It was no ted a bove t hat consc iousness is necessarily selective and
partial, i.e., that the content of consciousne ss is, at best, a small part
of truth abou t the self. But if this part be selected in any systematic
manner, it is certain that the partial truths of consc iousnes s will be,
in aggrega te, a distortion of the truth of some larger whole .
In the case of an iceberg, we may guess , from what is above
surface, what sort of stuff is below; but we cannot make the same
sort of extrapol ation from the conten t of consc iousnes s. It is not
merely the selectivity of prefe rence , wher eby the skeletons
accumulate in the Freud ian unconscious , that makes such
extrapola tion unsound . Such a selection by preference would only
promote optimism.
What is serious is the crosscu tting of the circuitry of the mind. If,
as we must believe, the total mind is an integra ted network (of
propos itions, images, processes , neural pathology, or what have you
—acco rding t o wha t scientific language you prefer to use), and if the
conten t of consciousne ss is only a sampling of differen t parts and
localities in this net-work ; then, inevitably, the consc ious v iew of the
network as a whole is a monstrous denia l of the integrat ion of that
whole. From the cutting of consc iousness , what appe ars above the
surface is arcs of circuits instead of either the complete circuits or
the larger complete circuits of circuits.
What the unaided consciousn ess (unaided by art, dreams, and the
like) can n ever apprec iate is the systemic nature of mind.
154

This notion can conv enien tly be illustrated by an analog y: the
living human body is a complex, cyberne tically integra ted system.
This system has been studied by scien tists—mostly medical men—
for many years. What they now know abou t the body may aptly be
compared with what the unaided consciou sness knows about the
mind. Being docto rs, they had purposes : to cure this and that. Their
research efforts were therefore focused (as attention focus es the
consciousne ss) upon those short trains of causa lity which they could
manipulate, by means of drugs or othe r intervent ion, to correct more
or less specific and identifiable states or symptoms. Whenever they
discover ed an effective "cure" for something , research in that area
ceased and attention was directed elsewhere. We can now preven t
polio, but nobod y knows much more abou t the systemic aspec ts of
that fascinating diseas e. Research on it has ceased or is, at best,
confined to improving the vaccines.
But a bag of tricks for curing or prevent ing a list of specified
disease s provides no overall wisdom. The ecolog y and popul ation
dynamics of the species has been disrup ted; parasites have been
made immune to antibiotics; the relationship between mother and
neon ate has b een a lmost d estroyed; and so on.
Char acteristically, errors occu r wherev er the altered causa l chain
is part of some large or small circuit structu re of system. And the
remainder of our technolog y (of which medical science is only a
part) bids fair to disrupt t he rest of our ecolog y.
The point, however , which I am trying to make in this pape r is
not an attack on medical science but a demonstr ation of an
inevitable fact: that mere purpos ive ration ality unaided by such
pheno mena as art, religion, dream, and the like, is neces sarily
pathogeni c and destructive of life; and that its virulen ce springs
specifically from the circumstance that life depends upon
interlock ing circuits of conting ency, while consciousnes s can see
only such short arcs of su ch circuits as human purpose may direct.
In a word, the unaided consciou sness must always involve man
in the so rt of stupid ity of which evolution was gu ilty when she urged
upon the dinos aurs the common-sen se values of an armaments race.
She inevitably realized her mistake a million years later and wiped
them out.
155

Unaided consciousnes s must alwa ys tend toward hate; not only
becaus e it is good common sense to exterminate the other fellow,
but for the more profound reason that, seeing only arcs of circuits,
the individua l is continua lly surpr ised and necessa rily anger ed when
his hardhe aded po licies re-turn to plague the inven tor.
If you use DDT to kill insect s, you may succeed in reducing the
insect popul ation so far that the insectivores will starve . You will
then have to use more DDT than be-fore to kill the insects which the
birds no longe r eat. More probabl y, you will kill off the birds in the
first round when they eat the poisoned insects. If the DDT kills off
the dogs, you will have to have more police to keep down the
burglars. The burglars will beco me better armed and more cunn ing
… and s o on.
That is the sort of world we live in—a world of circuit structur es
—and love can survive only if wisdo m (i.e., a sense or recogni tion
of the fact of circuitry) has an effective voice .
What has been said so far propose s questions abou t any
particular work of art somewhat different from those which have
been convent ional ly asked by anthropo logis ts. The "culture and
person ality schoo l," for example, has traditionally used pieces of art
or ritual as samples or probes to revea l particula r psychologic al
themes or states.
The quest ion has been: Does the art tell us abou t what sort of
person made it? But if art, as sugges ted above , has a positive
function in maintain ing what I called "wisdo m," i.e., in correcting a
too purpo sive view of life and making the view more systemic, then
the question to be asked of the given work of art beco mes: What
sorts of correction in the direction of wisdo m would be achieved by
creating or viewing this work of art?
The que stion beco mes dynamic rather than static.
Analys is of Baline se Pa inting
Turning now from the conside ration of epistemolog y to a specific
art style, we no te first what i s most gene ral and m ost obvious .
156

With almost no exceptions, the behav iors called art or their
products (also called art) have two characteristics: they require or
exhibit skill, and t hey contain redund ancy or pattern.
But those two characteristics are not separate: the skill is first in
maintain ing and then in modulat ing the redundanci es.
The matter is perhaps most clear where the skill is that of the
journe yman and the redundanc y is of comparatively low order. For
example, in the Balinese painting by Ida Bagus Djati Sura of the
village of Batuan , 1937 and in almost all painting of the Batuan
school, skill of a certain elementary but highly discip lined sort was
exercised or practiced in the background of foliage. The
redundanci es to be achieved involve rather uniform and rhythmical
repetition of leaf forms, but this redund ancy is, so to speak, fragile.
It would be broken or interrupted by smudges or irregularities of
size or tone in the painting o f the su ccess ive leaves .
When a Batuan artist look s at the work of anothe r, one of the first
things he examines is the techn ique of the leafy background . The
leaves are first drawn, in free outline in pencil; then each outline is
tightly redefined with pen and black ink. When this has been done
for all the leaves , the artist begins to paint with brush and Chinese
ink. Each leaf is cove red with a pale wash. When these washes are
dry, each leaf receives a smaller concent ric wash and after this
another still smaller , and so on. The final result is a leaf with an al-
most white rim inside the inked outline, and succes sive steps of
darker and d arker color toward t he center of the leaf.
A "good" picture has up to five or six such success ive washes on
every leaf. (This particular painting is not very "good" in this sense.
The leaves are don e in onl y three o r fou r steps.)
The skill and the patterning so far discu ssed depend upon
muscular rote and muscular accuracy—achieving the perhaps not
negligible artistic level of a well-laid ou t field of turnips.
I was watching a very gifted American carpen ter-arch itect at
work on the wood work of a house he had designed . I commented on
the sureness and accur acy of each step. He said, "Oh, that. That's
only like using a typewrit er. You have to be able to do that without
thinking ."
But on top of this level of redund ancy is another. The unifo rmity
of the lowe r-level redundan cy must be modulated to give higher
157

orders of redund ancy. The leaves i n one ar ea m ust be different from
the leaves in another area, and these differences must be in some
way mutually redundant : they must be pa rt of a larger pattern.
Indeed , the function and necessity of the first-level control is
precisely to make the second level possible. The perceiver of the
work of art must receive information that the artist can paint a
unifo rm area of leaves because without this information he will not
be able to accep t as significant the variations in that uniformity.
Only the violinist who can control the quality of his notes can
use variation s of that quality for musical purposes .
This princip le is basic and accoun ts, I sugges t, for the almost
unive rsal linkage in aesthetics between skill and pattern. The
excep tions—e .g., the cult of natural landscape s, "found objects,"
inkblots, scattergrams, and the works of Jackson Pollock—s eem to
exemplify the same rule in reverse. In these cases, a larger
patterning seems to propose the illusion that the details must have
been controlled. Inter-mediate cases also occur : e.g., in Balin ese
carving, the natural grain of the wood is rather frequent ly used to
sugges t de-tails of the form or surface of the subject. In these cases,
the skill lies not in the draftsmanship of the details, but in the artist's
placement of hi s design wi thin t he thr ee-dimensional s tructure o f the
wood. A speci al "effect" is achieved, not by the mere
represent ation alism, but by the perceiver's partial awareness that a
physical system other than that of draftsman- ship has contributed to
determine h is percep tion.
We now turn to more complex matters, still concentrating
attention upon the most obv ious a nd elementar y.
Comp osition
(1) The delineation of leaves and other forms does not reach to
the edge of the picture but shades of f into darkn ess so that almost all
around the rectangl e there is a band of undifferen tiated dark
pigment. In other word s, the picture is framed within its own fade-
out. We are allowed to feel that the matter is in some sense "out of
158

this world"; and this in spite of the fact that the scene depicted is
familiar—the starting ou t of a cremation p rocess ion.
(2) The picture is filled. The composition leaves no open spaces .
Not only is none of the pape r left unpa inted , but no considerab le
area is left in uniform wash . The largest such areas are the very dark
patches at the bottom between the legs of the men.
To Occidental eyes, this gives an effect of "fussiness." To
psychiatric eyes, the effect is of "anxiety" or "compulsivity." We
are all familiar with the strange look of those letters from cranks,
who feel that they must fill the page.
(3) But before trying too fast to diagnos e or evaluate , we have to
note that the composition of the lower half of the picture, apart from
this filling of backg round space, is turbulen t. Not merely a depic tion
of active figures, but a swirling composit ion mounting upwards and
closed off by the contrasting direction of the gestures of the men at
the top of the pyramid.
The upper half of the picture, in contrast, is serene. Indeed, the
effect of the perfectly balanced women with offerings on their heads
is so serene that, at first glanc e, it appea rs that the men with musical
instruments must surely be sitting. (They are supposed to be moving
in pro cession.)
But this compositional structure is the reverse of the usual
Occidenta l. We expect the lower part of a picture to be the more
stable and expec t to see action and movement in the upper part—if
anywhere.
(4) At this point, it is appropriate to examine the picture as a
sexual pun and, in this connection, the internal evidence for sexual
reference is at least as strong as it is in the case of the Tangaroa
figure discussed by Leach. All you have to do is to set your mind
in the correct posture and you will see an enormous phallic object
(the cremation tower) with two elephants' heads at the base. This
object must pass through a narrow entrance into a serene courtyard
and thence onward and upward through a still more narrow
passageway. Around the base of the phallic object you see a
turbulent mass of homunculi, a crowd in which
Was none who would be foremost To
lead such dire attack;
159

But those behind cried "Forward!" And
those before cried "Back!"
And if you are so minded, you will find that Macau lay's poem
about how Horat ius kept the bridg e is no les s sexu al than the presen t
picture. The game of sexua l interpr etation is easy if you wan t to pl ay
it. No doubt the snake in the tree _ to the left of the picture could
also be woven into the sexual story.
It is still possible, however , that something is added to our
unders tanding of a work of art by the hypothesis that the subject
matter is doub le: that the picture represents both the start of a
cremation process ion and a phallus with vagina. With a little
imagina tion, we could also see the picture as a symbolic
represent ation of Balin ese social organiza tion in which the smooth
relations of etique tte and gaiety metaphor ically cover the turbu lence
of passion . And, of cours e, "Horat ius" is very evident ly an idealized
myth of nineteenth -century imperia l Engl and.
It is probably an error to think of dream, myth, and art as being
about any one matter other than relationship . As was mentioned
earlier, dream is metaphori c and is not particularly about the relata
mentioned in the dream. In the conven tiona l interpr etation of dream,
anothe r set of relata, often sexual , is substituted for the set in the
dream. But perhaps by doing this we only create another dream.
There indeed is no a priori reason for supposing that the sexua l
relata are any more primary or basic than a ny other set.
In genera l, artists are very unwi lling to accept interpretations of
this sort, and it is not clear that their objec tion is to the sexua l natu re
of the interpre tation. Rathe r, it seems that rigid focusing upon any
single set of relata destroys for the artist the more profound
significance of the work . If the picture were only abou t sex or only
about social organization, it would be trivial. It is nontr ivial or
profound precisely becau se it is about sex and social organiza tion
and cremation, and other thing s. In a word , it is only abou t re-
lationsh ip and not about any identifiable relata.
(5) It is appropri ate then to ask how the artist has hand led the
identification of his subject matter within the picture . We note first
that the cremation tower which occupies almost one-t hird of the
picture is almost invisible. It does not stand out against its
160

background as it should if the artist want ed to assert unequ ivocal ly
"this is a cremation." Notab ly also, the coffin, which might be
expected to be a focal point, is approp riately placed just below the
center but even so does not catch the eye. In fact, the artist has
inserted details which label the picture as a cremation scene but
these details beco me almost whimsical asides , like the snake and the
little birds in the trees. The women are carrying the ritually correct
offerings on their head s, and two men approp riately bring bamboo
containers of pa lm toddy, but thes e deta ils, too, are on ly whimsically
added. The artist plays down the subje ct identification and thereb y
gives major stress to the contrast between the turbulent and the
serene mentioned in 3, above .
(6) In sum, it is my opinion that the crux of the picture is the
interwoven contrast between the serene and the turbul ent. And a
similar con trast or combination was a lso p resen t, as we have seen, in
the painting of the leaves . There, too, an exuber ant freedo m was
overlaid by precision .
In terms of this conclus ion, I can now attempt an answer to the
question posed above: What sorts of correction, in the direction of
systemic wisdo m, could be achieved by creating or viewing this
work of art? In final analysis, the picture can be seen as an
affirmation that to choos e either turbulen ce or serenity as a human
purpose would be a vulgar error. The conceiv ing and creating of the
picture must have provid ed an exper ience which exposed this error.
The unity and integration of the picture assert that neither of these
contrasting poles can be chosen to the exclusion of the other,
because the poles are mutually dependent . This profound and
general truth is simultaneous ly asserted for the fields of sex, social
organiza tion, and death.
161

Comment on Par t II
Since World War II, it has been fashionab le to engage in
"interdiscip linary" resear ch, and this usual ly means, for example,
that an ecologis t will need a geologis t to tell him abou t the rocks and
soils of the particula r terrain which he is investigating . But there is
anothe r sense in which scientific work may claim to be
interdisciplinary.
The man who studies the arrang ement of leaves and branches in
the growth of a flowering plant may note an analog y between the
formal relations between stems, leaves, and buds, and the formal
relations that obtain between different sorts of words in a sentence .
He will think of a "leaf" not as something flat and green but as
something related in a particular way to the stem from which it
grows and to the second ary stem (or bud) which is formed in the
angle between leaf and primary stem. Similarly the modern linguis t
thinks of a "noun" not as the "name of a person, place, or thing," but
as a member of a class of words de-fined by their relationship in
sentence s tructure to "verb s" and other parts.
Those who think first of the "things" which are related (the
"relata") will dismiss any analog y between grammar and the
anato my of plants as far-fetched. After all, a leaf and a noun do not
at all resemble each other in outwa rd appe arance . But if we think
first of the relationships and conside r the relata as defined solely by
their relationsh ips, then we begin to wonder . Is there a profound
analog y between grammar and anatomy? Is there an
interdisciplinary science which shou ld concern itself with such
analog ies? What would such a science claim as its subject matter?
And why should we expe ct such far-flung analog ies to have
significance?
In dealing with any analog y, it is importan t to define exactly
what is claimed when we say that the analog y is meaningfu l. In the
present e xample, it is not claimed that a noun shou ld look l ike a leaf.
It is not even claimed that the relation between leaf and stem is the
same as the relation between noun and verb. What is claimed is,
first, that in both anato my and grammar the parts are to be classified
162

according to the relations between them. In bo th fields, the relations
are to be thought of as somehow primary, the relata as secondar y.
Beyond this, it is claimed that the relations are of the sort gener ated
by proces ses o f information ex -change .
In other word s, the mysteriou s and polymorphic relation between
context and content obtains in both anato my and linguis tics; and
evolutionis ts of the nineteenth centur y, preoccupied with what were
called "homologies," were, in fact, studying precis ely the contex tual
structures of biological deve lopment.
All of this specula tion beco mes almost platitude when we realize
that both grammar and biologica l structure are products of
communica tiona l and organizational process . The anatomy of the
plant is a complex transform of genot ypic instructions , and the
"language" of the genes , like any other language , must of neces sity
have contextua l structure. More-ove r, in all communication , there
must be a relevance between the contex tual structure of the message
and some structuring of the recipien t. The tissues of the plant could
not "read" the geno typic instructions carried in the chromosomes of
every cell unless cell and tissue exist, at that given moment, in a
contextu al structure .
What has been said above will serve as sufficient definition of
what is here meant by "form and pattern." The focus of discuss ion
was upon form rather than content, upon context rather than upon
what occu rs "in" the given con-te xt, upon relation ship rather than
upon the related per-sons or pheno mena.
The essa ys included range fro m a discussion of "sch ismogenesis"
(1935) t o two essays written after the birth of cybernet ics.
In 1935 , I certainly had not clearly grasped the central
importance of "contex t." I though t that the proce sses of schis-
mogenesis were important and nontrivial because in them I seemed
to see evolution at work: if interaction between persons could
unde rgo progress ive qualitative change as in-tensity increas ed, then
surely this could be the very stuff of cultural evolution . It followed
that all directiona l change , even in biolog ical evolut ion and
phylogen y, might—or must —be due to progre ssive interaction
between organisms. Under natural selection, such change in
relationsh ips would favor progressive change in anatomy and
physiolog y.
163

The progressive increase in size and armament of the dinosaurs
was, as I saw it, simply an interactive armaments race—a
schismogenic process. But I could not then see that the evolution
of the horse from Eohip pus was not a one-sided adjustment to
life on grassy plains. Surely the grassy plains themselves were
evolved pari passe with the evolution of the teeth and hooves of
the horses and other ungulates. Turf was the evolving response of
the vegetation to the evolution of the horse. It is the context
which evolves.
The classification of schismogenic process into "symmetrical"
and "complementary" was already a classification of con-texts of
behavior; and, already in this essay, there is a proposal to exmine
the possible combinations of themes in complementary behavior.
By 1942, I had completely for-gotten this old proposal, but I
attempted to do precisely what I had proposed seven years
previously. In 1942 many of us were interested in "national
character" and the con-, trast between England and America
fortunately brought into focus the fact that "spectatorship" is in
England a filial characteristic, linked with dependency and
submission, while in America spectatorship is a parental
characteristic linked with dominance and succoring.
This hypothesis, which I called "end-linkage," marked a turning
point in my thinking. From that time on, I have consciously
focused upon the qualitative structure of con-texts rather than
upon intensity of interaction. Above all, the phenomena of end-
linkage showed that contextual structures could themselves be
messages—an important point which is not made in the 1942
article. An Englishman when he is applauding another is
indicating or signaling potential submission and/or dependency;
when he shows off or demands spectatorship, he is signaling.
dominance or superiority; and so on. Every Englishman who
writes a book must be guilty of this. For the American, the
converse must hold. His boasting is but a bid for quasiparental ap-
proval.
The notion of context reappears in the essay "Style, Grace, and
Information in Primitive Art," but here the idea of context has
evolved to meet the related ideas of "redundancy," "pattern," and
"meaning."
164

Part III: Form and Pathol ogy in
Relationship

Soci al Planning and
the Conc ept of
Deutero-Learning*
Let me take as focus for this comment the last item1 in Dr.
Mead' s summary of her pape r. To the layman who has not occupied
himself with the comparat ive study of human cultures, this
recommendation may appe ar strange ; it may appea r to be an ethical
or philo sophica l paradox, a suggest ion that we discard purpos e in
* This- article was my comment on Margaret Mead 's article "The Com parative
Study of Cultur e and the Pu rposive Culti vation o f Democratic Values," published as
Chapter IV of Science , Philosophy and Religion, Second Symposium, copyright
1942 by the Conference on Science, Philosoph y and Religion, New York. It is here
reprinted b y permission of the Conference and of Harper & Ro w, Inc.
I have italicized a parenthesis in footnote 5 which pre-figures the concept of the
"double bind."1 Dr. Mead writes: ". . those students who have de-voted themsel ves to studying
cultures as wholes, as systems of dynamic equilibriu m, can make the follo wing
contributions : . . .
"4. Implement plans for alterin g our present culture by recognizin g the
importance o f includin g the social scientist within his e xperim ental m aterial , and b y
recognizing that by working toward defined ends we commit oursel ves to the
manipulation of persons, and therefore to the negation of democrac y. Only by
working in terms of values which are limited to definin g a direction is it possible
for us to use scienti fic methods in the control of the process without the negation of
the moral autono my of the human spirit. " (Italics hers.)
166

order to achiev e our purpose; it may even call to mind some of the
basic aphorisms of Christianity and Taoism. Such aphorisms are
familiar enough; but the layman will be a little surprised to find
them coming from a scientist and dressed in all the paraphernalia
of analytic thought. To other anthropologists and social scientists,
Dr. Mead's recommendations will be even more surprising, and
perhaps more meaningless, be-cause instrumentality and
"blueprints" are an essential ingredient in the whole structure of
life as science sees it. Likewise, to those in political life, Dr.
Mead's recommendation will be strange, since they see decisions
as classifiable into policy-making decisions versus executive
decisions. The governors and the scientists alike (not to mention
the commercial world) see human affairs as patterned upon pur-
pose, means and ends, connation and satisfaction.
If anybody doubts that we tend to regard purpose and
instrumentality as distinctively human, let him consider the old
quip about eating and living. The creature who "eats to live" is the
highest human; he who "lives to eat" is coarser-grained, but still
human; but if he just "eats and lives," without attributing
instrumentality or a spurious priority in time sequence to either
process, he is rated only among the animals, and some, less kind,
will regard him as vegetable.
Dr. Mead's contribution consists in this—that she, fortified by
comparative study of other cultures, has been able to
transcend the habits of thought current in her own cul-
ture and has been able to say virtually this: "Before we apply
social science to our own national affairs, we must re-examine and
change our habits of thought on the subject of means and ends. We
have learnt, in our cultural setting, to classify behavior into
`means' and `ends' and if we go on . defining ends as separate from
means and apply the social sciences as crudely instrumental
means, using the recipes of science to manipulate people, we shall
arrive at a totalitarian rather than a democratic system of life." The
solution which she offers is that we look for the "direction," and "values" implicit in the means, rather than looking ahead to a
blueprinted goal and thinking of this goal as justifying or not
justifying manipulative means. We have to find the value of a
planned act implicit in and simultaneous with the act itself, not
167

separate from it in the sense that the act would derive its value
from reference to a future end or goal. Dr. Mead's paper is, in fact,
not a direct preachment about ends and means; she does not say
that ends either do or do not justify the means. She is talking not
directly about ends and means, but about the way we tend to think
about ways and means, and about the dangers inherent in our habit
of thought.
It is specifically at this level that the anthropologist has most to
contribute to our problems. It is his task to see the highest common
factor implicit in a vast variety of human phenomena, or inversely,
to decide whether phenomena which appear to be similar are not
intrinsically different. He may go to one South Sea community,
such as the Manus, and there find that though everything that the
natives do is concretely different from our own behavior, yet the
underlying system of motives is rather closely comparable with
our own love of caution and wealth accumulation; or again he may
go to another society such as Bali and there find that, while the
outward appearance of the native religion is closely comparable
with our own—kneeling to pray, incense, intoned utterances
punctuated by a bell, etc.—the basic emotional attitudes are
fundamentally different. In Balinese religion we find an approval
accorded to rote, nonemotional performance of certain acts instead
of the insistence upon correct emotional attitude, characteristic of
Christian churches.
In every case the anthropologist is concerned not with mere
description but with a slightly higher degree of abstraction, a wider
degree of generalization. His first task is the meticulous collection
of masses of concrete observations of native life—but the next step
is not a mere summarizing of these data; it is rather to interpret the
data in an abstract language which shall transcend and comprehend
the vocabulary and notions explicit and implicit in our own
culture. It is not possible to give a scientific description of a native
culture in English words; the anthropologist must devise a more
abstract vocabulary in terms of which both our own and the native
culture can be described.
This then is the type of discipline which has enabl ed Dr. Mead to
point out that a discrep ancy—a basic and fundamental discrepanc y
—exis ts between "soci al enginee ring," manipulat ing peop le in order
168

to achieve a plann ed blue-prin t society, and the ideals of democracy,
the "supr eme worth and moral responsib ility of the individua l
human per-son." The two confl icting motifs have long been implicit
in our cu lture, science ha s had ins trumental lean ings s ince befo re the
Industr ial Revolu tion, and emphasis upon individual worth and
responsib ility is even older. The threat of conflict between the two
motifs has only come recently, with increas ing consciousn ess of,
and emphasis upon, the democratic motif and simultaneou s spread
of the instrumental motif. Finally, the conflict is now a life-or-death
strugg le over the role which the social scienc es shall play in the
ordering of human relationsh ips. It is hardly an exaggeration to say
that this war is ideologic ally about just this— the role of the social
science s. Are we to reserve the techniques and the right to
manipulate peop le as the privilege of a few planning , goal-oriented,
and powe r-hungr y individua ls, to whom the instrumental ity of
science makes a natural appe al? Now that we have the technique s,
are we, in cold blood, going to treat people as things ? Or what are
we go ing to do with these technique s?
The problem is one of very great difficulty as well as urgency,
and it is doubly difficult because we, as scientists, are deeply
soaked in habits of instrumental thought those of us, at least, for
whom science is a part of life, as well as a beautiful and dignified
abstraction. Let us try to surmount this additional source of
difficulty by turning the tools of science upon this habit of
instrumental thought and upon the new habit which Dr. Mead
envisages—the habit which looks for "direction" and "value" in the
chosen act, rather than in defined goals. Clearly, both of these
habits are ways of looking at time sequences. In the old jargon of
psychology, they represent different ways of apperceiving
sequences of behavior, or in the newer jargon of gestalt
psychology, they might both be described as habits of looking for
one or another sort of contextual frame for behavior. The problem
which Dr. Mead—who advocates a change in such habits—raises
is the problem of how habits of this abstract order are learned.
This is not the simple type of question which is posed in most
psychologic al laboratories, "Under what circumstances will a dog
learn to salivate in response to a bell?" or, "What variables gove rn
success in rote learning?" Our question is one degree more abstract,
169

and, in a sense, bridges the gap between the expe rimental work on
simple learning and the approach of the gestalt psychologis ts. We
are asking, "How does the dog acquir e a habit of punc tuating or ap-
perceiving the infinitely complex stream of events (including his
own behavior ) so that this stream appear s to be made up of one type
of short sequen ces rather than an-othe r?" Or, substituting the
scientist for the dog, we might ask, "What circumstances determine
that a given scientist will punc tuate the stream of even ts so as to
conclud e that all is predete rmined, while another will see the stream
of even ts as so regula r as to be suscep tible of control?" Or, again , on
the same level of abstraction let us ask—and this question is very
relevant to the promotion of democrac y—"What circumstances
promote that specific habitu al phrasing of the universe which we
call `free will' and those other s which we call `responsib ility,'
`cons tructivene ss,' `energy,' `passiv ity,' `dominance,' and the rest?"
For all these abstract qualities, the essential stock-in -trade of the
educa tors, can be seen as various habits of punctuating the stream of
experience so that it takes on one or anothe r sort of coherence and
sense . They are abstractions which begin to assume some
opera tiona l meaning when we see them take their place on a
concep tual level between the statements of simple learning and
those of gestalt psycholog y.
We can, for example, put our finger very simply on the process
which leads to traged y and disillusion whenever men decide that the
"end justifies the means" in their efforts to achi eve either a Christian
or a bluepr inted heaven-on -earth. They ignore the fact that in social
manipula tion, the tools are not hammers and screwdr ivers . A
screwdriver is not serious ly affected when, in an emergenc y, we use
it as a wedge; and a ha mmer's outlook on life is not af fected becau se
we sometimes use its hand le as a simple lever. But in social
manipula tion our tools are peop le, and people learn, and they
acqui re habits which are more subtle and pervasive than the tricks
which the bluepr inter teaches them. With the best intentions in the
world , he may train children to spy upon their parents in order to
eradicate some tendenc y prejudicial to the success of his blueprint,
but because the children are people they will do more than learn
this simple trick—they will build this experience into their whole
philosophy of life; it will color all their future attitudes to-ward
170

authority. Whenever they meet certain sorts of con-text, they will
tend to see these contexts as structured on an earlier familiar
pattern. The blueprinter may derive an initial advantage from the
children's tricks; but the ultimate success of his blueprint may be
destroyed by the habits of mind which were learned with the trick.
(Unfortunately, there is no reason to believe that the Nazi blueprint
will break down for these reasons. It is probable that the un-
pleasant attitudes here referred to are envisaged as basic both to
the plan itself and to the means of achieving it. The road to hell
can also be paved with bad intentions, though well-intentioned
people find this hard to believe.)
We are dealing, apparently, with a sort of habit which is a by-
product of the learning process. When Dr. Mead tells us that we
should leave off thinking in terms of blue-prints and should instead
evaluate our planned acts in terms of their immediate implicit
value, she is saying that in the upbringing and education of
children, we ought to try to inculcate a sort of by-product habit
rather different from that which we acquired and which we daily
reinforce in ourselves in our contacts with science, politics,
newspapers, and so on.
She states perfectly clearly that this new shift in the emphasis
or gestalt of our thinking will be a setting forth into uncharted
waters. We cannot know what manner of human beings will result
from such a course, nor can we be sure that we ourselves would
feel at home in the world of 1980. Dr. Mead can only tell us that if
we proceed on the course which would seem most natural,
planning our applications of social science as a means of attaining
a defined goal, we shall surely hit a rock. She has charted the rock
for us, and advises that we embark on a course in a direction where
the rock is not; but in a new, still uncharted direction. Her paper
raises the question of how we are to chart this new direction.
Actu ally, scienc e can give us- something approaching a chart. I
indica ted above that we might see a mixed bunch of abstract terms
—free will, predes tination, respons ibility, constructivene ss,
passivity, dominance , etc.—as all of them descriptive of
apperceptive habits, habitual ways of looking at the stream of
events of which our own behavior is a part, and further that these
habits might all be, in some sense, by-products of the learning
171

process. Our next task, if we are to achieve some sort of chart, is
clearly to get something better than a random list of these possible
habits. We must reduce this list to a classification which shall show
how each of these habits is systematically related to the others.
We meet in common agreement that a sense of individual
autonomy, a habit of mind somehow related to what I have called
"free will," is an essential of democracy, but we are still not
perfectly clear as to how this autonomy should be defined
operationally. What, for example, is the relation between
"autonomy" and compulsive negativism? The gas stations which
refuse to conform to the curfew—are they or are they not showing
a fine democratic spirit? This sort of "negativism" is undoubtedly
of the same degree of abstraction as "free will" or "determinism";
like them it is an habitual way of apperceiving contexts, event
sequences and own behavior; but it is not clear whether this
negativism is a "subspecies" of individual autonomy; or is it rather
some entirely different habit? Similarly, we need to know how the
new habit of thought which Dr. Mead advocates is related to the
others.
Eviden tly our need is for something better than a rando m list of
these habits of mind. We need some systematic frame-work or
classification which shall show how each of th ese habits is related to
the others, and such a classification might provid e us with
something approach ing the chart we lack. Dr. Mead tells us to sail
into as yet unchar ted waters , adopting a new habit of though t; but if
we knew how this habit is related to others, we might be able to
judge of the bene fits and danger s, the possib le pitfalls of such a
cours e. Such a chart might provid e us with the answers to some of
the questions which Dr. Mead raises—as to how we are to judge of
the "direction" and value implici t in our planned a cts.
You must not expec t the social scientist to produce such a chart
or classification at a moment's notice, like a rabbit out of a hat, but I
think we can take a first step in this direction; we can sugges t some
of the basic themes—the cardinal point s, if you like—upon which
the final classification m ust be built.
We have noted that the sorts of habit with which we are
conce rned are, in some sense, by-product s of . the learning
proce sses, and it is therefo re natural that we look first to the
172

pheno mena of simple learning as likely to provid e us with a clue.
We are raising quest ions one degree more abstract than those chiefly
studied by the exper imental psycholog ists, but it is still to their
labora tories that we m ust look for our a nswers .
Now it so happens that in the psychologic al labora tories there is
a common pheno menon of a somewhat highe r degree of abstraction
or genera lity than those which the exper iments are planned to
elucid ate. It is a commonplace that the exper imental subjec t—
wheth er animal or man, beco mes a better subjec t after repeated
experiments. He not only learns to salivate at the appropri ate
moments, or to recite the appropriate nonsense syllables ; he also, in
some way, learn s to learn. He not only solves the probl ems set him
by the experimenter , where each solving is a piece of simple
learning; but, more than this, he becomes more and more skilled in
the so lving of proble ms.
In semigestalt or semianthropo morphic phraseo logy, we might
say that the subjec t is learning to orient himself to certain types of
contexts, or is acquiring "insight" into the contexts of proble m
solving . In the jargon of this pape r, we may say that the subject has
acquired a habit of looking for contexts and sequenc es of one type
rather than another, a habit of "punctua ting" the stream of even ts to
give repetitions o f a certain type of meaningful sequen ce.
The line of argument which we have followed has brought us to
a point at which statements about simple learning meet statements
abou t gestalt and contextu al structure , and we have reached the
hypothesis that "learning to learn" is a synonym for the acqu isition
of that class of abstract habits of though t with which this paper is
concerned ; that the states of mind which we call "free will,"
instrumental thinking, dominance, passivity, etc., are acqu ired by a
process which we may equat e with "learning to learn."
173

This hypothesis is to some extent new2 to psycholog ists as well
as to laymen, and therefore I must digress at this point to suppl y
techn ical readers with a more precise statement of my meaning. I
must demonstr ate at least my willingnes s to state this bridge
between simple learning and gesta lt in opera tiona l terms.
Let us coin two words, "proto-learning" and "deuterolearning," to
avoid the labor of defining operational ly all the other terms in the
field (transfe r of learning , genera lization, etc., etc.). Let us say that
there are two sorts of gradient discernibl e in all continued learning.
The gradien t at any poin t on a simple learning curve (e .g., a curve of
rote learning) we will say chiefly represents rate of proto -learning .
If, however , we inflict a series of similar learning exper iments on
the same subject, we shall find that in each successiv e expe riment
the subje ct has a somewhat steep er proto-learning gradien t, that he
learns somewhat more rapidly. This proges sive change in rate of
proto -learning we wil l call "deu tero-learning ."
From this point we can easily go on to represen t deute rolea rning
graph ically with a curve whose gradien t shall represent rate of
deute ro-learning. Such a repres entation might be obtained, for
example, by intersecting the series of protolearning curves at some
arbitarily chosen number of trials, and noting what propor tion of
succe ssful response s occur red in each exper iment at this point. The
curve of deute ro-learning would then be obtained by plotting these
numbers a gains t the serial nu mbers of the expe riments.3
2 Psycholo gical papers bearin g upon this problem of the relationship between
gestalt and simple learnin g are very numerous, if we include all who have worked
on the concepts of transfer of learning, generalization, irradiation, reaction
threshold (Hull), insight, and the like. Historicall y, one of the first to pose these
questions was Mr. Fran k (L. K. Frank, "The Problems of Learnin g," Psych. Re view,
1926, 33: 329–51; and Professor Maier has recentl y introduced a concept of
"direction " which is closel y related to the notion of "deutero -learning. " He says:
"direction … is the force which integrates memories in a particular manner without
being a memory itself. " (N. R. F. Maier , "The Behavior Mechanisms Concerned
with Problem Solving, "- Psych. Review, 1940, 47: 43–58 .) If for "force" we
substitute "habit," and for "memory" we substitute "experience of the stream of
events," the concept of deutero- learning can be seen as almost synonymous with
Professor Maier 's concept of "direction ."3 It will be noted that the operational definition of deutero- learnin g is
necessaril y somewhat easier than that of proto- learnin g. Actuall y, no simple
learnin g curve represents proto -learnin g alone. Even within the duration of the
174

In this defin ition of proto- and deutero-learning , one phras e
remains consp icuousl y vagu e, the phras e "a series of similar
experiments." For purposes of illustration, I imagined a series of
experiments in rote learning, each expe riment similar to the last,
except for the substitution of a new series of nonsense syllables in
single learning experiment we must suppose that some deutero-learning will occur,
and this will make the gradient at any point somewhat steeper than the hypothetical
gradient of "pure" proto- learnin g.
175

place of those already learned. In this example, the curve of
•deutero -learning represent ed in-creasing prof icien cy in the business
of rote learning , and, as an expe rimental fact, such increa se in rote
proficien cy can be demonstra ted.4
Apart from rote learning, it is much more difficult to de-fine
what we mean by saying that one learning contex t is "similar" to
anothe r, unles s we are content to refer the matter back to the
exper imental ists by saying that learning contex ts shall be consider ed
to be "similar" one to anothe r whenever it can be shown
exper imental ly that exper ience of learning in one contex t does, as a
matter of fact, promote speed of learning in anothe r, and asking the
exper imentalists to find out for us what sort of classification they
can build up by use of this criterion. We may hope that they will do
this; but we canno t hope for immediat e answer s to our questions ,
becaus e there are very serious difficulties in the way of such
exper imentat ion. Exper iments in simple learning are already
difficult enough to control and to per-form with critical exactness ,
and expe riments in deute rolearning are likely to prove almost
impossib le.
There is, howeve r, an alterna tive course open to us. When we
equat ed "learning to learn" with acquir ing apper ceptive habits, this
did not exclude the possibility that such habits might be acquired in
other ways. To suggest that the only method of acquiring one of
these habits is through repeated experience of learning contexts of
a given kind would be logically analogous to saying that the only
way to roast pig is by burning the house down. It is obvious that in
human education such habits are acquired in very various ways.
We are not concerned with a hypothetical isolated individual in
contact with an impersonal events stream, but rather with real
individuals who have complex emotional patterns of relationship
with other individuals. In such a real world, the individual will be
led to acquire or reject apperceptive habits by the very complex
phenomena of personal example, tone of voice, hostility, love, etc.
Many such habits, too, will be conveyed to him, not through his
own naked experience of the stream of events, for no human
4 C. Hull, Mathematico-Deductive Theory of Rote Learning, New Haven, Yale
Universit y Press, 1940 .
176

beings (not even scientists) are naked in this sense. The events
stream is mediated to them through language, art, technology, and
other cultural media which are structured at every point by
tramlines of apperceptive habit.
It therefore follows that the psychological laboratory is not the
only possible source of knowledge about these habits; we may turn
instead to the contrasting patterns implicit and explicit in the
various cultures of the world studied by the anthropologists. We
can amplify our list of these obscure habits by adding those which
have been developed in cultures other than our own.
Most profitably, I believe, we can combine the insights of the
experimental psychologists with those of the anthropologists,
taking the contexts of experimental learning in the laboratory and
asking of each what sort of apperceptive habit we should expect to
find associated with it; then looking around the world for human
cultures in which this habit has been developed. Inversely, we may
be able to get a more definite—more operational—definition of
such habits as "free will" if we ask about each, "What sort of
experimental learning context would we devise in order to
inculcate this habit?" "How would we rig the maze or problem-box
so that the anthropomorphic rat shall obtain a repeated and
reinforced impression of his own free will?"
The classification of contexts of experimental learning is as yet
very incomplete, but certain definite advances have been made.5 It
5 Various classifications have been devised for purposes of exposition. Here I
follow that of Hilgard and Marquis (E. R. Hilgard and D. G. Marquis,
Conditioning and Learning, New York, Appleton Century Co., 1940). These
authors subject their own classification to a brilliant critical analysis, and to
this analysis I am indebted for one of the formative ideas upon which this
paper is based. They insist that any learning context can be described in
terms of any theory of learning, if we are willing to stretch and
overemphasize certain aspects of the context to fit onto the Procrustean bed
of the theory. I have taken this notion as a cornerstone of my thinking, sub-
stituting "apperceptive habits" for "theories of learning," and arguing that
almost any sequence of events can be stretched and warped and punctuated
to fit in with any type of apperceptive habit. (We may suppose that ex-
perimental neurosis is what happens when the subject fails to achiev e this
assimilation .)
I am also indebted to Lewin's topological analysis of the contexts of
reward and punishment. (K. Lewin, A Dynamic Theory of Personality , New
177

is possible to classify the principal contexts of positive
learning (as distinct from negative learning or inhibition,
learning not to do things) under four heads, as follows:
(1)Classical Pavlovian contexts
These are characterized by a rigid time sequence in which the
condi tioned stimulus (e.g., buzzer) alwa ys pre-cedes the
uncondi tioned stimulus (e.g., meat powder) by a fixed interva l of
time. This rigid sequenc e of even ts is not altered by anything that
the animal may do. In these con-texts, the animal learns to respond
to the condi tioned stimulus with behavior (e.g., salivation ) which
was fo rmerly evoked on ly by the un condit ioned stimulus.
(2)Contexts of instrumental reward or escape
These are characterized by a sequence which depends upon the
animal's behavior . The uncondi tioned stimulus in these contexts is
usual ly vagu e (e.g., the whole sum of circumstances in which the
animal is put, the proble m-box) and may be internal to the animal
(e.g., hunger ). If and when, unde r these circumstances , the animal
perfo rms some act within its behavior al repertoire and previous ly
selected by the experimenter (e.g., lifts its leg), it is immediately
rewarded.
(3)Contexts of instrumental avoidance
These are also characterized by a cond itional sequence . The
uncondi tioned stimulus is usually definite (e.g., a warning buzz er)
and this is followed by an unpleasan t exper ience (e.g., electric
shock) unless in the interval the animal per-forms some selected
act (e.g., lifts leg).
(4)Contexts of serial and rote learning
These are characterized by the predominant condi tioned stimulus
being an act of the subjec t. He learns , for example, alwa ys to give
the cond itioned response (nons ense syllable B) after he has himself
uttered the condi tioned stimulus ( nonsense syllable A),.
York, McGraw-Hill Book Co., 1936.)
178

This small beginning of a classification6 will be sufficient to
illustrate the principles with which we are conce rned and we can
now go on to ask about the occur rence of the appropria te
appercep tive habits among men of variou s cultures. Of greatest
interest—be cause least familiar—a re the Pavlovian pat-terns and the
patterns o f rote. It is a little hard for members of Western civilization
to believe that whole systems of behav ior can be built on premises
other than our own mixture of instrumental reward and instru mental
avoidance . The Trobriand Islanders , however , appe ar to live a life
whose coheren ce and sense is based upon looking at even ts through
Pavlovian spectac les, only slightly tinted with the hope of
instrumental reward, while the life of the Baline se is sensib le if we
6 Many people feel that the conte xts of experim ental learning are so
oversi mplified as to have no bearin g upon the pheno mena of the real world.
Actuall y, expansion of this classi fication will give means of defining systematically
many hundreds of possible contexts of learning with their associated apperceptive
habits . The scheme ma y be expanded in the follo wing ways:
a.Inclusion of contexts of negative learning (inhibition).
b.Inclusion of mixed types (e.g., cases in which salivation, with its
physiological relevance to meat powder, is also instrumental in obtaining
the meat powder).
c.Inclusion of the cases in which the subject is able to deduce some sort of
relevance (other than the physiological) between some two or more
elements in the sequence. For this to be true, the subject must have
experience of contexts differing systematically one from another, e.g.,
contexts in which some type of change in one element is constantly
accompanied by a constant type of change in another element. These cases
can be spread out on a lattice of possibilities, according to which pair of
elements the subject sees as interrelated. There are only five elements (con-
ditioned stimulus, conditioned response, reward or punishment, and two
time intervals), but any pair of these may be interrelated, and of the
interrelated pair, either may be seen by the subject as determining the other.
These possibilities, multiplied for our four basic contexts, give forty-eight
types.
d.The list of basic types may be extended by includin g those cases (not as yet
investi gated in learnin g experiments but common in interpersonal relationships)
in which the roles of subject and experim enter are reversed. In these, the
learning partner pro vides the initial and f inal elements, w hile some other person
(or circumstance) provides the middle term. In these types, we see the buzzer and
the meat powder as the behavior of a person and ask: "What does this person
learn ?" A great part of the gamut of apperceptive habits associated with
authorit y and parenthood is based on conte xts of this general type.
179

accep t premises based upon combining rote with instrumental
avoidan ce.
Clear ly, to the "pure " Pavlov ian, only a very limited fatal-ism
would be possible. He would see all even ts as preordained and he
would see himself as fated only to search for omens, not able to
influence the course of even ts—abl e, at most, from his reading of
the omens, to put himself in the proper ly receptive state, e.g., by
salivation, before the inevitable occu rred. Trobriand culture is not so
purely Pavlovian as this, but Dr. Lee,7 analyzing Profe ssor
Malinowski's rich observations , has shown that Trobriand phrasings
of purpose, cause, and effect are profoundl y differen t from our own;
and though Dr. Lee does not use the sort of classification here
proposed , it appea rs from Trobriand magic that these peopl e
continual ly exhibit a habit of thinking that to act as if a thing were
so will make it so. In this sense, we may describe them as semi-
Pavlov ians who have decided that "salivation" is instrumental to
obtaining "meat powde r." Malinowski , for example, gives us a
dramatic description of the almost physiolog ical extremes of rage8
which the Trobriand black magician practices in his incanta tions ,
and we may take this as an illustration of the semi-Pavlov ian frame
of mind in contrast with the very various types of magical procedure
in other parts of the world , where , for example, the efficacy of a
spell may be associa ted not with the intensity but with the extreme
rote accurac y of the recitation.
Among the Balines e9 we find another pattern which contrasts
sharp ly both with our own and with that of the Trobriande rs. The
7 Doroth y Lee, "A Primiti ve System of Values ," Journal Philos. of Science,
1940, 7: 355-78.8 A It is possible that se mi-Pavlov ian phrasings of the stream of events tend, like the
experiments which are their protot ypes, to hinge particularl y upon autonom ic reactions
—that those who see events in these terms tend to see these reactions, which are only
partiall y subject to voluntary control, as peculiarly effective and powerful causes of
outside events. There may be some ironic al logic in Pavlovian fatalism which
predisposes us to believe that we can alter the course of events only by means of those
beha viors w hich we are least able to control .9 The Balinese material collected by Dr. Mead and my-self has not yet been published in
extenso, but a brief out-line of the theory here suggested is available—cf. G. Bateson,
"The Frustration-Aggression Hypothesis and Culture ," Psychological Review, 1941,
48: 350-55.
180

treatment of children is such that they learn not to see life as
composed of connat ive sequences ending in satisfaction, but rather
to see it as composed of rote sequences inherently satisfying in
themselves— a pattern which is to some extent related to that pattern
which Dr. Mead has recommended, of looking for value in the act
itself rather than regarding the .act as a means to an end. There is,
how-eve r, one very important difference between the Balinese
pattern and that recommended by Dr. Mead. The Balinese pattern is
essent ially deriv ative from contex ts of instrumental avoidance ; they
see the world as dange rous, and themselves as avoiding, by the
endless rote behavior of ritual and courtesy, the ever-present risk of
faux pas. Their life is built upon fear, albeit that in gene ral they
enjoy fear. The positive value with which they endow their
immediate acts, not looking for a goal, is somehow associated with
this enjoyment of fear. It is the acroba t's enjoyment both of the thrill
and of his own virtuos ity in avo iding disaster.
We are now, after a somewhat long and technical excursion into
psychological laboratories and foreign cultures, in a position to
examine Dr. Mead's proposal in somewhat more concrete terms.
She advises that when we apply the social sciences we look for
"direction" and "value" in our very acts, rather than orient
ourselves to some blueprinted goal. She is not telling us that we
ought to be like the Balinese, except in our time orientation, and
she would be the first to disparage any suggestion that fear (even
enjoyed fear) should be our basis for assigning value to our acts.
Rather, as I understand it, this basis should be some sort of hope—
not looking to some far-off future, but still some sort of hope or
optimism. In fact, we might summarize the recommended attitude
by saying that it ought to be formally related to instrumental
reward, as the Balinese attitude is related to instrumental
avoidance.
Such an attitude is, I believe, feasible. The Balinese attitude
might be defined as a habit of rote sequences inspired by a thrilling
sense of ever-imminent but indefinite danger, and I think that what
Dr. Mead is urging us toward might be defined in like terms, as a
habit of rote sequences inspired by a thrilling sense of ever-
imminent but undefined reward.
181

As to the rote component, which is almost certainly a necessary
concomitant of the peculiar time orientation advocated by Dr.
Mead, 1, personally, would welcome it, and I believe that it would
be infinitely preferable to the compulsive type of accuracy after
which we strive. Anxious taking-care and automatic, rote caution
are alternative habits which perform the same function. We can
either have the habit of automatically looking before we cross the
street, or the habit of carefully remembering to look. Of the two I
prefer the automatic, and I think that, if Dr. Mead's rec-
ommendation implies as increase in rote automatism, we ought to
accept it. Already, indeed, our schools are inculcating more and
more automatism in such processes as reading, writing, arithmetic,
and languages.
As to the reward component , this, too, shou ld not be beyond our
reach. If the Balin ese is kept busy and happ y by a nameless,
shape less fear, not located in space or time, we might be kept on our
toes by a nameless , shape less, unloca ted hope of enormous
achievement. For such a hope to be effective, the achievement need
scarcely be de fined .
All we need to be sure of is that, at any moment, achievement
may be just around the corner, and, true or false, this can never be
tested. We have got to be like those few artists and scientists who
work with this urgent sort of inspiration, the urgency that comes
from feeling that great discovery, the answer to all our problems,
or great creation, the perfect sonnet, is always only just beyond our
reach, or like the mother of a child who feels that, provided she
pay constant enough attention, there is a real hope that her child
may be that infinitely rare phenomenon, a great and happy person.
182

A Theory o f Play and Fantasy*
This resear ch was planned and started with an hypothesis to
guide ou r invest igations, the task of the inv estigato rs be ing to collect
relevan t observational data and, in the proces s, to amplif y and
modify the hypothes is.
The hypothes is wi ll here be descr ibed a s it has g rown in
our thinking .
Earlier funda mental work of Whitehe ad, Russe ll,10 Witt-
genstein,11 Carnap ,12 Whorf,13 etc., as well as my own at-tempt14 to
use this earlier thinking as an ep istemologica l base
for psychiat ric theory, led to a series of gene ralizations :
(1) That human verbal communicat ion can operate and alwa ys
does operate at many contrasting levels of abstraction. These range
in two directions from the seemingl y simple denot ative level ("The
cat is on the mat"). One range or set of these more abstract levels
include s those explicit or implicit messages where the subje ct of
discours e is the languag e. We will call these metalingu istic (for
example, "The verbal sound `cat' stand s for any member of such and
such class of objects," or "The word , `cat,' has no fur and cannot
scratch"). The other set of levels of abstraction we will call
metaco mmunicative (e.g., "My telling you wher e to find the cat was
friend ly," or "This is play"). In these, the subjec t of discourse is the
relationsh ip be tween t he sp eakers .
* This essay was read (by Jay Haley) at the A.P.A. Regional Research
Conference in Mexico City, March 11, 1954. It is here reprinted from A.P.A.
Psychiatric Research Reports, II, 1955, by permission of the American
Psychiatric Association.10 A. N. Whitehead and B. Russell, Principia Mathematica, 3 vols., 2nd ed.,
Cambrid ge, Cambrid ge Universit y Press, 1910- 13.11 L. Wittgenstein, Tractatus Logico-Philosophicus , Lon-don, Harcourt Brace,
1922 .12 R. Carnap, The Logical Syntax of Language, New York, Harcourt Brac e,
1937 .13 B. L. Whor f, "Science and Linguistics," Technolog y Review, 1940 , 44: 229-
48.14 J. Ruesch and G. Bateson , Com munication: The Social Matrix of Psychiatr y,
New York, Norton , 1951 .
183

It will be noted that the vast majority of both metalingu istic and
metacommunicat ive messages remain implic it; and also that,
espec ially in the psychiatric interview, there occu rs a further class of
implicit messages abou t how metaco mmunicative messages of
friendship and hostility are to be interpreted .
(2) If we specula te about the evolu tion of communica tion, it is
eviden t that a very importan t stage in this evolution occu rs when the
organism gradual ly cease s to respond quite "automatically" to the
mood-signs of anothe r and beco mes able to recogniz e the sign as a
signa l: that is, to recogn ize that the other individua l's and its own
signa ls are only signals, which can be trusted , distrusted, falsified,
denied , amplified, corrected , and so forth.
Clear ly this realization that signals are signals is by no means
complete even among the human species. We all too often respond
automatica lly to newspape r head lines as though these stimuli were
direct objec t-indica tions of events in our environment instead of
signa ls concoct ed and transmitted by creatures as complexl y
motivated as ourselve s. The nonhu man mammal is automatica lly
excited by the sexual odor of anoth er; and rightly so, inasmuch as
the secretion of that sign is an "invo lunta ry" mood-sign; i.e., an
outward ly percep tible event which is a part of the physiologi cal
proce ss which we have called a mood. In the human species a more
complex state of affairs begins to be the rule. Deodorants mask the
involuntar y olfactory signs, and in their place the cosmetic indus try
provid es the individual with perfu mes which are not involun tary
signs but volun tary signa ls, recognizabl e as such. Many a man has
been thrown off balance by a whiff of perfu me, and if we are to
believe the adver tisers, it seems that these signals, volunta rily worn,
have sometimes an automatic and autosugges tive effect even upon
the volun tary wearer.
Be that as it may, this brief digre ssion will serve to illustrate a
stage of evolu tion— the drama precipitated when organisms, having
eaten of the fruit of the Tree of Knowledge , discov er that their
signa ls are signa ls. Not only the charac teristically human inven tion
of language can then follow, but also all the complexit ies of
empathy, identification, projection, and so on. And with these comes
the possibility of communicating at the multiplicity of levels of
abstraction m entioned above .
184

(3) The first definite step in the formulation of the hypothesi s
guiding this resear ch occur red in January, 1952 , when I went to the
Fleishha cker Zoo in San Francis co to look for behav ioral criteria
which would indicate whether any given organis m is or is not able
to recogn ize that the signs emitted by itself and other members of
the species are signal s. In theory, I had though t out what such
criteria might look like—tha t the occurr ence of metaco mmunicative
signs (or signa ls) in the stream of interac tion between the animals
would indica te that the animals have at least some awareness
(consciou s or unconsc ious) that the signs about which they
metaco mmunicate are signa ls.
I knew , of cours e, that there was no likelihood of finding
deno tative messages among nonhu man mammals, but I was still not
awar e that the animal data would require an almost total revision of
my thinking . What I encounte red at the zoo was a pheno menon well
known to everybody: I saw two young monkeys playing, i.e.,
engaged in an interactive sequen ce of which the unit actions or
signal s were similar to but not the same as those of combat. It was
evident , even to the human observer, that the sequence as a whole
was not combat, and eviden t to the human observer that to the
participan t monkeys this was "no t combat."
Now, this pheno menon, play, could only occur if the participan t
organis ms were capable of some degree of meta-co mmunication,
i.e., of exchanging signals which would carry the message "this is
play."
(4) The next step was the examination of the message "This is
play," and the realization that this message contains those elements
which necessa rily genera te a paradox of the Russ ellian or
Epimenides type -a nega tive statement containing an implic it
negative metastat ement. Expanded, the statement "This is play"
looks something like this: "These actions in which we now engage
do not denote what those action s for which they stand would
deno te."
We now ask about the italicized words , "for which they stand."
We say the word "ca t" stands for any member of a certain class. That
is, the phra se "st ands for" i s a nea r synonym of "d enotes ." If we now
substitute "which they denote" for the words "for which they stand"
in the expand ed definition of play, the result is: "These actions, in
185

which we now engage, do not denote what would be de-no ted by
those actions which these actions denote ." The playful nip denotes
the bite, but it does not denot e what would be denot ed by the bite.
According to the Theory of Logical Types such a message is of
cours e inad missable , because t he word "deno te" is b eing used i n two
degre es of abstraction, and these two uses are treated as
synonymous. But all that we learn from such a criticism is that it
would be bad natur al history to expe ct the mental proce sses and
communicative habits of mammals to confor m to the logician's
ideal. Indeed, if human though t and communicat ion always
confor med to the ideal, Russel l would not in fact could not
have formulated t he ideal.
(5) A related problem in the evolution of communicat ion
conce rns the origin of what Korz ybski15 has called the map-territory
relation: the fact that a message , of whateve r kind, does not consis t
of those objec ts which it deno tes ("The word `cat' cannot scratch
us"). Rathe r, language bears to the objects which it denotes a
relationship comparabl e to that which a map bears to a territory.
Denotat ive communicat ion as it occu rs at the human level is only
possib le after the evolu tion of a complex set of metalingui stic (but
not verbalized)16 rules which govern how words and sentenc es shall
be related to objects and events . It is therefo re approp riate to look
for the evolu tion of such metalingu istic and/or meta-co mmunicative
rules at a prehuman and p reverb al level.
It appea rs from what is said above that play is a pheno menon in
which the actions of "play" are related to, or denote , other actions of
"not play." We therefore meet in play with an instanc e of signals
stand ing for other even ts, and it appea rs, therefor e, that the
evolu tion of play may have been an important step in the evolution
of communication.
(6) Threat is another pheno menon which resembles play in that
actions deno te, but are different from, other actions . The clenched
fist of threat is different from the punch , but it refers to a possible
15 A. Korzybski, Science and Sanity, New York, Science Press, 1941.16 The verbalization of these metalinguistic rules is a much later
achievement which can only occur after the evolution of a
nonverbalized meta-metalinguistics.
186

future (but at present nonex isten t) punch. And threat also is
commonly recogn izabl e among non-human mammals. Indeed it has
lately been argued that a great part of what appea rs to be combat
among members of a single species is rather to be regarded as threat
(Tinbergen,17 Lorenz18).
(7) Histrionic behav ior and deceit are other examples of the
primitive occu rrence of map-territory differen tiation. And there is
evidence that dramatization occu rs among birds: a jackdaw may
imitate her own mood-signs (Lorenz19), and deceit has been
observed among howler monkeys (Carp enter20).
(8) We might expect threat, play, and histrion ics to be three
independen t pheno mena all contribu ting to the evolution of the
discrimination between map and territory. But it seems that this
would be wrong , at least so far as mammalian communica tion is
concerned . Very brief analysis of childhood behavio r shows that
such combinations as histrion ic play, bluff, playful threat, teasing
play in response to threat, histrionic threat, and so on form togeth er
a single total complex of pheno mena. And such adult phenomena as
gambling and playing with risk have their roots in the combinat ion
of threat and play. It is eviden t also that not only threat but the
reciprocal of threat— the behavio r of the threatened individual—a re
a part of this complex. It is probab le that not only histrionics but
also spectatorship should be included within this field. It is also
appropri ate to mention self-pity.
(9) A further exten sion of this thinking leads us to includ e ritual
within this gene ral field in which the discrimination is drawn, but
not complete ly, between denot ative action and that which is to be
deno ted. Anthropolog ical studies of peace-m aking ceremonies , to
cite only one ex ample, sup-po rt this conc lusion .
In the Anda man Islands, peace is conclud ed after each side has
been given ceremonial freedo m to strike the other. This example,
howeve r, also illustrates the labile nature of the frame "This is play,"
17 N. Tinber gen, Social Behavior in Animals with Special Reference to
Vertebrat es, London, Methuen, 1953.18 K. Z. Lorenz, King Solomon 's Ring, Ne w York, Crowell, 1952.19 Ibid.20 C. R. Carpenter, "A Field Study of the Behavior and Social Relations of
Howling Monkeys," Comp. Psychol. Mono gr., 1934, 10: 1-168.
187

or "This is ritual." The discrimination between map and territory is
alwa ys liable to break down, and the ritual blows of peace -making
are always liable to be mistaken for the "real" blows of combat. In
this event , the peace -making ceremony beco mes a battle (Radcl iffe-
Brown21) .
(10) But this leads us to recognit ion of a more complex form of
play; the game which is const ructed not upon the premise "This is
play" but rather around the quest ion "Is this play?" And this type of
interac tion also has its ritual forms, e.g., in the hazing of initiation.
(11) Paradox is doubl y present in the signa ls which are
exchanged within the context of play, fantas y, threat, etc. Not only
does the playful nip not denote what would be denoted by the bite
for which it stands, but, in addition, the bite itself is fictiona l. Not
only do the playing animals not quite mean what they are saying
but, also, they are usually communicat ing about something which
does not exist. At the human level, this leads to a vast variety of
complications and inver sions in the fields of play, fantasy, and art.
Conjure rs and pain ters of the trompe l'oeil school c oncentrate upon
acqui ring a virtuosi ty whose only reward is reached after the viewer
detects that he has been deceiv ed and is forced to smile or marvel at
the skill of the deceiv er. Hollywood film-makers spend millions of
dollars to increase the realism of a shadow . Other artists, perhaps
more realistically, insist that art be nonrepre senta tiona l; and poke r
players achieve a strange addictive realism by equa ting the chips for
which they play with dollars. They still insist, however , that the
loser accept his loss as part of the ga me.
Finally, in the dim region where art, magic, and r eligion meet and
overlap, human beings have evolved the "metaphor that is meant,"
the flag which men will die to save, and the sacrament that is felt to
be more than "an outward and visible sign, given unto us." Here we
can recogn ize an attempt to deny the difference between map and
territory, and to get back to the absolu te innocenc e of
communication by means of pure mood-signs.
(12) We face then two pecu liarities of play: (a) that the messages
or signa ls exchanged in play are in a certain sense untrue or not
21 A. R. Radcliffe-Brown, The Andaman Islanders, Cambridge, Cambridge
University Press, 1922.
188

meant; and (b) that that which is deno ted by these signals is
nonex istent. These two pecu liarities so metimes co mbine st rangel y to
a reverse a conc lusion reached abov e. It was stated (4) that the
playful nip de-no tes the bite, but does not deno te that which would
be denoted by the bite. But there are other instances where an
oppos ite pheno menon occurs. A man experiences the full intensity
of subjec tive t error when a spea r is flung at h im out of th e 3D screen
or when he falls head long from some peak created in his own mind
in the intens ity of nightmare. At the moment of terror there was no
question ing of "reality," but still there was no spear in the movie
hous e and no cliff in the bedroom. The images did not denote that
which they seemed to denote , but these same images did really
evok e that terror which would have been evoked by a real spea r or a
real precipice. By a similar trick of self-con tradiction, the film-
makers of Hollywood are free to offer to a puritanical public a vast
range of pseudo sexual fantasy which otherwi se would not be
tolerated . In David and Bathsheba, Bath sheba can be a Troilistic
link between David and Uriah . And in Hans Christian Ander sen, the
hero starts out accompanied by a boy. He tries to get a woman, but
when he is defea ted in this attempt, he return s to the boy. In all of
this, there is, of course, no homosexuali ty, but the choic e of these
symbolis ms is associated in these fantasies with certain
characteristic ideas, e.g., abou t the hopelessness of the heterosexual
masculine position when faced with certain sorts of women or with
certain sorts of male author ity. In sum, the pseudoho mosexua lity of
the fan tasy does no t stand f or an y real ho mosexual ity, but does s tand
for and express attitudes which might accompany a real
homosexua lity or feed its etiologica l roots. The symbols do not
deno te homosexual ity, but do deno te ideas for which homosexuali ty
is an approp riate symbol. Evident ly it is necessa ry to re-examine the
precis e semantic validity of the interpre tations which t he ps ychiat rist
offers to a patient, and, as preliminar y to this analysis, it will be
necessa ry to examine the nature of the frame in which these
interpret ation s are offered .
(13) What has previousl y been said about play can be used as an
introdu ctory example for the discu ssion of frames and contexts. In
sum, it is our hypothesis that the message "This is play" establishe s
189

a paradoxi cal frame comparabl e to Epimenides' paradox. This frame
may be diagra mmed thus:
The first statement within this frame is a self-con tradictory
propos ition abou t itself. If this first statement is true, then it must be
false. If it be false, then it must be true. But this first statement
carries with it all the other statements in the frame. So, if the first
statement be true, then all the othe rs must be false; and vice vers a, if
the first statement be un true then all the other s must be true.
(14) The logically minded wi ll notice a non-sequitur. It could be
urged that even if the first staement is false, there remains a logical
possib ility that some of the other statements in the frame are untrue.
It is, however , a characte ristic of unconsc ious or "primary-process"
thinking that the think er is unable to discriminate between "some"
and "all," and unable to discriminate between "not all" and "none ." It
seems that the achiev ement of these discriminations is performed by
highe r or more consc ious mental proce sses which serve in the
nonps ychotic individua l to correct the black-a nd-white thinking of
the lower levels. We assume, and this seems to be an orthodox
assumption, that primary process is continual ly operating, and that
the psycholog ical validity of the paradoxical play frame depends
upon t his part of the mind.
(15) But, conve rsely, while it is necessary to invoke the primary
proce ss as an explan atory principle in order to delete the notion of
"some" from between "all" and "none ," this does not mean that play
is simply a primary-process pheno menon . The discrimination
between "play" and "nonp lay," like the discrimination between
fantasy and nonfan tasy, is certainly a function of seconda ry process,
or "ego." Within the dream the dreamer is usually unaware that he is
dreaming, and with in "play" he must often be reminded that "This is
play."
190

Similar ly, within dream or fantasy the dreamer does not opera te
with the concept "untru e." He operates with all sorts of statements
but with a curious inability to achieve meta-statements. He cannot ,
unless close to waking , dream a statement referring to (i.e., framing)
his dream.
It therefo re follows that the play frame as here used as an
explanatory principle implies a special combination of primary and
secondar y proces ses. This, howeve r, is related to what was said
earlier, when it was argued that play marks a step forward in the
evolution of communicat ion—the crucial step in the discover y of
map-territory relations. In primary process, map and territory are
equated; in seconda ry proces s, they can be discriminated . In play,
they are bo th equ ated and d iscriminated .
(16) Anoth er logical anomaly in this system must be mentioned :
that the relationship between two proposi tions which is commonly
describ ed by the word "premise" h as be come intrans itive. In genera l,
all asymmetrical relation ships are transitive. The relationship
"greater than" is typical in this respect ; it is convent ional to argue
that if A is greate r than B, and B is greater than C, then A is greate r
than C. But in psychologi cal processe s the transitivity of
asymmetrical relations is not observed . The propos ition P may be a
premise for Q; Q may be a premise for R; and R may be a premise
for P. Spec ifically, in the system which we a re conside ring, the circle
is still more contracted . The message , "All statements within this
frame are untrue" is itself to be taken as a premise in evalu ating its
own truth or untruth. (Cf. the in-transitivity of psychologic al
prefer ence discussed by McCul loch.22 The paradig m for all
paradoxes of this genera l type is Russel l's23 "class of classes which
are not members of them-selves." Here Russel l demonstrates that
paradox is genera ted by treating the relationsh ip, "is a member of,"
as an in-transitive.) With this caveat , that the "premise" relation in
psycholog y is likely to be intransi tive, we shall use the word
"premise" to denote a dependenc y of one idea or message upon
another comparable to the dependenc y of one proposition upon
22 W. S. McCulloch , "A Heterarch y of Values, etc.," Bulletin of Math.
Biophys., 1945, 7: 89-93.23 Whiteh ead and Russell, op. cit.
191

anothe r which is referred to in logic by saying that the propo sition P
is a premise for Q.
(17) All this, howeve r, leave s uncle ar what is meant by "frame"
and the related notion of "context." To clarify these, it is necessary
to insist first that these are psychologica l concepts. We use two sorts
of analogy to discuss these notions: the physical analogy of the
picture frame and the more abstract, but still not psychologi cal,
analog y of the mathematical set. In set theory the mathematici ans
have developed axioms and theore ms to discus s with rigor the
logical implications of membership in overlapping categori es or
"sets." The relationsh ips between sets are commonly illustrated by
diagrams in which the items or members of a larger univers e are
represent ed by dots, and the smaller sets are delimited by imaginar y
lines enclos ing the members of each set. Such diagrams then
illustrate a topologica l approach to the logic of classification. The
first step in defining a psychologica l frame might be to say that it is
(or delimits) a class or set of messages (or meaningfu l actions ). The
play of two individuals on a certain occasion would then be defined
as the set of al l messages exch anged by them within a limited period
of time and modified by the paradox ical premise system which we
have descr ibed. In a set-theor etical diagram these messages m ight be
represent ed by dots, and the "set" enclosed by a line which would
separate these from other dots represen ting nonpl ay messages. The
mathematical analogy break s down, however , becau se the
psycholog ical frame is not satisfactorily represent ed by an
imagina ry line. We assume that the psychologica l frame has some
degre e of real existence . In many instance s, the frame is consc iousl y
recogn ized and even represent ed in vocabu lary ("play," "movie,"
"interview," "job," "language ," etc.). In other cases, there may be no
explicit verbal referenc e to the frame, and the subject may have no
consc iousness of it. The analyst, however , finds that his own
thinking is simplified if he uses the notion of an unconsciou s frame
as an explanatory principle; usually he goes further than this and
infers its existen ce in the sub ject's uncon scious .
But while the analogy of the mathematica l set is perhaps over
abstract, the analog y of the picture frame is excess ively concr ete.
The psycholog ical concept which we are trying to define is neither
physical nor l ogica l. Ra ther, the actua l physical f rame is, we believe,
192

added by human beings to physical pictures because these human
beings operate more easily in a universe in which some of their
psychologic al characteristics are externalized. It is these
characteristics which. we are tr ying to discuss , using the exte rnaliza-
tion as an illustrative device.
(18) The common functions and uses of psychological frames
may now be listed and illustrated by reference to the analogies
whose limitations have been indicated in the previous paragraph:
(a)Psychologica l frames are exclusiv e, i.e., by including certain
messages (or meaningful actions) within a frame, certain other
messages a re excluded .
(b)Psycholog ical frames are inclusive , i.e., by exclud ing certain
messages certain others are included . From the point of view of set
theory these two functions are synonymous, but from the point of
view of psycholog y it is nece ssary to list them separately. The frame
around a picture, if we consider this frame as a message intended to
order or organiz e the percept ion of the viewe r, says, "Attend to what
is within and do not attend t o what is outside." Figure and g round, as
these terms are used by gesta lt psychologis ts, are not symmetrically
related as are the set and nonset of set theory. Percep tion of the
ground must be positively inhib ited and percep tion of the figure (in
this case the picture ) must be positively enhanc ed.
(c)Psychologica l frames are related to what we have called
"premises." The picture frame tells the viewe r that he is not to use
the sa me sort of th inking i n interpre ting t he pic ture t hat h e might use
in interpre ting the wallp aper outside the frame. Or, in terms of the
analogy from set theory, the messages enc losed within the imaginar y
line are defined as members of a class by virtue of their sharing
common premises or mutual relevan ce. The frame itself thus
becomes a part of the premise system. Either, as in the case of the
play frame, the frame is involved in the evaluation of the messages
which it contains, or the frame merely assists the mind in
unde rstand ing the contained messages by reminding the think er that
these messages are mutually relevant and the messages outside the
frame may be ignored.
(d)In the sense of the previous paragraph , a frame is
metaco mmunicative. Any message, which either explicitly or
implic itly defines a frame, ipso facto gives the receive r instructions
193

or aids in his attempt to unde rstand th e messages i ncluded within the
frame.
(e)The conver se of (d) is also true. Ever y meta-co mmunicative or
metalinguis tic m essage def ines, ei ther expl icitly or implic itly, the set
of messages about which it communicat es, i.e., every
metacommunicat ive message is or de-fine s a psycholog ical frame.
This, for example, is very eviden t in regard to such small
metacommunicat ive signals as punctuation marks in a printed
message , but applies equa lly to such complex metacommunicat ive
messages as the psychiatrist's definition of his own curative role in
terms of which his contribution s to the whole mass of messages in
psychothe rapy are to be under stood.
(f)The relation between psychologica l frame and perceptu al
gestalt needs to be conside red, and here the analogy of the picture
frame is usefu l. In a painting by Roual t or Blake , the human figures
and other objec ts represen ted are outlined. "Wise men see outlines
and therefore they draw them." But outside these lines, which
delimit the percep tual gestalt or "figure ," there is a background or
"ground" which in turn is limited by the picture frame. Similarly, in
set-theo retical diagrams, the larger unive rse within which the
smaller sets are drawn is itself enclosed in a frame. This doub le
framing is, we believe, not merely a matter of "frames within
frames" but an indication that mental proces ses resemble logic in
needing an outer frame to delimit the ground agains t which the
figures are to be perceived. This need is often unsatisfied, as when
we see a piece of sculptu re in a junk shop window , but this is
unco mfortable. We sugges t that the need for this outer limit to the
ground is related to a preferenc e for avoiding the paradox es of
abstraction. When a logical class or set of items is defined—fo r
example, the class of matchboxes— it is necess ary to delimit the set
of items which are to be excluded, in this case, all those things
which are not matchboxes. But the items to be included in the
background s et must be o f the s ame degree of ab straction, i.e., of t he
same "logica l type" as those within the set itself. Spec ifically, if
paradox i s to b e avoided , the " class of m atchboxes" a nd the "c lass o f
nonmatchboxes" (even though both these items are clearly not
matchboxes ) must not be regarded as members of the class of
nonmatchboxes . No class can be a member, of itself. The picture
194

frame then, becaus e it delimits a backg round, is here regard ed as an
externa l representa tion of a very special and importan t type of
psycho-log ical frame—na mely a frame whose function is to delimit
a log ical t ype. This, in fac t, is what was ind icated above when it was
said that the picture frame is an instruction to the viewer that he
shou ld not extend the premises which obtain between the figure s
within the picture to the., wall paper behind it.
But, it is precisely this sort of frame that precipitates paradox.
The rule for avoid ing paradox es insists that the items outside any
enclosing line be of the same logical type as those within , but the
picture frame, as analyzed above , is a line dividing items of one
logica l type from those of anoth er. In passing, it is interesting to not e
that Russel l's rule canno t be stated withou t breaking the rule. Russel l
insists that al l items of inapprop riate logica l type be ex luded (i.e., by
an imaginary line) from the background of any class, i.e., he insists
upon the drawing of an imagina ry line of prec isely the sort which he
prohibi ts.
(19) This whole matter of frames and paradox es may be
illustrated in terms of animal behavio r, where three types of message
may be recognized or deduced : (a) Message s of the sort which we
here call mood-sign s; (b) messages which simulate mood-signs (in
play, threat, histrioni cs, etc.) ; and (c) messages which enable the
receiv er to discriminate between mood-signs and those other signs
which resemble them. The message "This is play" is of this third
type. It tells the receive r that certain nips and other meaningful
action s are no t messages of the first type.
The message "This is play" thus sets a frame of the sort which is
likely to precipitate paradox : it is an attempt to discriminate
between, or to draw a line between, categor ies of different logical
types.
(20) This discuss ion of play and ps ycholog ical fr ames establishes
a type of triadic constellation (or system of relationships ) between
messages. One instance of this constellation is analyzed in
paragraph 19, but it is evident that conste llations of this sort occur
not only at the nonhu man level but also in the much more complex
communica tion of human beings . A fantasy or myth may simulate a
deno tative narrative, and, to discriminate between these types of
discours e, people use messages of the frame-setting type, and s o on.
195

(21) In conclu sion, we arrive at the complex task of applying this
theoretical appro ach to the particula r pheno mena of psychother apy.
Here the lines of our thinking may most briefly be summarized by
presenting and partially answer ing these questions :
(a)Is there any indica tion that certain forms of psycho-pa tholog y
are specifically characterized by abnormalities in the patient's
handl ing of frames and paradox es?
(b)Is there any indication that the techn iques of psycho-the rapy
neces sarily depend upon the manipula tion of frames and pa radoxes?
(c)Is it possible to describe the proce ss of a given psychotherapy
in terms of the interaction between the patient's abno rmal use of
frames and t he therapist's manipulation of them?
(22) In reply to the first question, it seems that the "word salad"
of schizophren ia can be described in terms of ,the patient's failure to
recogn ize the metaphor ic nature of his fantasies. In what shou ld be
triadic const ellation s of messages ., the frame-set ting message (e.g.,
the ph rase "a s if") is o mitted , and the m etaphor or f antas y is nar rated
and acted upon in a manner which would be approp riate if the
fantasy were a message of the more direct kind. The absenc e of
metacommunicat ive framing which was noted in the case of dreams
(15) is characteristic of the waking communications of the
schizophren ic. With the loss of the ability to set metaco mmunicative
frames, there is also a loss of ability to achiev e the more primary or
primitive message. The metaphor is treated directly as a message of
the more primary type. (This matter is discussed at greater length
in the paper given by Jay Haley at this Conference.)
(23) The dependence of psychotherapy upon the manipulation
of frames follows from the fact that therapy is an attempt to change
the patient's metacommunicative habits. Before therapy, the patient
thinks and operates in terms of a certain set of rules for the making
and understanding of messages. After successful therapy; he
operates in terms of a different set of such rules. (Rules of this sort
are in general, unverbalized, and unconscious both before and
after.) It follows that, in the process of therapy, there must have
been communication at a level meta to these rules. There must
have been communication about a change in rules.
But such a communication about change could not conceivably
occur in messages of the type permitted by the patient's
196

metacommunicative rules as they existed either be-fore or after
therapy.
It was suggested above that the paradoxes of play are
characteristic of an evolutionary step. Here we suggest that similar
paradoxes are a necessary ingredient in that process of change
which we call psychotherapy.
The resemblance between the process of therapy and the
phenomenon of play is, in fact, profound. Both occur within a
delimited psychological frame, a spatial and temporal bounding of
a set of interactive messages. In both play and therapy, the
messages have a special and peculiar relationship to a more
concrete or basic reality. Just as the pseudocombat of play is not
real combat, so also the pseudolove and pseudohate of therapy are
not real love and hate. The "transfer" is discriminated from real
love and hate by signals invoking the psychological frame; and
indeed it is this frame which permits the transfer to reach its full
intensity and to be discussed between patient and therapist.
The formal characte ristics of the therap eutic proces s may be
illustrated by building up a model in stages . Imagine first two
players who engage in a game of canasta accord ing to a standard set
of rules. So long as these rules govern and are unques tioned by both
players, the game is unchang ing, i.e:, no therapeut ic change will
occur. (Indeed many at-tempts at psychotherap y fail for this reason. )
We may imagine, however , that at a certain moment the two canas ta
players cease to play canasta and start a discussion of the rules.
Their discourse is now of a different logical type from that of their
play. At the end of this discussion, we can imagine that they return
to playing but with modified rules.
This sequence- of events is, however, still an imperfect model
of therapeutic interaction, though it illustrates our contention that
therapy necessarily involves a combination of discrepant logical
types of discourse. Our imaginary players avoided paradox by
separating their discussion of the rules from their play, and it is
precisely this separation that is impossible in psychotherapy. As we
see it, the process of psychotherapy is a framed interaction
between two persons, in which the rules are implicit but subject to
change. Such change can only be proposed by experimental action,
but every such experimental action, in which a proposal to change
197

the rules is implicit, is itself a part of the ongoing game. It is this
combination of logical types within the single meaningful act that
gives to therapy the character not of a rigid game like canasta but,
instead, that of an evolving system of interaction. The play of
kittens or otters has this character.
(24) In regard to the specific relation ship between the way in
which the patient handles frames and the way in which the therapist
manipula tes them, very little can at present be said. It is, howeve r,
sugges tive to observe that the psychologica l frame of therapy is an
analogu e of the frame-setting message which the schizophren ic is
unable to achieve. To talk in "word salad" within the psycholog ical
frame of therap y is, in a sense, not patho logica l. Indeed the neuro tic
is specifically encou raged to do precisely this, narrating his dreams
and free assoc iations so that patient and therapi st may achiev e an
unders tanding of this materia l. By the process of interpre tation, the
neuro tic is driven to insert an "as if" clause into the produc tions of
his primary process thinking, which product ions he had previousl y
depre cated or re-pressed . He must learn that fantasy contains truth.
For the schizophrenic the problem is somewhat different. His
error is in treating the metaphors of primary process with the full
intensity of literal truth. Through the discovery of what these
metaphors stand for he must discover that they are only metaphors.
(25) From the point of view of the project, howeve r,
psychothe rapy constitutes only one of the many fields which we are
attempting to inves tigate. Our central thesis may be summed up as a
statement of the necess ity of the paradoxes of abstraction. It is not
merely bad natural history to suggest that peop le might or should
obey the Theory of Logica l Types in their communicat ions; their
failure to do this is not due to mere carelessness or ignor ance.
Rathe r, we believe that the paradoxe s of abstraction must make their
appea rance in all communicat ion more complex than that of mood-
signa ls, and that without these paradoxes the evolu tion of
communication would be at an end. Life would then be an endle ss
interchang e of stylized messages , a game with rigid rules,
unrelieved by change o r humor.
198

Epidemiology o f a Schizo phreni a*
If we are to discu ss the epidemiology of mental cond itions, i.e.,
cond itions partly induced by exper ience , our first task is to pinpoin t
a defect of an ideational system sufficiently so that we can go on
from that pinpoin ting to postulate what sort of contex ts of learning
might indu ce this formal defect.
It is convent ional ly said that schizoph renics have "ego
weaknes s." I now defin e ego weakness as troub le in identifying and
interpret ing those signa ls which shou ld tell the individu al what sort
of a message a message is, i.e., trouble with the signals of the same
logica l type as the signal "This is play." For example, a patient
comes into the hosp ital can-te en and the girl behind the counte r
says, "What can I do for you?" The patient is in doubt as to what
sort of a message this is—is it a message about doing him in? Is it
an indication that she wants him to go to bed with her? Or is it an
offer of a cup of coffee? He hears the message and does not know
what sort or order of a message it is. He is unab le to pick up the
more abstract labels which we are most of us able to use
conventionally but are most of us unable to identify in the sense
that we don't know what told us what sort of a message it was. It is
as if we some-how make a correct guess. We are actually quite
unconscious of receiving these messages which tell us what sorts
of message we receive.
Difficulty with signals of this sort seems to be the center of a
syndrome which is characteristic for a group of schizophrenics, so
therefore we can reasonably look for an etiology starting from this
symptomatology as formally defined.
When you begin thinking in this way, a great deal of what the
schizophrenic says falls into place as a description of his
* This is an edited version of a talk, "How the Deviant Sees His Societ y," given in
May, 1955 , at a conference on "The Epid emiolo gy of Mental Health " held at
Brighton, Utah, sponsored by the Departments of Psychiatry and Psychology of the
University of Utah, and the Veterans Administration Hospital , Fort Douglas
Division, of Salt Lake City, Utah. A rough transcript of the talks at this conference
was mimeographed and circulated by the organizers.
199

experience. That is, we have a second lead toward the theory of
etiology or transmission. The first lead is from the symptom. We
ask, "How does a human individual acquire an imperfect ability to
discriminate these specific signals?" and when we look at his
speeches, we find that, in that peculiar language which is
schizophrenic salad, he is de-scribing a :traumatic situation which
involves a metacommunicative tangle.
A patient, for example, has a central notion, that "some-thing
moved in space," and that that is why he cracked up. I somehow,
from the way he spoke about "space," got an idea that space is his
mother and said so. He said, "No, space is the mother." I suggested
to him that she might be in some way a cause of his troubles. He
said, "I never condemned her." At a certain point he got angry, and
he said—this is verbatim—"If we say she had movement in her
because of what she caused, we are only condemning ourselves."
Something moved in space that made him crack up. Space is not
his mother, it is the mother. But now we focus upon his mother
whom he says he never condemned. And he now says, "If we say
that she had movement in her because of what she caused, we are
only condemning our-selves."
Look very carefully at the logical structure of that last
quotation. It is circular. It implies a way of interaction and chronic
cross-purposes with the mother such that for the child to make
those moves which might straighten out the misunderstanding was
also prohibited.
On another occasion he had skipped his therapy session in the
morning, and I went over to the dining hall at supper time to see
him and assure him that he would see me next day. He refused to
look at me. He looked away. I made some remark about 9.30 the
next morning—no answer. Then, with great difficulty, he said,
"The judge disapproves." Be-fore I left him, I said, "You need a
defense attorney," and when I found him on the grounds next
morning I said, "Here is your defense attorney," and we went into
session together. I started out by saying, "Am I right in supposing
that the judge not only disapproves of your talking to me but also
disapproves of your telling me that he disapproves?" He said,
"Yes!" That is, we are dealing with two levels here. The "judge"
disapproves of the attempt to straighten out the confusions and
200

disapproves of communicating the fact of his (the judge's)
disapproval.
We have to look for an etiology involving multiple levels of
trauma.
I am not talking at all about the content of these traumatic
sequences, whether they be sexual, or oral. Nor am I talking about
the age of the subject at the time of trauma, nor about which parent
is involved. That is all episodic as far as I'm concerned. I'm only
building up toward the statement that the trauma must have had
formal structure in the sense that multiple logical types were
played against each other to generate this particular pathology in
this individual.
Now, if you look at our conventional communication with one
another, what you find is that we weave these logical types with
incredible complexity and quite surprising facility. We even make
jokes, and these may be difficult for a foreigner to understand.
Most jokes, both canned and spontaneous, and nearly anywhere,
are weavings of multiple logical types. Kidding and hazing
similarly depend upon the unresolved question whether the kid-ee
can identify that this is kidding. In any culture, the individuals
acquire quite extraordinary skill in handling not only the flat
identification of what sort of a message a message is but in dealing
in multiple identifications of what sort of a message a message is.
When we meet these multiple identifications we laugh, and we
make new psychological discoveries about what goes on inside
ourselves, which is perhaps the reward of real humor.
But there are people who have the utmost difficulty with this
problem of multiple levels, and it seems to me that this unequal
distribution of ability is a phenomenon which we can approach
with the questions and terms of epidemiology. What is needed for a
child to acquire, or to not acquire, a skill in the ways of
interpreting these signals?
There is not only the miracle that any of them acquire the skills
—and a lot of them do—there is also the other side, that a great
many people have difficulty. There are people, for example, who,
when Big Sister in the soap opera suffers from a cold, will send a
bottle of aspirin to the radio station or recommend a cure for Big
Sister's cold, in spite of the fact that Big Sister is a fictitious
201

character within a radio soap opera. These particular members of
the audience are apparently a little bit askew in their identification
of what sort of a communication this is that is coming from their
radio.
We all make errors of that kind at various times. I'm not sure
that I've ever met anybody that doesn't suffer from "schizophrenia
P" more or less. We all have some difficulty in deciding sometimes
whether a dream was a dream or not, and it would not be very easy
for most of us to say how we know that a piece of our own fantasy
is fantasy and not experience. The ability to place an experience in
time is one of the important cues, and referring it to a sense organ
is another.
When you look at the mothers and fathers of patients for an
answer to this etiological question, you meet with several sorts of
answers.
First of all there are answers connected with what we may call
the intensifying factors. Any disease is made worse or more
probable by various circumstances, such as fatigue, cold, the
number of days of combat, the presence of other diseases, etc.
These seem to have a quantitative effect upon the incidence of
almost any pathology. Then there are those factors which I
mentioned—the hereditary characteristics and potentialities. To get
confused about the logical types, one presumably has to be
intelligent enough to know that there is something wrong, and not
so intelligent as to be able to see what it is that is wrong. I presume
that these characteristics are hereditarily determined.
But the nub of the probl em, it seems to me, is to identify what
real circumstances lead to the spec ific patholog y. I acknowledg e that
the bacteria are not really by any means the sole determinant of a
bacterial disease, and grant also therefo re that the occu rrence of
such traumatic sequence s or contexts is not by any means the sole
determinant of mental illness . But still it seems to me that the
identification of those contexts is the nub of unders tanding the
disease, as identifying the bacteria is essen tial to unde rstand ing a
bacterial disease.
I have met the mother of the patient whom I mentioned earlier.
The family is not badly off. They live in a nice tract house. I went
there with the patient, and when we arrived nobody was home. The
202

newspaper boy had tossed the evening paper out in the middle of
the lawn, and my patient wanted to get that paper from the middle
of that perfect lawn. He came to the edge of the lawn and started to
tremble.
The house looks like what is called a "model" home—a house
which has been furnished by the real estate people in order to sell
other houses to the public. Not a house furnished to live in, but
rather furnished to look like a furnished house.
I discussed his mother with him one day, and suggested that
perhaps she was a rather frightened person. He said, "Yes." I said,
"What is she frightened of?" He said, "The appeariential
securities."
There is a beautiful, perfectly centered mass of artificial, plastic
vegetation on the middle of the mantle. A china pheasant here and
a china pheasant there, symmetrically arranged. The wall-to-wall
carpet is exactly as it should be.
After his mother arrived, I fe lt a little unco mfortable , intrud ing in
this hous e. He had not visited there for about five years, but thing s
seemed to be going all right, so I decided to leave him there and to
come back when it was time to go back to the hospi tal. That gave
me an hour in the streets with absolutely nothing to do, and I began
to think what I would like to do to this setup. What and how could I
communica te? I decided that I would like to put into it something
that was both beaut iful and untid y. In trying to imple ment that
decision, I decided that flowers were the answe r, so I bought some
gladio luses . I took the gladioluses, and, when I went to get him, I
present ed them to the mother with a speech t hat I wanted her to have
in her house something that was "both beaut iful and untid y." "Ohl"
she said, "Those are not untid y flowe rs. As each one withe rs, you
can sn ip it off."
Now, as I see it, what is interesting is not so much the castrative
statement in that speech, but the putting me in the position of
having apologized when in fact I had not. That is, she took my
message and reclassified it. She changed the label which indicated
what sort of a message it was, and that is, I believe, what she does
all the time. An endless taking of the other person's message and
replying to it as if it were either a statement of weakness on the
203

part of the speaker or an attack on her which should be turned into
a weakness on the part of the speaker; and so on.
What the patient is up against today—and was up against in
childhood—is the false interpretation of his messages. If he says,
"The cat is on the table," she replies with some reply which makes
out that his message is not the sort of message that he thought it
was when he gave it. His own message identifier is obscured or
distorted by her when the message comes back. And her own
message identifier she continually contradicts. She laughs when
she is saying that which is least funny to her, and so on.
Now there is a regular maternal dominance picture in this
family, but I am not concerned at the moment to say that this is the
necessary form of the trauma. I am only concerned with the purely
formal aspects of this traumatic constellation; and I presume the
constellation could be made up with father taking certain parts of
it, mother taking certain other parts of it, and so forth.
I am trying to make only one point: that there is here a
probability of trauma which will contain certain formal char-
acteristics. It will propagate a specific syndrome in the patient
because the trauma itself has impact upon a certain element in the
communicational process. That which is at-tacked is the use of
what I have called the "message-identifying signals"—those
signals without which the "ego" dare not discriminate fact from
fantasy or the literal from the metaphoric.
What I tried to do was pinpoint a group of syndromata, namely
those syndromata related to an inability to know what sort of a
message a message is. At one end of the classification of those,
there will be more or less hebephrenic individuals for whom no
message is of any particular definite type but who live in a sort of
chronic shaggy-dog story. At the other end are those who try to
overidentify, to make an overly rigid identification of what sort of
a message every message is. This will give a much more paranoid
type of picture. Withdrawal is another possibility.
Finally, it seems to me that with a hypothesis of this kind, one
could look for the determinants in a population which might lead
to the occurrence of that sort of constellation. This would seem to
me an appropriate matter for epidemiological study.
204

Toward a Theory of Schizophr enia*
Schizophren ia—i ts nature, etiology, and the kind of therapy to
use for it—re mains one of the most puzz ling of the mental illnesse s.
The theory of schizophr enia presented here is based on
communica tions analysis, and specifically on the Theory of Logica l
Types. From this theory and from observations of schizophr enic
patients is deriv ed a description, and the necessar y cond itions for, a
situation called the "doub le bind"—a situation in which no matter
what a person does, he "can' t win." It is hypothes ized that a person
caught in the doub le bind may develop schizophren ic symptoms.
How and why the double bind may arise in a family situation is dis-
cussed, togeth er with illustration s from clinica l and experimental
data.
This is a report24 on a research project which has been
formulating and testing a broad , systematic view of the natur e,
etiology, and therap y of schizoph renia . Our research in this field
has proceeded by discussion of a varied body of data and
ideas, with all of us contributing according to our varied
experience in anthropology, communications analysis,
psychotherapy, psychiatry, and psychoanalysis. We have now
* This paper by Gregor y Bateson, Don D. Jackson, Jay Hale y, and. John H.
Weakland is here reproduced from Behavioral Science, Vol. I, No. 4, 1956, by
permission of Behavioral Science24 This paper derives from hypotheses first developed in a research project
financed by the Rockfeller Foundation from 1952-54, administered by the
Department of Sociology and Anthropology at Stanford University and directed by
Gregor y Bateson . Since 1954 the project has financed by the Josiah Macy, Jr.
Foundation. To Jay Haley is due credit for recognizing that the symptoms of
schizophrenia are suggestive of an inability to discriminate the Logical
Types, and this was amplified by Bateson, who added the notion that the
symptoms and etiology could be formally described in terms of a double bind
hypothesis. The hypothesis was communicated to D. D. Jackson and found to fit
closely with his ideas of family homeostasis. Since then Dr. Jackson has
worked closely with the project. The study of the formal analogies between
hypnosis and schizophrenia has been the work of John H. Weakland and Jay
Haley.
205

reached common agreement on the broad outlines of a
communicational theory of the origin and nature of
schizophrenia; this paper is a preliminary report on our con-
tinuing research.
The Bas e in Commun ications Theo ry
Our approach is based on that part of communication s theory
which Russ ell has called the Theory of Logic al Types.25 The central
thesis of this theory is that there is a discontinuity between a class
and its members. The class cannot be a member of itself nor can one
of the members be the class, since the term used for the class is of a
different level of abstraction—a differen t Logi cal Type—fro m
terms used for members. Although in formal logic there is an at-
tempt to maintain this discon tinuity between a class and its
members, we argue that in the psycholog y of real communications
this discon tinuity is cont inual ly and inev itably breached ,26 and that a
priori we must expect a patho logy to occu r in the human organis m
when certain formal pat-terns of the breach ing occu r in the
communication between mother and child. We shall argue that this
patho logy at its extreme will have symptoms whose formal
characteristics would lead the patho logy to be classified as a
schizophren ia.
Illustrations of how human beings handle communicat ion
involving multiple Logical Types can be derived from the following
fields:
1. The use of various communica tiona l modes in human
communication . Examples are play, nonpla y, fantasy, sacrament,
metapho r, etc. Even among the lower mammals there appea rs to be
an exchange of signal s which identify certain meaningfu l behavior
as "play," etc.27 These signa ls are evidently of highe r Logi cal Type
than the messages they classify. Among human beings this framing
25 A. N. Whiteh ead and B. Russell , Principia Mathematica, Cambridge,
Cambridge University Press, 1910.26 G. Bateson , "A Theory of Play and Fantas y," Psychiatric Research Reports,
1955, 2: 39-51.
206

and labeling of messages and meaningful actions reach es
consider able complexit y, with the pecu liarity that our vocabula ry for
such discrimination is still very poorly developed , and we rely
preponderantly upon nonve rbal media of postu re, gesture , facial ex-
pression , intonat ion, and the context for the communication of these
highly abstract, but vitally importan t, labels.
2. Humor . This seems to be a method of explo ring the implicit
themes in though t or in a relationship . The method of exploration
involves the use of messages which are characterized by a
cond ensat ion of Logic al Types or communication al modes. A
discover y, for example, occu rs when it sudden ly beco mes plain that
a message was not only metaphori c but also more literal, or vice
versa. That is to say, the explos ive moment in humor is the moment
when the labeling of the mode under goes a disso lution and
resynthesis . Commonly, the punch line compels a re-evaluation of
earlier signa ls which ascrib ed to certain messages a particular mode
(e.g., literalness or fantasy). This has the pecul iar effect of
attribut ing mode to those signals which had previous ly the status of
that highe r Logi cal Type which c lassifies the m odes.
3. The falsification of mode-i dentifying signa ls. Among human
beings mode identifiers can be falsified, and we have the artificial
laugh, the manipul ative simulation of friend liness, the confidence
trick, kidding, and the like. Similar falsifications have been reco rded
among mammals.28 Among human beings we meet with a strange
pheno menon—the unconsciou s falsification of these signals . This
may occur within the self—the subjec t may conc eal from himself
his own real hostility under the guise of metaphor ic play—or it may
occur as an unconscious falsification of the subjec t's unde rstand ing
of the other person's mode-id entifying signals. He may mistake
shyness fo r con tempt, etc. Indeed most of the errors of self-reference
fall under this head.
4. Learn ing. The simplest level of this pheno menon is
exemplified by a situation in which a subj ect rec eives a message and
27 A film prepared by this project , "The Nature of Play; Part I, River Otters," is
available.28 C. R. Carpenter, "A Field Study of the Behavior and Social Relations of
Howling Monkeys," Comp. Psychol. Monogr ., 1934, 10: 1–168; also K. Z. Lorenz,
King Solomon's Ring, New York, Crowell, 1952.
207

acts appropri ately on it: "I heard the clock strike and knew it was
time for lunch . So I went to the table." In learning experiments the
analogu e of this sequen ce of even ts is observed by the exper imenter
and commonly treated as a single message of a higher type. When
the dog salivates between buzze r and meat powde r, this sequence is
accep ted by the experi menter as a message indicating that "The dog
has learned that buzzer means meat powder ." But this is not the end
of the hierarchy of types involved . The exper imental subject may
beco me more skilled in learning . He may learn to learn,29 and it is
not inconceivable that still higher orders of learning may occur in
human being s.
5. Multiple levels of learning and the Logical Typing of signa ls.
These are two insep arable sets of pheno mena—inseparable because
the ability to handle the multiple types of signals is itself a learn ed
skill and t herefore a function of the multiple levels of learning.
According to our hypothes is, the term "ego funct ion" (as this
term is used when a schizoph renic is described as having "weak ego
function") is precisely the process of discriminating
communicational modes either within the self or between the
self and others. The schizophr enic exhibits weakness in three areas
of such function: (a) He has difficulty in assign ing the correct
communication al mode to the messages he receives from other
person s. (b) He has difficulty in assign ing the correct
communication al mode to t hose m essages whi ch he hi mself utters or
emits nonverb ally. (c) He has difficulty in assign ing the correct
communication al mode to his own thoughts , sensa tions , and
percepts.
At this point it is appropr iate to compare what was said in the
previous paragraph with von Domarus'7 approa ch to the systematic
description of schizophren ic utterance. He sugge sts that the
messages (and though t) of the schizophrenic are deviant in
syllogis tic structure. In place of structures which derive from the
syllogis m, Barba ra, the schizophren ic, accord ing to this theory, uses
29 G. Bateson , "Social Planning and the Concept of Deutero-Learning,"
Conference on Science, Philosophy and Religion , Second Symposium , New York,
Harper , 1942. (See above, p. 159) ; also H. F. Harlo w, "The Formation of Learning
Sets, " Psychol . Review , 1949 , 56: 51–65; also C. L. Hull, et al., Mathematico-
deductive Theory of Rote Learning , New Haven, Yale Universit y Press, 1940.
208

structures which identify predica tes. An example of such a distorted
syllogis m is:
Men die.
Grass dies.
Men are grass.
But as we see it, von Domarus30 formulation is only a more
precis e—and therefor e valuab le—wa y of saying that schizophrenic
utterance is rich in metaphor . With that gene ralization we agree. But
metaphor is an indispens able tool of though t and expre ssion—a
characteristic of all human communication , even of that of the
scientist. The conc eptua l models of cybernet ics and the energy
theories of psycho-anal ysis are, after all, only labeled metaphors .
The pecul iarity of the schizophr enic is not that he uses metaphor s,
but that he uses unlabeled metaphors. He has speci al difficulty in
hand ling signal s of that class whose members assign Logica l Types
to other signa ls.
If our formal summary of the symptomatolog y is correct and if
the schizoph renia of our hypothesi s is essentially a result of family
interaction, it shou ld be possible to arrive a priori at a formal
descrip tion of these sequen ces of experience which would induc e
such a symptomatolog y. What is known of learning theory combines
with the evident fact that human beings use context as a guide for
mode discrimination. Therefo re, we must look not for some specific
traumatic experience in the infantile etiology but rather for
characteristic sequen tial patterns . The specificity for which we
search is to be at an abstract or formal level. The sequen ces must
have this chara cteristic: that from them the patient will acqu ire the
mental habits which are exemplified in schizophren ic
communica tion. That is to say, he must live in a universe where the
sequences of events are such that his unconven tional
communicational habits will be in some sense appropriate. The
hypothesis which we offer is that sequen ces of this kind in the
externa l expe rience of the patient are respons ible for the inner
30 E. von Domarus, "The Specific Laws of Logic in Schizophrenia," Language
and Thought in Schizophrenia, J. S. Kasanin , ed., Berkeley, Universit y of
California Press, 1944.
209

conflicts of Logic al Typing. For such unreso lvabl e sequences of
exper iences , we use the term "double bind."
The D ouble Bind
The neces sary ingredien ts for a double bind situation, as we see
it, are:
1. Two or more persons. Of these, we designate one, for
purposes of our definition, as the "victim.". We do not assume that
the double bind is inflicted by the mother alone, but that it may be
done either by mother alone or by some combination of mother,
father, and/or siblings.
2. Repea ted experienc e. We assume that the doub le bind is a
recurrent theme in the expe rience of the victim. Our hypothes is does
not invoke a single traumatic expe rience, but such repeated
exper ience that the doub le bind structure comes to be an habitual
expec tation.
3. A primary nega tive injunction. This may have either of two
forms: (a) "Do not do so and so, or I will punish you," or (b) "If you
do not do so and so, I will punish you." Here we select a context of
learning based on avoidan ce of punishment rather than a context of
reward seek ing. There is perhap s no formal reason for this selection.
We assume that the punish ment may be either the withdrawa l of
love or the expre ssion of hate or anger—or most devast ating— the
kind of abandon ment that results from the parent's expres sion of
extreme helplessn ess.31
4. A seconda ry injunct ion conflicting with the first at amore
abstract level, and like the first enforc ed by punishments or signals
which threaten surviv al. This second ary injunct ion is more difficult
to descr ibe than the primary for two reasons . First, the seconda ry
injun ction is commonly communicated to the child by nonverba l
means. Postur e, gestu re, tone of voice , meaningful action, and the
implications concea led in verbal comment may all be used to
31 Our concept of punishment is being refined at present. It appears to us to
involve perc eptual experience in a way that cannot be encompassed b y the notion of
"trauma."
210

conv ey this more abstract message. Second , the secondar y injunc-
tion may impinge upon any element of the primary prohibition.
Verbalization of the second ary injunc tion may, there-fore , includ e a
wide variety of forms; for example, "Do no t see this as punish ment";
"Do not see me as the punishing agent" ; "Do not submit to my
prohibi tions" ; "Do not think of what you must not do"; "Do not
question my love of which the primary prohib ition is (or is not) an
example"; and so on. Other examples beco me possib le when the
doub le bind is inflicted not by one individual but by two. For ex-
ample, one parent may negat e at a more abstract level the
injunc tions of the other.
5. A tertiary negat ive injunct ion prohibiting the victim from
escaping from the field. In a formal sense it is perhaps unneces sary
to list this injunct ion as a separa te item since the reinfor cement at
the other two levels involves a threat to surviva l, and if the double
binds are imposed during infanc y, escape is natura lly impossible .
However , it seems that in some cases the escape from the field is
made impossib le by certain device s which are not purely negat ive,
e.g., capricious promises of love, and the like.
6. Finally, the complete set of ingredien ts is no longer necessary
when the victim has learned to perceive his universe in doub le bind
patterns. Almost any part of a double bind sequen ce may then be
sufficient to precipitate pan ic or rage.
The pattern of conflicting injunc tions may even be taken over by
hallucinato ry voices.32
The Ef fect of the Double Bind
In the Eastern religion , Zen Buddhis m, the goal is to achieve
enlightenment. The Zen master attempts to bring about
32 J. Perceval, A Narrati ve of the Treatment Experienced by a Gentle man During
a State of Mental Derangement, Designed to Explain the Causes and Nature of
Insanit y, etc., London, Effingham Wilson, 1836 and 1840. (See biblio graphic item,
1961 a.)
211

enlighten ment in his pupil in various ways. One of the thing s he
does is to hold a stick over the pupil's head and say fiercely, "If you
say this stick is real, I will strike you with it. If you say this stick is
not real, I will strike you with it. If you don't say anything, I will
strike you with it." We feel that the schizophreni c finds himself
continual ly in the same situation as the pupil, but he achieve s
something like disorienta tion rather than enlighten ment. The Zen
pupil might reach up and take the stick away from the master—who
might accept this response , but the schizophreni c has no such choice
since with him there is no not caring abou t the relationsh ip, and his
mother's aims and awa reness are not like the m aster' s.
We hypothesi ze that there will be a breakdown in any
individua l's ability to discriminate between Logica l Types wheneve r
a doub le bind situation occu rs. The general characteristics of this
situation a re the following:
(A) When the individua l is involved in an intense relationsh ip;
that is, a relationship in which he feels it is vitally importan t that he
discriminate accura tely what sort of m essage is being communicated
so that he may respond appropr iately.
(B) And, the individual is caught in a situation in which the othe r
person in the relationsh ip is express ing two orders of message and
one o f these denies the oth er.
(C) And, the individua l is unab le to comment on the messages
being expressed to correct his discrimination of what order of
message to respond to, i.e., he cannot make a metaco mmunicative
statement.
We have sugg ested that this is the sort of situation which occurs
between the preschizoph renic and his mother, but it also occu rs in
normal relationsh ips. When a person is caugh t in a doub le bind
situation, he will respond defen sively in a manner similar to the
schizophren ic. An individua l will take a metaphori cal statement
literally when he is in a situation where he must respond, where he
is faced with contradictory messages, and when he is unab le to
comment on the contradictions. For example, one day an employee
went home during office hours . A fellow emplo yee called him at his
home, and said lightly, "Well, how did you get there?" The
employee replied, "By automobile." He responded literally because
he was faced with a message which asked him what he was doing at
212

home when he should have been at the office, but which denied that
this question was being asked by the way it was phrased. (Since the
speaker felt it wasn't really his busine ss, he spoke metaphori cally.)
The relationship was intense enough so that the victim was in doubt
how the information would be used, and he therefore responded
literally. This is characteristic of anyone who feels "on the spot," as
demonstrated by the careful literal replies of a witne ss on the stand
in a court trial. The schizoph renic feels so terribly on the spot at all
times that he habitually responds with a defensiv e insisten ce on the
literal level when it is quite inapp ropriate, e.g., when someone is
joking.
Schizophren ics also confus e the literal and metaphor ic in their
own utterance when they feel themselves caugh t in a double bind.
For example, a patient may wish to criticize his therapist for being
late for an appoin tment, but he may be unsur e what sort of a
message that act of being late was—pa rticula rly if the therapis t has
anticipated the patient's reaction and apolog ized for the event. The
patient canno t say, "Why were you late? Is it because you don't want
to see me today?" This would be an accusa tion, and so he shifts to a
metaphoric al statement. He may then say, "I knew a fellow once
who missed a boat , his name was Sam and the boat almost sunk , . . .
etc.," Thus he develop s a metaphor ical story and the therapist may
or may not discove r in it a comment on his being late. The
conv enien t thing about a metaphor is that it leaves it up to the
therap ist (or mother) to see an accus ation in the statement if he
choo ses, or to ignore it if he chooses . Should the therapis t accep t the
accusat ion in the metaphor , then the patient can accept the statement
he has made abou t Sam as metaphorical. If the therapist points out
that this doesn't sound like a true statement about Sam, as a way of
avoiding the accus ation in the story, the patient can argue that there
really was a man named Sam. As an answe r to the doubl e bind
situation, a shift to a metaphoric al statement brings safety. However ,
it also preven ts the patient from making the accusation he wants to
make. But instead of getting over his accusat ion by indicating that
this is a metaphor , the schizophren ic patient seems to try to get over
the fact that it is a metaphor by making it more fantas tic. If the
therap ist should ignore the accusa tion in the story about Sam, the
schizophr enic may then tell a story about going to Mars in a rocke t
213

ship as a way of putting over his accusation. The indica tion that it is
a metaphor ical statement lies in the fantas tic aspect of the metaphor ,
not in the signals which usual ly accompany metaphors to tell the
listene r that a metaphor is being used.
It is not only safer for the victim of a doub le bind to shift to a
metaphor ical order of message, but in an impossib le situation it is
better to shift and beco me somebod y else, or shift and insist that he
is somewhere else. Then the doub le bind cannot work on the victim,
becaus e it isn't he and besides he is in a different place. In other
words, the statements which show that a patient is disoriented can be
interpr eted as ways of defending himself agains t the situation he is
in. The patho logy enters when the victim himself either does not
know that his response s are metaphorical or canno t say so. To
recogn ize that he was speaking metaphor ically he would need to be
aware that he was defending himself and therefor e was afraid of the
other person . To him such an awarenes s would be an indictment of
the other person a nd th erefo re provoke d isaster.
If an individual has spent his life in the kind of doub le bind
relationship described here, his way of relating to people after a
psychotic break would have a systematic pat-tern. First, he would
not share with normal peopl e those signal s which accompany
messages to indicate what a person means. His metaco mmunica tive
system—the communicat ions about communication—would have
broken down, and he would not know what kind of message a
message was. If a person said to him, "What would you like to do
today?" he would be unable to judge accura tely by the contex t or by
the tone of voice or gestu re whether he was being condemned for
what he did yesterda y, or being offered a sexual invitation, or just
what was meant. Given this in-ability to judge accurat ely what a
person really means and an excessive concern with what is really
meant, an individual might defend himself by choosing one or more
of severa l alternat ives. He might, for example, assume that behind
every statement there is a conce aled meaning whi ch is detrimental to
his welfare. He would then be excess ively conc erned with hidden
meanings and determined to demonstrate that he could not be
deceived—as he had been all his life. If he chooses this alternative,
he will be continual ly searching for meanings behind what peopl e
214

say and behind chan ce occur rences in the environ ment, and he will
be characte ristically suspi cious and d efian t.
He might choo se another alternative, and tend to accept literally
everything peop le say to him; when their tone or gestu re or context
contradicted what they said, he might establish a pattern of laugh ing
off these metacommunicative signa ls. He would give up trying to
discriminate between levels of message and treat all messages as
unimportant or to be laughed at.
If he didn't beco me suspicious of metacommunicat ive messages
or attempt to laugh them off, he might choo se to try to ignore them.
Then he would f ind it nece ssary to see and hea r less and le ss of what
went on around him, and do his utmost to avoid provoking a
response in his environment. He would try to detach his interest
from the external world and concent rate on his own interna l
processe s and, therefore, give the appea rance of being a withd rawn,
perhaps mute, indiv idual .
This is another way of saying that if an individua l doesn't know
what sort of message a message is, he may defend himself in ways
which have been descr ibed as paranoid , hebephren ic, or catatoni c.
These three alterna tives are not the only ones. The point is that he
cannot choose the one alternative which would help him to discover
what people mean; he cannot , withou t consid erabl e help, discu ss the
messages of others. Withou t being able to do that, the human being
is like any self-corr ecting system which has lost accept the
accusation in the metaphor, then the patient can accept the
statement he has made about Sam as metaphorical. If the therapist
points out that this doesn't sound like a true statement about Sam,
as a way of avoiding the accusation in the story, the patient can
argue that there really was a man named Sam. As an answer to the
double bind situation, a shift to a metaphorical statement brings
safety. However, it also prevents the patient from making the
accusation he wants to make. But instead of getting over his
accusation by indicating that this is a metaphor, the schizophrenic
patient seems to try to get over the fact that it is a metaphor by
making it more fantastic. If the therapist should ignore the
accusation in the story about Sam, the schizophrenic may then tell
a story about going to Mars in a rocket ship as a way of putting
over his accusation. The indication that it is a metaphorical
215

statement lies in the fantastic aspect of the metaphor, not in the
signals which usually accompany metaphors to tell the listener that
a metaphor is being used.
It is not only safer for the victim of a doub le bind to shift to a
metaphor ical order of message, but in an impossib le situation it is
better to shift and beco me somebod y else, or shift and insist that he
is somewhere else. Then the doub le bind cannot work on the victim,
becaus e it isn't he and besides he is in a different place. In other
words, the statements which show that a patient is disoriented can be
interpr eted as ways of defending himself agains t the situation he is
in. The patho logy enters when the victim himself either does not
know that his response s are metaphorical or canno t say so. To
recogn ize that he was speaking metaphor ically he would need to be
aware that he was defending himself and therefor e was afraid of the
other person . To him such an awarenes s would be an indictment of
the other person a nd th erefo re provoke d isaster.
If an individual has spent his life in the kind of doub le bind
relationship described here, his way of relating to people after a
psychotic break would have a systematic pat-tern. First, he would
not share with normal peopl e those signal s which accompany
messages to indicate what a person means. His metaco mmunica tive
system—the communicat ions about communication—would have
broken down, and he would not know what kind of message a
message was. If a person said to him, "What would you like to do
today?" he would be unabl e to judge accurately by the contex t or by
the tone of voice or gestu re whether he was being condemned for
what he did yesterda y, or being offered a sexual invitation, or just
what was meant. Given this in-ability to judge accurat ely what a
person really means and an excessive concern with what is really
meant, an individual might defend himself by choosing one or more
of severa l alternat ives. He might, for example, assume that behind
every statement there is a conce aled meaning whi ch is detrimental to
his welfare. He would then be excess ively conc erned with hidden
meanings and determined to demonstrate that he could not be
deceived—as he had been all his life. If he chooses this alternative,
he will be continual ly searching for meanings behind what peopl e
say and behind chance occurren ces in the environment, and he will
be charac teristically suspicious and defiant.
216

He might choo se another alternative, and tend to accept literally
everything peop le say to him; when their tone or gestu re or context
contradicted what they said, he might establish a pattern of laugh ing
off these metacommunicative signa ls. He would give up trying to
discriminate between levels of message and treat all messages as
unimportant or to be laughed at.
If he didn't beco me suspicious of metacommunicat ive messages
or attempt to laugh them off, he might choo se to try to ignore them.
Then he would f ind it nece ssary to see and hea r less and le ss of what
went on around him, and do his utmost to avoid provoking a
response in his environment. He would try to detach his interest
from the external world and concent rate on his own interna l
processe s and, therefore, give the appea rance of being a withd rawn,
perhaps mute, indiv idual .
This is another way of saying that if an individua l doesn't know
what sort of message a message is, he may defend himself in ways
which have been descr ibed as paranoid , hebephren ic, or catatoni c.
These three alterna tives are not the only ones. The point is that he
cannot choose the one alternative which would help him to discover
what people mean; he cannot , withou t consid erabl e help, discu ss the
messages of others. Withou t being able to do that, the human being
is like any self-co rrecting system which has lost its governo r; it
spirals into neve r-ending , but always systematic, distortions.
A Description of the Family Situation
The theore tical poss ibility of doub le bind situation s stimulated us
to look for such communicat ion sequences in the schizophren ic
patient and in his family situation. Toward this end we have studied
the written and verbal reports of psychotherap ists who have treated
such patients intens ively; we have studied tape record ings of
psychotherap eutic inter-views, both of our own patients and others;
we have inter-viewed and taped parent s of schizophren ics; we have
had two mothers and one father participate in intensive psy-
chotherap y; and we have interviewed and taped parents and patients
seen con jointly.
217

On the basis of these data we have deve loped a hypothes is about
the fa mily situation which ultimatel y leads to an individual suf fering
from schizophren ia. This hypothes is has not been statistically tested;
it selects and emphasizes a rather simple set of interactiona l
pheno mena and does not attempt to describe comprehensive ly the
extraord inary complexity of a family relation ship.
We hypothesize that the family situation of the schizophrenic has
the following general characteristics:
(1) A child whose mother becomes anxious and with-d raws if the
child respond s to her as a loving mother. That is, the child's very
existence has a special meaning to the mother which arouses her
anxie ty and hostility when she is in danger of intimate contac t with
the child.
(2) A mother to whom feeling s of anxiety and hostility toward
the child are not accept able, and whose way of denying them is to
expre ss over t loving behavio r to persuade the child to respond to her
as a loving mother and to with-draw from him if he does not.
"Loving behav ior" does not necessarily imply "affection" ; it can, for
example, be set in a framework of doing the proper thing , instilling
"goodness ," and th e like.
(3) The absence of anyone in the family, such as a strong and
insightfu l father, who can interven e in the relationsh ip between the
mother and child and support the child in the face of the
contradic tions involved.
Since this is a formal descrip tion we are not specifically
conce rned with why the mother feels this way abou t the child, but
we suggest that she could feel this way for variou s reason s. It may
be that merely having a child arouses anxie ty abou t herself and her
relationships to her own family; or it may be importan t to her that
the child is a boy or a girl, or that the child was born on the
annive rsary of one of her own sibling s,33 or the child may be in the
same sibling position in the family that she was, or the child may be
special to her for other reason s related to her own emotional
probl ems.
33 R. Hilgard, "Anniversar y Reactions in Parents Precipitated by Children ,"
Psychiatry, 1953, 16: 73-80.
218

Given a situation with these characte ristics, we hypothesize that
the mother of a schizophreni c will be simultaneousl y expressing at
least two order s of message . (For simplicity in this presentation we
shall confine ourselv es to two orders.) These orders of message can
be rough ly characterized as (a) hostile or withdrawing behavio r
which is aroused when-e ver the child approaches her, and (b)
simulated loving or approaching behav ior which is aroused when
the child respond s to her hostile and withdrawing behav ior, as a wa y
of denying that she is withd rawing. Her proble m is to control her
anxiety by controlling the closeness and distance between herself
and her child. To put this another way, if the mother begin s to feel
affectiona te and close to her child, she beg ins to feel endange red and
must withdraw from him; but she canno t accep t this hostile act and
to deny it must simulate affection and closene ss with her child. The
important point is that her loving behav ior is then a comment on
(since it is compensator y for) her hostile behav ior and consequen tly
it is of a different order of message than the hostile behav ior— it is a
message about a sequence of messages. Yet by its nature it denies
the existence of those messages which it is about , i.e., the hostile
withdrawal .
The mother uses the child's respons es to affirm that her behav ior
is loving, and since the loving behavio r is simulated, the child is
placed in a position where he must not accurat ely interpret her
communica tion if he is to maintain his relationship with her. In
other words, he must not discriminate accurately between orders of
message, in this case the difference between the expression of
simulated feelings (one Logical Type) and real feelings (another
Logical Type). As a result the child must systematically distort his
perception of metacommunicative signals. For ex-ample, if mother
begins to feel hostile (or affectionate) to-ward her child and also
feels compelled to withdraw from him, she might say, "Go to bed,
you're very tired and I want you to get your sleep." This overtly
loving statement is intended to deny a feeling which could be.
verbalized as "Get out of my sight because I'm sick of you." If the
child correctly discriminates her metacommunicative signals, he
would have to face the fact that she both doesn't want him and is
deceiving him by her loving behavior. He would be "punished" for
learning to discriminate orders of messages accurately. He
219

therefore would tend to accept the idea that he is tired rather than
recognize his mother's deception. This means that he must deceive
himself about his own internal state in order to support mother in
her deception. To survive with her he must falsely discriminate his
own internal messages as well as falsely discriminate the messages
of others.
The probl em is compounded for the child becaus e the mother is
"benevo lently" defining for him how he feels; she is express ing
overt maternal conce rn ove r the fact that he is tired. To put it anothe r
way, the mother is contro lling the child's definitions of his own
messages , as well as the defin ition of his response s to her (e.g., by
saying, "You don't really mean to say that," if he shou ld criticize
her) by insisting that she is not concerned about herself but only
about him. Consequent ly, the easiest path for the child is to accep t
mother's simulated loving behav ior as real, and his desires to
interpr et what is going on are under mined. Yet the result is that the
mother is withdr awing from him and defining this withd rawal as the
way a loving relation ship should be.
Howeve r, accepting mother's simulated loving behavior as real
also is no solution for the child. Should he make this false
discrimination , he would approach her; this move to-wa rd closeness
would provoke in her feelings of fear and helples sness , and she
would be compelled to withdraw . But if he then withdrew from her,
she would take his withdrawal as a statement that she was not a
loving mother and would either punish him for withdr awing or
approa ch him to bring him close r. If he then approa ched, she would
respond by putting him at a distanc e. The child is punish ed for
discriminat ing a ccurate ly what she is expressing, and he is punish ed
for discrimina ting i naccura tely—he i s caught in a doubl e bind .
The child might try various means of escaping from this
situation. He might, for example, try to lean on his father or some
other member of the family. However , from our preliminary
observ ation s we think it is likely that the fathers of schizophr enics
are not substantial enough to lean on. They are also in the awkward
position where if they agreed with the child abou t the nature of
mother's decep tions , they would need to recogn ize the nature of
their own relation-ship s to the mother, which they could not do and
remain attached to her in the modus operand i they have worked out.
220

The need of the mother to be wanted and loved also preven ts the
child from gaining suppor t from some other person in the
environment, a teacher , for example. A mother with these
characteristics would feel threatened by any other attachment of the
child and would break it up and bring the child back closer to her
with cons equent anxie ty when t he child bec ame dependen t on her.
The only way the child can really escap e from the situation is to
comment on the contradictory position his mother has put him in.
However , if he did so, the mother would take this as an accusat ion
that she is unloving and both punish him and insist that his
percept ion of the situation is distorted. By prevent ing the child from
talking about the situation, the mother forbids him using the
metaco mmunicative level—the level we use to correct our
percept ion of communicat ive behavio r. The ability to communica te
abou t communication, to comment upon the meaningfu l actions of
oneself and other s, is essen tial for success ful social inter-course. In
any normal relationship there is a constant inter-change of
metaco mmunicative messages such as "What do you mean?" or
"Why did you do that?" or "Are you kidding me?" and so on. To
discriminate accura tely what people are really expressing, we must
be able to comment directly or indirectly on that expres sion. This
metaco mmunicative level the schizophr enic seems unab le to use
successfu lly.34 Given these characteristics of the mother, it is
apparent why. If she is denying one order of message , then any
statement abou t her statements endangers her and she must forb id it.
Therefore , the child grows up unsk illed in his ability to com-
municate about communication and, as a result, unsk illed in
determining what peopl e really mean and unski lled in expressing
what he really means, whi ch is essential for normal relationships .
In summary, then, we suggest that the doubl e bind nature of the
family situation of a schizophr enic results in placing the child in a
position where, if he responds to his mother's simulated affection,
her anxiety will be arous ed and she will punish him (or insist, to
protec t herself, that his overtures are simulated , thus confusing him
abou t the nature of his own messages ) to defend herself from close-
34 G. Bateson, ". . . Play and Fantas y," op. cit.
221

ness with him. Thus the child is blocked off from intimate and
secure associa tions with his mother. Howeve r, if he does not make
overtures of affection , she will feel that this means she is not a
loving mother and her anxie ty will be aroused. Therefore, she will
either punish him for with-dr awing or make overtures toward the
child to insist that he demonstra te that he loves her. If he then
respond s and shows her affection, she will not only feel endangered
again , but she may resent the fact that she had to force him to
respond . In either case in a relationship , the most importan t in his
life and the model for all others , he is punished if he indica tes love
and affection and punished if he does not; and his escape routes
from the situation , such as gaining support from other s, are cut off.
This is the basic nature of the doub le bind relationsh ip between
mother and child. This descrip tion has not depicted , of course, the
more complicated interlocking gestalt that is the "family" of which
the "mother" is one importan t part.35
Illustrations f rom Cl inica l Data
An analysis of an inciden t occu rring between a schizophren ic
patient and his mother illustrates the doub le bind situation. A young
man who had fairly well recove red from an acute schizophr enic
episod e was visited in the hosp ital by his mother . He was glad to see
her and impulsive ly put his arm around her shoulde rs, whereupon
she stiffened. He withdrew his arm and she asked , "Don't you love
me any more?" He then blushed , and she said, "Dear , you must not
be so easily embarra ssed and afraid of your feelings." The patient
was able t o stay with her on ly a few m inutes m ore and fol lowing her
depar ture he assaulted an aide and was put in the tubs.
Obviousl y, this result could have been avoided if the young man
had been able to say, "Mother, it is obvious that you beco me
unco mfortable when I put my arm around you, and that you have
35 D. D. Jackson, "The Questi on of Family Homeostasis," presented at the Americ an Psychiatric Associati on Meeting, St. Louis, May 7, 1954; also Jackson , "Some Factors Influencing the Oedipus Complex," Psychoanalytic Quarterly, 1954, 23: 566-81.
222

difficulty accepting a gestu re of affection from me." Howeve r, the
schizophr enic patient doesn' t have this possib ility open to him. His
intens e dependenc y and training prevents him from commenting
upon his mother's communicat ive behavio r, though she comments
on his and forces him to accept and to attempt to deal with the com-
plicated sequence . The complication s for the patient includ e the
following :
(1) The mother's reaction of not accepting her son's affectionate
gesture is masterful ly covered up by her conde mnation of him for
withdrawing , and the patient denies his perception of the situation
by accep ting her conde mnation .
(2) The statement "Don' t you love me any more" in this context
seems to imply:
(a) "I am lovab le."
(b) "Y ou should love me and if you don' t you are bad o r at fault."
(c) "Where as you did love me previousl y you don't any longe r,"
and thus focus is shifted from his expressing affection to his
inability to be affectiona te. Since the patient has also hated her, she
is on good ground here, and he responds appropr iately with guilt,
which she then attacks .
(d) "What you just express ed was not affection," and in order to
accept this statement, the patient must deny what she and the culture
have taught him abou t how one expre sses affection . He must also
question the times with her, and with others, when he thought he
was expe riencing affection and when they seemed to treat the
situation as if he had. He exper ience s here loss-o f-suppor t pheno m-
ena and is put in doubt about the reliability of past experienc e.
(3) The statement, "You must not be so easily embarrassed and
afraid of your feelings," seems to imply:
(a) "You are not like me and are different from other nice or
normal peop le because we express ou r feelings."
(b) "The feeling s you express are all right, it's only that you can't
accept them." However , if the stiffening on her part had indicated
"Thes e are unaccept able feelings," then the boy is told that he
shou ld not be embarrass ed by unaccep table feelings . Since he has
had a long training in what is and is not acceptab le to both her and
society, he again comes into conflict with the past. If he is unafraid
of his own feelings (which mother implies is good) , he should be
223

unafr aid of his affection and would then notice it was she who was
afraid, but he must not notice that be-cause her whole approach is
aimed at cover ing up this sho rt-coming in herself.
The impossib le dilemma thus beco mes: "If I a m to keep m y tie to
mother, I must not show her that I love her, but if I do not show her
that I love her, then I will lose her."
The importance to the mother of her speci al method of control is
striking ly illustrated by the interfamily situation of a young woman
schizophren ic who greeted the therap ist on their first meeting with
the remark, "Mother had to get married and now I'm here." This
statement meant to the therap ist that:
(1) The patient was the result of an illegitimate pregnancy.
(2) This fact was related to her present psychosis (in her
opinion).
(3) "Here" referred to the psychiat rist's office and to the patient's
presence on earth for which she had to be eternal ly indeb ted to her
mother, espec ially since her mother had sinned and suffered in order
to bring her into the wor ld.
(4) "Had to get married" referred to the shotgun natur e of
mother's wedding and to the mother's respons e to pressure that she
must marry, and the reciprocal, tha t she resented the forced natu re of
the situation and blamed the patient for it.
Actual ly, all these suppos itions subsequent ly proved to be
factually correct and were corrobo rated by the mother during an
abortive attempt at psychotherap y. The flavo r of the mother's
communication s to the patient seemed essent ially this: "I am
lovab le, lov ing, and satisfied with m yself. You are lovab le when y ou
are like me and when you do what I say." At the same time the
mother indica ted to the daught er both by words and behavio r: "You
are physically delicate, unintelligent, and differen t from me (`not
normal'). You need me and me alone becau se of these handicaps ,
and I will take care of you and love you." Thus the patient's life was
a series of beginnings , of attempts at exper ience , which would resu lt
in failure and withdrawa l back to the maternal hearth and boso m
becaus e of the collusion between her and her mother.
It was noted in collaborat ive therapy that certain areas importan t
to the mother's self-esteem were especially conflictual situation s for
the patient. For example, the mother needed the fiction that she was
224

close to her family and that a deep love existed between her and her
own mother . By analogy the relationship to the grand mother served
as the prototype for the mother's relationsh ip to her own daugh ter.
On one occasion when the daughter was seven or eight years old,
the grandmother in a rage threw a knife which barely missed the
little girl. The mother said nothing to the grandmother but hurried
the little girl from the room with the words, "Grand mommy really
loves you." It is sign ificant tha t the grandmother took the at titude to-
ward the patient that she was not well enough controlled, and she
used to chide her daugh ter for being too easy on the child. The
grandmother was living in the hous e during one of the patient's
psychotic episodes, and the girl took great delight in throwing
various objects at the mother and grandmother while they cower ed
in fear.
Mother felt herse lf very attractive as a girl, and she felt that her
daugh ter resembled her rather closel y, although by damning with
faint praise, it was obviou s that she felt the daughte r definitely ran
second. One of the daught er's first acts during a psychotic period
was to announce to her mother that she was going to cut off all her
hair. She proce eded to do this while the mother pleaded with her to
stop. Subsequen tly the mother would show a picture of herself as a
girl and expla in to peop le how the patient would look if she only
had her beau tiful hair.
The mother, apparently without awareness of the significance of
what she was doing, would equate the daughter's illness with not
being very bright and with some sort of organic brain difficulty.
She would invariably contrast this with her own intelligence as
demonstrated by her own scholastic record. She treated her
daughter with a completely patronizing and placating manner
which was insincere. For example, in the psychiatrist's presence
she promised her daughter that she would not allow her to have
further shock treatments, and as soon as the girl was out of the
room she asked the doctor if he didn't feel she should be
hospitalized and given electric shock treatments. One clue to this
deceptive behavior arose during the mother's therapy. Although the
daughter had had three previous hospitalizations, the mother had
never mentioned to the doctors that she herself had had a psychotic
episode when she discovered that she was pregnant. The family
225

whisked her away to a small sanitarium in a nearby town, and she
was, according to her own statement, strapped to a bed for six
weeks. Her family did not visit her during this time, and no one
except her parents and her sister knew that she was hospitalized.
There were two times during therapy when the mother showed
intense emotion. One was in relating her own psychotic
experience; the other was on the occasion of her last visit when she
accused the therapist of trying to drive her crazy by forcing her to
choose between her daughter and her husband. Against medical
advice, she took her daughter out of therapy.
The father was as involved in the homeostatic aspec ts of the
intrafamily situation as the mother. For example, he stated that he
had to quit his posi tion as an important attorney in orde r to bring his
daughte r to an area where competent psychiat ric help was available.
Subsequent ly, acting on cues from the patient (e.g., she frequen tly
referred to a chara cter named "Nervou s Ned") , the therap ist was a ble
to elicit from him that he had hated his job and for years had been
trying to "get out from under ." Howeve r, the daught er was made to
feel that the move was i nitiated for her.
On the basis of our examination of the clinical data, we have
been impressed by a number of observations including:
(1) The helplessness , fear, exaspe ration, and rage which a doub le
bind situation provok es in the patient, but which the mother may
serenely and un-und erstanding ly pass over. We have noted reactions
in the father that both create double bind situations, or extend and
amplify those created by the mother , and we have seen the father,
passive and outraged , but helpless , beco me ensna red in a similar
manner t o the patient.
(2) The psychosis seems, in part, a way of dealing with doub le
bind situations to overco me their inhibiting and con-t rolling effect.
The psychotic patient may make astute, pithy, often metaphor ical
remarks that reveal an insight into the forces binding him.
Contra riwise , he may beco me rather expe rt in setting doubl e bind
situations h imself.
(3) Accord ing to our theory, the communicat ion situation
described is essen tial to the mother's secur ity, and by inference to
the family homeostas is. If this be so, then when psychother apy of
the patient helps him beco me less vulnerab le to mother's attempts at
226

control, anxiety will be produced in the mother. Similarly, if the
therap ist interpr ets to the mother the dynamics of the situation she is
setting up with the patient, this should produce an anxiety response
in her. Our impress ion is that when there is a perdu ring con-tact
between patient and family (especi ally when the patient lives at
home during psychothe rapy), this leads to a disturb ance (often
severe ) in the mother and sometimes in both mother and father and
other siblings .36
Current Position and Future Prospe cts
Many writers have treated schizophren ia in terms of the most
extreme contrast with any other form of human thinking and
behavior . While it is an isolable phenomenon, so much emphasis on
the differen ces from the normal—ra ther like the fearful physical
segrega tion of psychotic s—does not help in understanding the
proble ms. In our appro ach we assume that schizophren ia involves
general princip les which are importan t in all communication and
therefore many in-formative similari ties can be found in "normal"
communication situations.
We have -been particula rly interested in various sorts of
communica tion which involve both emotiona l significance and the
necessi ty of discriminating between orders of message. Such
situations include play, humor, ritual, poetry, and fiction. Play,
especia lly among animals, we have studied at some length.37 It is a
situation which s triking ly illustrates t he occu rrenc e of m etamessage s
whose correct discrimination is vital to the coope ration of the
individu als involved ; for ex-ample, false discrimination could easily
lead to combat. Rather closely related to play is humor, a continuing
subject of our – research. It involves sudden shifts in Logical Types
as well as discrimination of those shifts. Ritual is a field in which
unusu ally real or literal ascriptions of Logical Type are made and
36 D. D. Jackson, "An Episode of Sleep walking," Journal of the American
Psychoanal ytic Association, 1954, 2: 503—508; also Jackson, "Some Factors . . . ,"
Psycho-analytic Quarterl y, 1954, 23: 566—581 .37 Bateson, "A Theor y of Play …" op. cit.
227

defended as vigorousl y as the schizophren ic defend s the "reality" of
his delusions . Poetry exemplifies the communicative power of
metaphor—ev en very unusua l metaphor—when labeled as such by
various signs, as contrasted to the obscu rity of unlabeled
schizophren ic metaphor. The entire field of fictiona l communication,
defined as the narration or depiction of a series of events with more
or less of a label of actuality, is most relevan t to the investigation of
schizophren ia. We are not so much concerned with the content
interpr etation of fiction—al though analysis of oral and destructive
themes is illuminating to the studen t of schizophren ia—as with the
formal problems involved in simultan eous existence of multipl e
levels of message in the fictional presentation of "reality." The
drama is especially interesting in this respect, with both performers
and spectators responding to messages about both the actual and
the theatrical reality.
We are giving extens ive attention to hypnosis. A great array of
pheno mena that occur as schizophren ic symptoms—halluc inations,
delus ions, alterations of personal ity, amnesias, and so on—can be
produced temporar ily in normal subjects with hypnosis . These need
not be directly suggest ed as specific pheno mena, but can be the
"spont aneous" result of an arranged communication sequenc e. For
example, Erickson38 will produ ce a hallucin ation by first inducing
catalepsy in a subject's hand and then saying, "There is no
conce ivable way in which your hand can move, yet when I give the
signa l, it must move." That is, he tells the subjec t his hand will
remain in place, yet it will move, and in no way the subject can
consc iousl y conceiv e. When Erickson gives the signal , the subjec t
hallucina tes the hand moved, or hallucinates himself in a different
place and therefor e the hand was m oved. This use of hallucination to
resolve a problem posed by contradic tory commands which cannot
be discussed seems to us to illustrate the solution of a double bind
situation via a shift in Logi cal Types. Hypnotic respon ses to direct
sugges tions or statements also commonly involve shifts in type, as
in accepting the words "Her e's a glass of water" or "You feel tired"
as external or internal reality, or in literal response to metaphor ical
38 M. H. Erickson, P ersonal communication, 1955.
228

statements, much like schizophren ics. We hope that further study of
hypnotic induc tion, pheno mena, and waking will, in this con-
trollable situation, help sharpen our view of the essen tial
communica tiona l sequence s which produce phenomena like those of
schizophr enia.
Anothe r Erick son exper iment seems to isolate a doub le bind
communica tiona l sequence without the specific use of hypnosis .
Erickson arranged a seminar so as to have a young chain smoker sit
next to him and to be withou t cigare ttes; other participant s were
briefed on what to do. All was ordered so that Erickson repeat edly
turned to offer the young man a cigarette, but was alwa ys
interrupted by a question from someone so that he turned away,
"inadvertently" withdrawing the cigarettes from the young man's
reach. Later another participant asked this young man if he had
received the cigarette from Dr. Erickson. He re-plied, "What
cigarette?", showed clearly that he had forgot-ten the whole
sequence, and even refused a cigarette offered by another member,
saying that he was too interested in the seminar discussion to
smoke. This young man seems to us to be in an experimental
situation paralleling the schizophrenic's double bind situation with
mother: an important relationship, contradictory messages (here of
giving and taking away), and comment blocked—because there
was a seminar going on, and anyway it was all "inadvertent." And
note the similar outcome: amnesia for the double bind sequence
and reversal from "He doesn't give" to "I don't want."
Although we have been led into these collateral areas, our main
field of observ ation has been schizophren ia itself. All of us have
worked directly with schizophren ic patients and much of this case
materia l has been recorded on tape for detailed study. In addition,
we are recording interviews held jointly with patients and their
families, and we are taking sound motion pictures of mothers and
disturbed, presumabl y preschizoph renic , children. Our hope is that
these opera tions will provide a clearly evident record of the con-
tinuing , repetitive doub le binding which we hypothes ize goes on
steadily from infantile beginnings in the family situation of
individu als who beco me schizophren ic. This basic family situation,
and the overtly communication al charac teristics of schizophr enia,
have been the major focus of this pape r. However , we expect our
229

concep ts and some of these data will also be useful in future work
on other probl ems of schizoph renia , such as the variety of other
symptoms, the charac ter of the "adjus ted state" before schizophren ia
beco mes manifest, and the nature and circumstances of the
psychotic break.
Therapeut ic Implications of this Hypothe sis
Psychothe rapy itself is a context of multilevel communication,
with explo ration of the ambiguous lines between the literal and
metaphor ic, or reality and fantasy, and i ndeed , various forms of play,
drama, and hypnosis have been used extensively in therapy. We
have been interested in therapy, and in addition to our own data we
have been collecting and examining recordings, verbatim
transcripts, and personal accounts of therapy from other therapists.
In this we prefer exact records since we believe that how a
schizophrenic talks depends greatly, though often subtly, on how
another person talks to him; it is most difficult to estimate what
was really occurring in a therapeutic interview if one has only a
description of it, especially if the description is already in
theoretical terms.
Except for a few gene ral remarks and some specu lation,
however , we are not yet prepa red to comment on the relation of the
double bind to psychothe rapy. At presen t we c an on ly note:
(1) Doubl e bind situations are created by and within the
psychothe rapeu tic setting and the hospital milieu. From the point of
view of this hypothesis , we wonder about the effect of medical
"benevo lence" on the schizophr enic patient. Since hospi tals exist for
the benefit of personnel as well as—as much as—more than— for
the patient's benefit, there will be contradictions at times in
sequenc es where actions are taken "benevolen tly" for the patient
when actually they are intended to keep the staff more comfortable .
We would assume that whenever the system is organized for hospi tal
purpose s and it is announc ed to the patient that the actions are for
his benef it, then the schizoph renogeni c situation is being
230

perpetua ted. This kind of decep tion will provoke the patient to
respond to it as a doub le bind situation, and his response will be
"schizoph renic" in the sense that it will be indirect and the patient
will be unab le to comment on the fact that he feels that he is being
deceived . One vigne tte, fortuna tely amusing, illustrates such a
response . On a ward with a dedicated and "benevo lent" physician in
charge there was a sign on the physician's door which said "Doctor' s
Office. Please Knock ." The docto r was driven to distraction and
finally capitu lation by the obedien t patient who carefully knocked
every time he passed th e doo r.
(2) The under stand ing of the doub le bind and its communicat ive
aspects may lead to innovat ions in therapeu tic technique . Just what
these innova tions may be is difficult to say, but on the basis of our
invest igation we are assuming that doub le bind situations occur
consistently in psychotherapy. At times these are inadvertent in the
sense that the therapist is imposing a double bind situation similar
to that in the patient's history, or the patient is imposing a double
bind situation on the therapist. At other times therapists seem to
impose double binds, either deliberately or intuitively, which force
the patient to respond differently than he has in the past.
An incident from the expe rienc e of a gifted psychother apist
illustrates the intuitive unders tanding of a doub le bind
communica tiona l sequence. Dr. Frieda Fromm-Reich mann39 was
treating a young woman who from the age of seven had built a
highly complex religion of her own replete with power ful gods. She
was very schizoph renic and quite hesitant abou t entering into a
therapeu tic situation. At the be-ginning of the treatment she said,
"God R says I shou ldn't talk with you." Dr. Fromm-Reich mann
replied, "Look , let's get something into the record. To me God R
doesn't exist, and that whol e world of yours doesn't exist. To you it
does, and far be it from me to think that I can take that away from
you, I have no idea what it means. So I'm willing to talk with you in
terms of that world , if only you know I do it so that we have an
unde rstand ing that it doesn't exist for me. Now go to God R and tell
him that we have to talk and he shou ld give you permission. Also
you must tell him that I am a docto r and that you have lived with
39 F. Fromm-Reich mann, Personal com munication, 1956
231

him in his kingdo m now from seven to sixteen— that's nine years —
and he hasn't helped you. So now he must permit me to try and see
whether you and I can do that job. Tell him that I am a doctor and
this is what I want t o try."
The the rapis t has he r patient i n a "the rapeu tic doub le bind ." If t he
patient is rendered doubtfu l abou t her belief in her god, then she is
agreeing with Dr. Fromm-Reichmann, and is admitting her
attachment to ther apy. If she ins ists tha t God R is rea l, then she m ust
tell him that Dr. Fromm-Reich mann is "more power ful" than he—
again admitting her involvement wi th the therap ist.
The difference between the therapeut ic bind and the origina l
double bind situation is in part the fact that the therapis t is not
involved in a life and death strugg le himself. He can therefore se t up
relatively benevolen t binds and gradual ly aid the patient in his
emancipa tion from them. Many of the unique ly appropr iate
therapeut ic gambits arranged by therap ists seem to be intuitive. We
share the goal of most psychothe rapis ts who strive toward the day
when such strokes of genius will be well enough unde rstood to be
systematic and co mmonplace .
Addi tiona l Refe rences
J. Hale y, "Paradoxes in Play, Fantas y, and Psychotherap y," Psychiatric
Research Reports, 1955, 2: 52-8.
J. Ruesch and G. Bateson , Communication: The Social Matrix of Psychiatry, New
York, Norton, 1951.
232

The Grou p Dyna mics of Schizophr enia*
First, I intend to attach very specific meaning to the title of this
paper. An essential no tion a ttach ed to th e word "group" as I shall use
it is the idea of relatedn ess between members. Our concern is not
with the sort of phenomena which occur in exper imental ly formed
groups of gradu ate student s who have no previou sly determined
habits of communicat ion—no habitual differentiations of role. The
group to which I mostly refer is the family; in genera l, those fami-
lies in which the paren ts maintain an adjustment to the world around
them withou t being recogniz ed as grossly devian t, while one or
more of their offspring differ consp icuous ly from the normal
popu lation in the frequ ency and obvious nature of their responses . I
shall also be thinking of other groups analogou s to these, i.e., ward
organiza tions , which work in such a way as to promote
schizophr enic or schizophrenoid behavio r in some of the m embers.
The word "dynamics" is loosel y and conv ention ally used for all
studie s of person al interaction and espec ially when they stress
change or learning exhibited by the subjec ts. De-sp ite our following
its convent ional use, this word is a misnomer. It evokes analog ies
with ph ysics whi ch are totally false.
"Dynamics" is principa lly a language devised by physicists and
mathematic ians for the descr iption of certain even ts. In this strict
sense, the impact of one billiard ball upon an-othe r is subject matter
for dynamics, but it would be an error of language to say that
billiard balls "behave ." Dynamics appropr iately describe those
events whose descriptions can be checked by asking whethe r they
contraven e the First Law of Thermodynamics, the Law of the
* The ideas in this lecture represent the combined thinking of the staff of The
Project for the Study of Schizophrenic Communication. The staff consisted of
Gregory Bateson, Jay Haley, John H. Weakland, Don D. Jackson, M.D., and William F.
Fry, M.D
The article is reprinted from Chronic Schizoph renia: Explorations in Theory
and Treatment , edited by L. Appleby, J. M. Scher, and J. Cumming, The Free Press,
Glencoe , Illinois, 1960; reprinted by permission.
233

Conservation of Energy. When one billiard ball strikes anothe r, the
motion of the second is energized by the impact of the first, and
such transferences of energy are the central subject matter of
dynamics. We, howeve r, are not conce rned with event sequence s
which have this characteristic. If I kick a ston e, the movement of the
stone is energized by the act, but if I kick a dog, the behavior of the
dog may indeed be partly conse rvative—he may travel along a
Newtonian trajectory if kicked hard enough, but this is mere
physics. What is importan t is that he may exhib it responses which
are energized not by the kick but by his metabolis m; he may turn
and b ite.
This, I think, is what peop le mean by magic. The realm of
pheno mena in which we are interes ted is alwa ys characterized by
the fact that "ideas" may influence events . To the physicist, this is a
gross ly magical hypothesi s. It is one which cannot be tested by
asking questions about the conse rvation o f ene rgy.
All this, however , has been better and more rigorous ly said by
Berta lanffy, which makes it easier for me to further explore this
realm of pheno mena in which communication occur s. We shall settle
for the conven tiona l term "dynamics" provided it is clearly
unders tood that we are not talking abou t dynamics in the physical
sense .
Robert Louis Stevenson40 in "The Poor Thing" has achieved
perhap s the most vivid cha racterization of this magical realm:
"In my thought one thing is as good as anothe r in this world ; and
a shoe of a horse will do." The word "yes" or a whol e performance
of Hamlet, or an injection of epinephr ine in the right place on the
surface of the brain may be interchangeab le objects. Any one of
them may, ac-cord ing to the convent ions of communicat ion
established at that moment, be an affirmative (or a negative) answer
to any question. In the famous message , "One if by land; two if by
sea," the objects actua lly used were lamps, but from the point of
view o f communicat ions theor y, they could h ave be en an ything f rom
aardv arks to zygomatic arche s.
40 R. L. Stevenson, "The Poor Thing," Novels and Tales of Robert Louis
Stevenson, Vol. 20, New York, Scribners, 1918, pp. 496-502.
234

It might well be sufficiently confus ing to be told that, accord ing
to the conv ention s of co mmunication in use at the m oment, an ything
can stand for anything else. But this realm of magic is not that
simple. Not only can the shoe of a horse stand for anything else
according to the conventions of communication , it can also and
simultaneou sly be a signal which will alter the convent ions of
communication. My finge rs crossed behind my back may alter the
whole tone and implication of everything. I recall a schizophren ic
patient who, like many other schizophren ics, had difficulty with the
first person pronoun; in particular, he did not like to sign his name.
He had a number of aliases, alterna tive named aspects of self. The
ward organization, of which he was a part, required that he sign his
name to obtain a pass, and for one or two weekends he did not
receiv e a pass becaus e he insisted on signing one of his aliases . One
day he remarked that he was going out the next weekend . I said, "Oh, did you sign?" He said, "Yes," with an odd grin. His real name,
we will say, was Edward W. Jones. What he had actually signed was
"W. Edward Jones ." The ward officials did not notice the difference.
It appeared to them that they had won a battle and had succe eded in
forcing him to act sanely. But to himself the message was, "He (the
real me) did not sign." He had won the battle. It was as if his finger s
were crossed b ehind his back.
All communicat ion has this characteristic— it can be magically
modified by accompanying communica tion. In this conference , we
have been discussing various ways of interacting with patients ,
describ ing what we do and what our strategy seems to us to be. It
would have been more difficult to discu ss our actions from the
patients' point of view. How do we qualify our communicat ions to
the patients, so that the exper ience which they receive will be
therapeu tic?
Appleby, for example, described a set of procedures on his
ward, and if I were a schizophrenic listening to him, I would have
been tempted to say, "It all sounds like occupational therapy to
me." He tells us very convincingly and with figures that his
program is successful, and in documenting his success he is no
doubt telling the truth. If this is so, then his description of the
program must necessarily be incomplete. The experiences which
the program provides for the patients must be something a little
235

more alive than the dry bones of the program which he has
described. The whole series of therapeutic procedures must have
been qualified, possibly with enthusiasm or with humor, with some
set of signals which altered the mathematical sign—plus or minus
—of what was being done. Appleby has told us only about the shoe
of the horse, not about the multitude of realities which determined
for what that horseshoe stood.
It is as if he had related that a given musical composition was
set in the key of C major, and asked us to believe that this skeletal
statement was a sufficient description to enable us to understand
why this particular composition altered the mood of the listener in
a particular way. What is omitted in all such descriptions is the
enormous complexity of modulation of communication. It is this
modulation which is music.
Let me shift from a musical to a wide biolog ical analogy in orde r
to examine further this magical realm of communication . All
organisms are partially determined by gene tics, i.e., by complex
const ellation s of messages carried principa lly in the chromosomes.
We are product s of a communicational process, modified and
qualified in various ways by environmental impact. It follows,
therefor e, that the differences between related organisms, say, a crab
and a lobster, or between a tall pea and a short pea, must alwa ys be
the sort of difference s that can be created by changes and
modulat ions in a constellation of messages . Sometimes these
changes in the message system will be relatively concrete—a shift
from "yes" to "no" in the answer to some question governing a
relatively superf icial detail of the anatomy. The total picture of the
animal may be altered by as little as one spot in the whole halftone
block , or the change may be one which modifies or modulates the
whole system of gene tic messages , so that every message in the
system takes on a different look while retaining its former relation –
ship to all neighbo ring messages. It is, I believe, this stability of the
relationship between messages unde r the impact of the change in
one part of the const ellation that provides a basis for the French
aphor ism "Plus get chang e, plus c'est la même chose." It is a
recogn ized fact that the skulls of the various anthropoids can be
drawn upon diversely skewed coordina tes 'to demonstr ate the
236

fundamental similar ity of relations and the systematic nature of the
transformation from one spe cies to anoth er.41
My father was a genet icist, and he used to say, "It's all
vibrations ,"42 and to illustrate this he would point out that the
striping of the common zebra is an octave higher than that of
Grev y's zebra. While it is true that in this particular case the
"frequenc y" is doubled , I don't think that it is entirely a matter of
vibrations as he ende avored to ex-pl ain it. Rather, he was trying to
say that it is all a matter of the sort of modifica tions which could be
expected among systems whose determinants are not a matter of
physics in the crude sense , but a matter of messages and modulated
systems of messages .
It is worth noting, too, tha t perhaps organic forms are beautiful to
us and the systematic biolog ist can find aesthetic satisfac tion in the
differences between related organisms simply because the
differences are due to modulation s of communication , while we
ourselve s are both organis ms who communicat e and whose forms
are determined by cons tellations of gene tic messages. This is not the
place, howeve r, for such a revision of aesthetic theory. An exper t in
the theory of mathematical groups could make a major contribut ion
in this field.
All messages and parts of messages are like phras es or segments
of equations which a mathematici an puts in bracke ts. Outside the
bracket s there may always be a qualifier or multiplier which will
alter the whole tenor of the phrase. More-ove r, these qualifiers can
always be added , even years later.
They do not have to precede the phrase inside the bracke ts.
Othe rwise , there could be no psychotherap y. The patient would be
entitled and even compelled to argue, "My mother slapp ed me down
in such and such ways, and, therefore, I am now sick; and because
those traumata occu red in the past they cannot be altered , and I,
therefore, cannot get well." In the realm of communication, the
events of the past constitute a chain of old horseshoes so that the
41 D. W. Thompson, On Growth and Form , Vol. 2, Ox-ford, Oxford Universit y
Press, 1952 .42 Beatrice C. Bateson, William Bateson , Naturalist , Cambridg e, Cambridg e
Universit y Press, 1928.
237

meaning of that chain can be changed and is continual ly being
changed . What exists today are only messages abou t the past which
we call memories, and these messages can al-ways be framed and
modulated from moment to moment.
Up to this point the realm of communication appea rs to be more
and more complex, more flexible, and less amenable to analysis.
Now the introduct ion of the group concep t—the consideration of
many person s—suddenl y simplifie s this confused realm of slipping
and sliding meanings. If we shak e up a number of irregula r stones in
a bag, or subjec t them to an almost rando m beating by the waves on
the seasho re, even at the crudely physical level, there will be a
gradua l simplification of the system—the stone s will resemble each
other. In the end, they will all become spherical, but in practice we
usual ly encount er them as partly rounded pebbles. Certain forms of
homogeniza tion result from multiple impact even at the crude
physical level, and when the impacting entities are organis ms
capab le of complex learning and communication, the total system
opera tes rapidl y to-ward either unifo rmity or toward systematic
differen tiation—an increase of simplicit y—which we call
organization. If there are differenc es between the impacting entities,
these differenc es will under go chang e, either in the direction of
reduc ing the difference , or in the direction of achiev ing a mutual
fitting or comple mentarity. Among groups of peop le, whether the
direction of change is toward homogenei ty or toward
complementarity, the a chiev ement is a s haring o f pre mises rega rding
the meaning and appropri atene ss of messages and other acts in the
contex t of the relationsh ip.
I shall not go into the complex problems of learning involved in
this proce ss but shall proceed to the problem of schizophren ia. An
individua l, i.e., the identified patient, exists within a family setting,
but when we view him singula rly, certain pecularities of his
communication al habits are no ted.
These pecu liarities may be partly determined by gene tics or
physiologica l accident , but it is still reasonab le to question the
function of these pecu liarities within the communicational system of
which they are a part the family. A number of living creatures
have been, in a sense, shaken up together and one of them has come
out appa rently differen t from the rest; we have to ask not only abou t
238

differences in the m aterial of which this par ticular ind ividua l may be
made, but also how his particula r characteristics were developed in
this family system. Can the pecu liarities of the identified patient be
seen as appropriate, i.e., as either homogeneous with, or
comple mentar y to, the charac teristics of the other members of the
group? We do not doub t that a large part of schizophr enic.
symptomatolog y is, in some sense , learned or determined by
experien ce, but an organis m can learn only that which it is taught by
the circumstance s of living and the expe rience s of exchang ing
messages with those around him. He cannot learn at rando m, but
only to be like or unlike those around him. We have, therefore , the
necessa ry task of looking at the experi ential setting of
schizophr enia.
We shall outline briefly what we have been calling the doub le
bind hypothes is, which has been more fully described elsewher e.43
This hypothes is contains two parts; a formal description of the
communicational habits of the schizophrenic, and a formal
description of the sequences of experience which would
understandably train the individual in his peculiar distortions of
communication. Empirically we find that one description of the
symptoms is, on the whole, satisfactory, and that the families of
schizophrenics are characterized by the behavioral sequences
which are predicted by the hypothesis.
Typically, the schizophr enic will eliminate from his messages
everything that refers explicitly or implicitly to the relation ship
43 G. Bateson, D. D. Jackson, J. Haley, and J. H. Weak-land, " Toward a Theory of
Schizophrenia ," Behavioral Science, 1956 , 1: 251–64; also G. Bateson , "Language
and Psychotherap y, Frieda Fromm-Reichmann 's Last Project ," Psychiatry , 1958 ,
21: 96–100; also G. Bateson (moderator), "Schizophrenic Distortions of
Communication," Psychotherapy of Chronic Schizophr enic Patients , C. A.
Whitacker, ed., Boston and Toronto, Little, Brown and Co., 1958 , pp. 31–56; also G.
Bateson , "Analysis of Group Therap y in an Admission Ward, United States Naval
Hospital , Oakland, California, " Social Psychiatry in Action, H. A. Wilmer,
Spring field, Ill., Charles C. Thomas, 1958 , pp. 334–49; also J. Hale y, "The Art of
Psychoanalysis," etc., 1958, .15: 190–200; also J. Haley, "An Interactional
Explanation of Hypnosis, " American Journal of Clinical Hypnosis , 1958, 1: 41–57;
also J. H. Weakland and D. D. Jackson, "Patient and Therapist Observations on the
Circumstances of a Schizophrenic Episode," AMA Archives of Neur ological
Psychiatry , 1958 , 79: 554–74.
239

between himself and the person he is addressing. Schizophren ics
commonly avoid the first and second person pronouns. They avoid
telling you what sort of a message they are transmitting—whether it
be literal or metaphor ic, ironic or direct and they are likely to
have difficulty with all messages and meaningful acts which imply
intimate contact between the self and some other. To receive food
may be almost impossib le, but so also may be the repud iation of
food.
When leaving for the A.P.A. meetings in Honolu lu, I told my
patient that I would be away and where I was going. He looked out
the window and said, "That plane flies awfully slowl y." He could
not say, "I shall miss you," because he would thus be identifying
himself in a relationship to me, or me in relationsh ip to himself. To
say, "I shall miss you" would be to assert a basic premise abou t our
mutual relationsh ip by defin ing the sorts of messages which should
be charac teristic of that relationship .
Observab ly, the schizophreni c avoids or distorts anything which
might seem to identify either himself or the person whom he is
addre ssing . He may eliminate anything which implies that his
message refers to, and is a part of, a relationship between two
identifiable people , with certain styles and premises gove rning their
behavio r in that relationship . He may avoid anything which might
enabl e the other to interpre t what he says. He may obscure the fact
that he is speaking in metaphor or in some special code, and he is
likely to distort or omit all referenc e to time and place. If we use a
Western Union telegram form as an analog y, we might say that he
omits what would be put on the procedural parts of the telegraph
form and will modify the text of his message to distort or omit any
indication of these metacommunicat ive elements in the total normal
message . What remains is likely to be a metaphor ic statement
unlabe lled as to contex t. Or, in extreme cases, there may be nothing
left but a stolid acting out of the message, "There is no relationship
between us."
This much is observable and may be summarized by saying that
the schizoph renic communica tes as if he expe cted to be punished
every time he indicates that he is right in his view of the contex t of
his own message .
240

The "double bind," which is central to the etiologica l half of our
hypothesis , may now simply be summarized by saying that it is an
experien ce of being punished precisely for being right in one's own
view of the contex t. Our hypothesis assumes that repeated
experien ce of punishment in sequences of this kind will lead the
individu al to behave h abitu ally as if he expect ed such punish ment.
The mother of one of our patients poured out blame upon her
husb and for refusing for fifteen years to hand over control of the
family finances to her. The father of the patient said, "I admit that it
was a great mistake of me not to let you handl e it, I admit that. I
have corrected that. My reasons for thinking it was a mistake are
entirely different from yours, but I admit that it was a very seriou s
error on m y part."
Mother: Now, you're just being facetious.
Father: No, I am not being facetiou s.
Mother: Well, anyway I don't care because when you come right
down to it the debts were incurred, still there is no reason why a
person would not be told of them. I think the woman should be told.
Father: It may be the same reason why when Joe (their psychotic
son) comes home from school and he has had trouble he doesn't tell
you.
Mother: Well, that's a good dodge.
The pattern of such a sequence is simply the succe ssive
disqual ification of each of the father's contributions to the
relationsh ip. He is continuousl y being told that the messages are not
valid. They are receiv ed as if they were in some way different from
that which h e though t he intended .
We may say that he is pena lized either for being right about his
views of his own intentions, or he is penal ized whenever his reply is
appropri ate to what sh e said.
But, per contra, from her viewpoint, it seems that he is endlessly
misinte rpreting her, and this is one of the most pecul iar
characteristics of the dynamic system which surrounds—o r is—
schizophr enia. Every therapist who has dealt with schizophren ics
will recogn ize the recurrent trap. The patient endeavors to put the
therap ist in the wrong by his interpre tation of what the therapist
said, and the patient does this because he expe cts the therapis t to
misinte rpret what he (the patient) said. The bind becomes mutual. A
241

stage is reached in the relationship in which neither person can
afford to receiv e or emit metacommunicat ive messages without
distortion.
There is, howeve r, usually, an asymmetry in such relationsh ips.
This mutual doubleb inding is a type of strugg le and commonly one
or the other has the upper hand . We have delibera tely chosen to
work with familie s where one of the offspring is the identified
patient, and, partly for this reason, in our data, it is the supposedl y
normal paren ts who have the upper hand over an identifiably
psychotic younger member of the group. In such cases, the
asymmetry takes the curious form that the identified patient
sacrifices h imself to m ainta in the sa cred i llusion th at what t he paren t
says makes sense. To be close to that parent, he must sacrifice his
right to indica te that he sees any metaco mmunicative
incong ruenci es, even when his perception of these incongruencie s is
correct. There is, therefore, a curious dispari ty in the distribu tion of
awarenes s of what is happening . The patient may know but must not
tell, and thereby enabl es the parent to not know what he or she is
doing . The patient is an accomplice in the parent's unconsciou s
hypocrisy. The result may be very grea t unhapp iness and very gross,
but al-ways systematic, distortion s of communication.
Moreov er, these distortions are alwa ys precisely those which
would seem approp riate when the victims are faced with a trap to
avoid which would be to destroy the very nature of the self. This
parad igm is neatly illustrated by a pas-s age which is worth quoting
in full from Festing Jon es' life of Samuel Butle r.44
Butle r went to dinne r at Mr. Seeboh m's wher e he met Skertchley,
who told them about a rat-trap invented by Mr. Tylor's coach man.
DUNK ETT'S RA T-TRAP
Mr. Dunkett found all his traps fail one after another, and was in
such despai r at the way the corn got eaten that he resolved to inven t
a rat-trap. He began by putting himself as nearly as possible in the
rat's place.
44 H. F. Jones, Samuel Butler: A Memoir, Vol. 1, Lon-don, Macmillan,
1919.
242

"Is there anything," he asked himself, "in which , if I were a rat, I
shou ld have such complete confidenc e that I could not suspe ct it
without suspec ting everything in the world and being unab le
hencefor th to move fearlessly in any direction? "
He pond ered for a while and had no answer, till one night the
room seemed to beco me full of light and he hears a voice from
heaven s aying:
"Dra in-pipes."
Then he saw his way. To suspec t a common drain-p ipe would be
to cease to be a rat. Here Skertchley enlarged a little, expl aining that
a spring was to be concealed inside , but that the pipe was to be open
at both ends; if the pipe were closed at one end, a rat would
naturally not like going into it, for he would not feel sure of being
able to get ou t again; on which I [But ler] interrupted and said:
"Ah, it was just this which stopped me from going in-to the
Church ."
When he [Butler] told me this I [Jones] knew what was in his
mind, and that, if he had not been in such respectab le company, he
would have said: "It was just this which stopped me from getting
married ."
Notice th at Dunket t could on ly invent t his doub le bind f or rats by
way of an hallucinato ry exper ience , and that both Butler and Jones
immediately regarded the trap as a paradig m for human relations.
Indeed, this sort of dilemma is not rare and is not confined to the
contexts of schizophren ia.
The quest ion which we must face, therefore , is why these
sequences are either specia lly frequent or specially destructive in
those families which conta in schizophren ics. I do not have the
statistics to assert this; howeve r, from limited but intense
observat ion of a few of these families, I can offer an hypothes is
abou t the group dynamics which would deter-mine a system of
interaction, such that doubl e bind expe riences must recur ad
nauseam. The probl em is to construct a model which will necessar ily
cycle to recreate these p atterned sequenc es ove r and over again.
243

Such a model is provided in Von Neumann's and Mor-
genstern's45 theory of games, presented here not, indeed, with its
full mathematical rigor, but at least in terms some-what technical.
Von Neumann was concerned with mathematical study of the
formal condi tions under which entities, with total intelligenc e and a
preference for gain, would form coalitions among themselves in
order to maximize the profits which coalition members might
receive at the expense of the non-m embers. He imagined these
entities as engaged in some-thing like a game and proce eded to ask
about the for mal characteristics of th e rules which would co mpel the
totally intelligent but gain-orien ted players to form coalitions. A
very curiou s conc lusion emerged, and it is this conc lusion which I
would p ropose a s a model.
Evidently, coalition between players can only emerge when
there are at least three of them. Any two may then get together to
exploit the third, and if such a game be symmetrically devised, it
evidently has three solutions which we may represent as
AB vs. C
BC vs. A
AC vs. B
For this three-pe rson system, Von Neumann demonstrates that
once formed, any one of these coalitions will be stable. If A and B
are in alliance, there is nothing C can do about it. And, interes tingly
enough, A and B will necess arily develop conv ention s
(supp lementar y to the rules) which will, for example, forbid them
from listening to C's app roaches .
In the five-person game, the position beco mes quite different;
there will be a variety of possib ilities. It may be that four players
conte mplate a combination against one, illustrated in the following
five patterns:
A vs. BCDE
B vs. ACDE
C vs. ABDE
45 J. Von Neumann and O. Morgenstern, Theory of Games and Economic
Behavior, Princeton, Princeton University Press, 1944.
244

D vs. ABCE
E vs. ABCD
But none of these would be stable. The four players within the
coalition must, neces sarily, engage in a subga me in which they
maneuver against each other to achieve an unequal division of the
gains which the coalition could squeez e out of the fifth player. This
must lead to a coal ition pattern which we may describe as 2 vs. 2 vs.
1, i.e., BC vs. DE vs. A. In such a si tuation, it would become poss ible
for A to approach and join one of these pairs, so that the coalition
system will become 3 vs. 2.
But in the system 3 vs. 2, it would be advantageous for the three
to recruit over to their side one of the two, in order to make their
gains more certain. Now we are back to a 4 vs. 1 system—not
necessa rily the particular line-up that we started from but at any rate
a system having the same general properties. It, in turn, must break
down i nto 2 vs. 2 vs. 1, and so on .
In other word s, for ever y possib le pattern of coalitions , there is a t
least one other pattern which will "dominate" it—to use Von
Neumann's term—and the relationship of domination between
solutions is intransi tive. There will al-ways be a circular list of
alternative solutions so that the system will never cease from
passing on from solution to solution , alwa ys selecting another
solution which is preferable to that which preceded it. This means,
in fact, that the robot s (owing to their total intelligence) will be
unab le to decide upon a single "play" of the game.
I offer this model as being reminiscent of what happ ens in
schizophr enic families . No two members seem able to get together
in a coalition stable enough to be decisive at the given moment.
Some other member or members of the family will alwa ys
intervene. Or, lacking such interven tion, the two members who
contemplate a coalition wi ll feel guilty vis-a-vis what the third might
do or say, and will draw b ack f rom the coalition.
Notice that it takes five hypothet ical entities with total
intelligence to achieve this particular sort of instability or oscillation
in a Von Neumannian game. But three human beings seem to be
enough . Perhap s they are not totally intelligen t or perhaps they are
245

systematica lly inconsis tent regarding the sort of "gain" in terms of
which t hey are motivated .
I want to stress that in such a system, the experience of each
separate individual will be of this kind: every move which he makes
is the common-sense move in the situation as he correctly sees it at
that moment, but his every move is subsequen tly demonstrated to
have been wrong by the moves which other members of the system
make in response to his "right" move. The individual is thus caught
in a perpetu al sequence of what we have called doub le bind
exper iences .
I do not know how valid this model may be, but I offer it for two
reasons. First, it is proposed as a sample of trying to talk about the
larger system—the family—instead of talking, as we habitually do,
about the individua l. If we are to unde rstand the dynamics of
schizophren ia, we must devise a language adequ ate to the
pheno mena which are emergent in this larger system. Even if my
model is inapprop riate, it is still worthwhile to try to talk in the sort
of language which we shall need for describing these emergent
pheno mena. Second ly, conceptua l models, even when incorrect, are
usefu l to the extent that criticism of the model may point to new
theoretical deve lopments.
Let me, therefore, point out one criticism of this model, and
consid er to what ideas it will lead. There is no theorem in Von
Neumann's book which would indicate that his entities or robots ,
engaged in this infinite dance of changing coalitions, would ever
beco me schizophr enic. Acco rding to the abstract theory, the entities
simply remain totally intelligen t ad infinitum.
Now, the major difference between people and von Neumann's
robot s lies in the fact of learning. To be infinitely intelligen t implies
to be infinitely flexib le, and the players in the dance which I have
described could never exper ience the pain which human beings
would feel if continu allyproven wrong whenever they had been
wise. Human beings have a commitment to the solutions which
they discover, and it is this psychological commitment that makes
it possible for them to be hurt in the way members of a
schizophrenic family are hurt.
It appears then, from consideration of the model, that the
double bind hypothesis, to be explanatory of schizophrenia, must
246

depend upon certain psychological assumptions about the nature of
the human individual as a learning organism. For the individual to
be prone to schizophrenia, individuation must comprise two
contrasting psychological mechanisms. The first is a mechanism of
adaptation to demands of the personal environment; and the
second, a process or mechanism whereby the individual becomes
either briefly or enduringly committed to the adaptations which the
first process has discovered.
I think that what l am calling a brief commitment to an
adaptation is what Bertalanffy called the immanent state of action;
and that the more enduring commitment to adaptation is simply
what we usually call "habit."
What is a person ? What do I mean when I say "I?" Perhaps what
each of us means by the "self" is in fact an aggregate of habits of
percept ion and adap tive action plus, from moment to moment, our
"immanent states of action." If somebody attacks the habits and
immanent states which chara cterize me at the given moment of
dealing with that somebod y—that is, if they attack the very habits
and immanent states which have been called into being as part of my
relationsh ip to them at that moment—the y are nega ting me. If I care
deeply about that other person, the nega tion of me will be still more
painful .
What we have said so far is enough to indicate the sorts of
strategy—or perhaps we should say symptoms—which are to be
expected in that strange institution, the schizophren ic family. But it
is still surpr ising to observe how these strateg ies may be continua lly
and habitually practiced without friends and neighbors noticing that
something is wrong . From theor y we may predi ct that every
participan t member o f such an institution must be def ensive of his or
her own immanent states of action and enduring adaptive habits;
protec tive, that is, of the self.
To illustrate with one example: a colleague had been work ing for
some weeks with one of these families , particularly with the father,
the mother , and their adult schizophr enic son. His meetings were on
the conjoint pattern—the members of the family being present
togethe r. This appar ently provoked some anxiety in the mother and
she reques ted face-to -face interviews with me. This move was
discussed a t the next con joint meeting and i n due cour se she ca me to
247

her first session. Upon arrival she made a coup le of conve rsationa l
remarks, and then opened her purse and from it handed me a piece
of paper, saying, "It seems my husband wrote this." I unfold ed the
paper and found it to be a single sheet of single-sp aced typescrip t,
starting with the words , "My husband and I much apprecia te the
opportun ity of discuss ing our proble ms with you," etc. The
docu ment then went on to outl ine cer tain spec ific ques tions which "I
would l ike to raise.
It appea red that the husband had, in fact, sat down at his
typewriter the night before and had written this letter to me as
though it were written by his wife, and in it he outlined the questions
for her to discu ss wi th me.
In normal daily life this sort of thing is common enough ; it
passe s muster . When attention is focused upon the characteristic
strateg ies, however , these self-pro tecting and self-destroying
maneuvers beco me conspicuous . One suddenl y discov ers that in
such familie s these strategies seem to pre-do minate over all others .
It becomes hardly surprising that the identified patient exhib its
behavio r which is almost a caricature of that loss of identity which i s
characteristic of all the family members.
I believe that this is the essence of the matter, that the
schizophrenic family is an organization with great ongoing
stability whose dynamics and inner workings are such that each
member is continually undergoing the experience of negation of
self.
248

Minimal Require ments f or a Theory of
Schizophr enia*
Every scienc e, like eve ry person , has a du ty toward i ts neighbors ,
not perhaps to love them as itself, but still to lend them its tools, to
borrow tools from them, and, generally, to keep the neighbo ring
science s straigh t. We may perhap s judge of the importance of an
advance in any one science in terms of the changes which this
advance co mpels the neigh -boring scien ces to m ake in the ir methods
and in their thinking. But always there is the rule of parsimony. The
changes which we in the behav ioral sciences may ask for in
genetics, or in philosoph y, or in information theory must always be
minimal. The unity of science as a whole is achieved by this system
of minimal demands imposed by each science upon its neighbo rs,
and—no t a little—b y the lending of concep tual tools and patterns
which occu rs among the variou s sciences .
My purpose , therefore, in the present lecture is not so much to
discuss the particular theory of schizoph renia which we have been
developing at Palo Alto. Rather, I want to indicate to you that this
theory and others like it have impact upon ideas about the very
nature of explana tion. I have used the title "Minimal Requir ements
for a Theory of Schizophren ia," and what I had in mind in choo sing
this title was a discuss ion of the implic ation s of the doub le bind
theory for the wider field of behav ioral scienc e and even , beyond
that, its effect upon evolutionar y theory and biologica l episte-
mology. What minimal changes does this theory demand in related
science s?
I want to deal with question s about the impact of an expe riential
theory of schizophr enia upon that triad of related scienc es, learning
theory, gene tics, and evo lution.
* Second Annual Albert D. Lasker Memorial Lecture, delivered at the Institute for
Psychoso matic and Psychiatric Research and Training of the Michael Reese
Hospital , Chica go, April 7, 1959. This lecture is here reprinted by permission of the
A.M.A. Archives of General Psychiatry where it appeared in 1960, Vol. 2, pp.
477-491.
249

The hypothes is may first be briefly descr ibed. In its essentials,
the idea appe als only to everyday experience, and elementary
common sense. The first propos ition from which the hypothesis is
derived is that learning occu rs alwa ys in some context which has
formal charac teristics. You may think, if you will, of the formal
characteristics of an instrumental avoid ance sequence , or of the
formal charac teristics of a Pavlovian experiment. To learn to lift a
paw in a Pavlovi an context is differen t from learning the same
action in a context of instrumental reward .
Furthe r, the hypothesis depends upon the idea that this structu red
contex t also occurs within a wider context—a metacon text if you
will—and that this sequence of contexts is an open , and conce ivabl y
infinite, series.
The hypothesis also assumes that what occu rs within the narrow
contex t (e.g., instrumental avoidance ) will be affected by the wider
contex t within which this smaller one has its being . There may be
incong ruence or conflict between contex t and metacon text. A
contex t of Pavlovian learning may, for example, be set within a
metacon text which would punish learning of this kind, perhaps by
insisting upon insight. The organism is then faced with the dilemma
either of being wrong in the pr imary context or of being right for the
wrong reasons or in a wrong way. This is the so-called double bind.
We are inves tigating the hypothesis that schizophr enic
communication is learned and be-comes habitual as a result of
continual traumata o f this kind.
That is all there is to it.
But even these "common-sen se" assumptions break away from
the classical rules of scientific epistemology. We have learned from
the paradig m of the freely falling body—and from many similar
parad igms in many other scienc es—to approach scientific proble ms
in a peculiar way: the proble ms are to be simplified by ignoring—or
postpon ing consider ation of—the possibility that the larger contex t
may influen ce the smalle r. Our hypothesis runs counte r to this rule,
and is focused precisely upon the determining relations between
larger and s maller contexts .
Even more shocking is the fact that our hypothesis sugge sts —
but does not stand or fall with the sugg estion—tha t there may be an
infinite regres s of such relevant contexts.
250

In all of this, the hypothes is require s and reinforc es that revision
in scientific thought which has been occu rring in many fields, from
physics to biolog y. The observ er must be included within the focus
of observation, and what can be studied is always a relationship or
an infinite regress of relationsh ips. Never a "thing ."
An example will make clear the relevance of the larger contex ts.
Let us consider the larger contex t within which a learning
experiment might be conducted using a schizoph renic as a subjec t.
The schizophren ic is what is called a patient, vis-a-vis a member of
a superio r and unlov ed organization, the hospi tal staff. If the patient
were a good pragmatic Newtonian , he would be able to say to
himself: "The cigarettes which I can get by doing what this fellow
expects me to do are after all only cigarettes, and as an applied
scientist I will go ahead and do what he want s me to do. I will solve
the experimental proble m and obtain the cigarettes." But human
beings, and espec ially schizoph renics, do not alwa ys see the matter
this way. They are affected by the circumstance that the experiment
is being conducted by somebody whom they would rather not
please . They may even feel that there would be a certain
shamelessnes s about seeking to pleas e some one whom they dislike.
It thus comes about that the sign of the signa l which the
experimenter emits, giving or withholding cigarettes, is reversed .
What the exper imenter thought was a reward turns out to be a
message of partial indigni ty, and what the experimenter thought was
a punish ment beco mes in part a sourc e of satisfaction.
Consid er the acute pain of the mental patient in a large hospital
who is momentarily treated as a human being by a member of the
staff.
To explain the observed pheno mena we always have to conside r
the wider context of the learning exper iment, and every transac tion
between pe rsons is a con text of learning .
The doub le bind hypothesis , then, depends upon attributing
certain characte ristics to the learning process. If this hypothesis is
even approximately true, roo m must be made for it wi thin the theo ry
of learning. In particular, learning theory must be made
discont inuous so as to accommodate the discon tinuities of the
hierarchy of the contexts of learning to whi ch I have referred.
251

Moreover, these discontinuities are of a peculiar nature. I have
said that the larger context may change the sign of the
reinforcement proposed by a given message, and evidently the
larger context may also change the mode—may place the message
in the category of humor, metaphor, etc. The setting may make the
message inappropriate. The message may be out of tune with the
larger context, and so on. But there are limits to these
modifications. The context may tell the recipient anything about
the message, but it cannot ever destroy or directly contradict the
latter. "I was lying when I said `The cat is on the mat' " tells the
vis-a-vis nothing about the location of the cat. It tells him only
something about the reliability of his previous information. There
is a gulf between context and message (or between metamessage
and message) which is of the same nature as the gulf between a
thing and the word or sign which stands for it, or between the
members of a class and the name of the class. The context (or
metamessage) classifies the message, but can never meet it on
equal terms.
In order to fit these discontinuities into learning theory, it is
necessary to enlarge the scope of what is to be included within the
concept of learning . What the experimenters have described as "learning" are in general changes in what an organism does in
response to a given signal. The experimenter observes, for
example, that at first the buzzer evokes no regular response, but
that after repeated trials in which the buzzer has been followed by
meat powder, the animal will begin to salivate whenever it hears
the buzzer. We may say loosely that the animal has begun to attach
significance or meaning to the buzzer.
A change has occur red. In order to construct a hierarch ic series,
we pick on the word "chang e." Series such as we are interested in
are in gener al built in two ways. Within the field of pure
communication s theory, the steps of an hierarchic series may be
const ructed by success ive use of the word "abou t," or "meta." Our
hierarch ic series will then consi st of message, metamessage, meta-
metamessage , and so on. Where we deal with phenomena marginal
to communicat ions theory, similar hierarch ies may be constructed
by the piling up of "change" upon "change." In classic al physics, the
sequenc e: position; velocity (i.e., change in position) ; acceleration
252

(i.e., change in veloc ity or change in change of position); chang e of
accele ration, etc., is an example of such a hierarchy.
Further complication s are added—rar ely in classical physics but
commonly in human communica tion—b y noting that messages may
be about (or "meta" to) the relationsh ip between messages of
different levels. The smell of the experimental harness may tell the
dog that the buzzer will mean meat powder . We will then say that
the message of the harness is meta to the message of the buzze r. But
in human relations another sort of complexi ty may be generated;
e.g., messages may be emitted forbidding the subjec t to make the
meta connect ion. An alcohol ic paren t may punish a child for
showing that he knows that he should look out for storms whenever
the parent gets the bottle out of the cupboa rd. The hierarchy of
messages and contexts thus beco mes a complex branching s tructure .
So we can construct a similar hierarchic classification within
learning theo ry in subs tantially the sa me way as th e physicists. What
the experimenters have investigated is chang e in the receip t of a
signal . But, clearly, to receive a signa l already denot es change—a
change of a simpler or lower order than that which the
experimenters have investigated. This gives us the two first steps in
a hierarch y of learning , and above these an infinite series can be
imagined. This hierarch y46 can now be laid ou t as follows :
(1) The Receip t of a Signal I am work ing at my desk on which
there is a pape r bag, containing my lunch. I hear the hosp ital
whis tle, and from this I know that it is twelve o'clock . I reach out
and take my lunch . The whist le may be regarded as an answer to a
question laid down in my mind by previous learning of the second
order; but the single event—the receiving of this piece of
information—is a piece of learning, and is demonstrated to be so
by the fact that having received it, I am now changed and respond
in a special way to the paper bag.
(2) Those Learning s Which Are Changes in (1) These are
exemplified by the classical learning expe riments of various kinds :
46 ' 1971. In my final version of this hierarch y of orders of learnin g, published in
this volum e as "The Logical Categories of Learning and Communication," (see p.
283) I have used a different system of numbering. The receipt of a signal is there
called "Zero Learning "; changes in Zero Learning are called Learning I; "deutero-
learning " is called Learnin g II, etc
253

Pavlov ian, instrumental reward, instrumental avoidance , rote, and so
on.
(3) Those Learnings Which Const itute Changes in Second-Or der
Learning I have in the past, unfortuna tely, called these pheno mena
"deut ero-l earning," and have translated this as "learn ing to learn." It
would have been more correct to coin the word trito-learning and to
transla te it as "learn ing to learn to receive signals ." These are the
pheno mena in which the psychiatrist is preponde rantly interested ,
namely, the chang es whereb y an individua l comes to expect his
world to be structured in one way rather than an-othe r. These are the
pheno mena which under lie "transference"—th e expe ctation on a
patient's part that the relationsh ip with the therapis t will conta in the
same sorts of contex ts of learning that the pat ient has prev iousl y met
with in dealing wi th his parents .
(4) Changes in Those Processes of Change Refer red to in (3)
Whether learning of this fourth order occurs in human being s is
unknown. What the psychothe rapis t attempts to produ ce in his
patient is usually a third-order learning , but it is possib le, and
certainly conce ivable , that some of the slow and unconsc ious
changes may be shifts in sign of some higher derivative in the
learning proce ss.
At this point it is necessary to compare three types of hierarch y
with which we are faced : (a) the hierarch y of orders of learning ; (b)
the hierarch y of contex ts of learning, and (c) hierarchies of circuit
structu re which we may—indeed, must—expect to find in a
telencepha lized br ain.
It is my contention that (a) and (b) are synonymous in the sense
that all statements made in terms of contex ts of learning could be
transla ted (without loss or gain) into statements in terms of orde rs of
learning , and, further, that the classification or hierarchy of contex ts
must be isomorphic with the classification or hierarch y of orders of
learning . Beyond this, I believe that we shou ld look forward to a
classification or hierarchy of neurophysiolog ical structures which
will be isomorphic wi th the other two c lassification s.
This synonymy between statements about contex t and statements
about orders of learning seems to me to be self-eviden t, but
exper ience shows that it must be spelled out. "The truth canno t be
254

said so as to be unde rstood , and not be believed," but, conv ersely, it
cannot b e believed until it is said so as to be under stood .
It is necess ary first to insist that in the world of communication
the only relevan t entities or "realities" are messages, including in
this term parts of messages , relation s between messages, significant
gaps in m essages , and so on. The perception of an event or objec t or
relation is real. It is a neuroph ysiolog ical message . But the 'even t
itself or the object itself cannot enter this world and is, therefo re,
irrelevan t and, to that exten t, unreal. Converse ly, a message has no
reality or relevance qua message, in the Newtonian world: it there is
reduced to sound waves or printer's ink .
By the same token , the "contex ts" and "contexts of con-t exts"
upon which I am insisting are only real or relevant insofar as they
are communicat ional ly effective, i.e., function as messages or
modifiers of messages .
The differenc e between the Newtoni an world and the world of
communica tion is simply this: that the Newtonian world ascribes
reality to objects and achieves its simplicity by exclud ing the
context of the context— exclud ing indeed all metarel ation ships—a
fortiori excluding an infinite regress of such relations . In contrast,
the theorist of communication insists upon examining the
metarel ation ships while achieving its simplicity by exclud ing all
objects.
This world, of communication , is a Berkel eyan world, but the
good bishop was guilty of unde rstatement. Relevance or reality
must be denied not only to the sound of the tree which falls unhea rd
in the forest but also to this chai r which I can see and on which I am
sitting. My percep tion of the chair is communica tiona lly real, and
that on which I sit is, for me, only an idea, a message in which I put
my trust.
"In my though t, one thing is as good as ano ther in this world , and
the shoe of a horse will do," because in though t and in experience
there are no things , but only messages and the like.
In this world, indeed, I, as a material objec t, have no relevanc e
and, in this sense, no reality. "I," however , exist in the
communica tiona l world as an essent ial element in the syntax of my
experien ce and in the expe rienc e of others, and the communication s
255

of others may damage my identity, even to the point of breaking up
the organization o f my expe rienc e.
Perhaps one day, an ultimate synthesis will be achieved to
combine the Newtonian and the communicational worlds. But that
is not the purpose of the present discussion. Here I am concerned
to make clear the relation between the con-texts and the orders of
learning, and to do this it was first necessary to bring into focus the
difference between Newtonian and communicational discourse.
With this introductory statement, however, it becomes clear that
the separation between contexts and orders of learning is only an
artifact of the contrast between these two sorts of discourse. The
separation is only maintained by saying that the contexts have
location outside the physical individual, while the orders of
learning are located inside. But in the communicational world, this
dichotomy is irrelevant and meaningless. The contexts have
communicational reality only insofar as they are effective as
messages, i.e., insofar as they are represented or reflected
(correctly or with distortion) in multiple parts of the
communicational system which we are studying; and this system is
not the physical individual but a wide network of pathways of
messages. Some of these pathways happen to be located outside
the physical individual, others inside; but the characteristics of the
system are in no way dependent upon any boundary lines which
we may superpose upon the communicational map. It is not
communicationally meaningful to ask whether the blind man's
stick or the scientist's microscope are "parts" of the man who uses
them. Both stick and microscope are important pathways of
communication and, as such, are parts of the network in which we
are interested; but no boundary line—e.g., halfway up the stick—
can be relevant in a description of the topology of this net.
However, this discarding of the boundary of the physical
individual does not imply (as some might fear) that com-
municational discourse is necessarily chaotic. On the contrary, the
proposed hierarchic classification of learning and/or context is an
ordering of what to the Newtonian looks like chaos, and it is this
ordering that is demanded by the double-bind hypothesis.
256

Man must be the sort of animal whose learning is characterized
by hierarchic discontinuities of this sort, else he could not become
schizophrenic under the frustrations of the double bind.
On the evidential side, there is beginning to be a body of
experiment demonstrating the reality of third-order learning47; but
on the precise point of discontinuity between these orders of
learning there is, so far as I know, very little evidence. The
experiments of John Stroud are worth quoting. These were tracking
experiments. The subject is faced with a screen on which a spot
moves to represent a moving target. A second spot, representing
the aim of a gun, can be controlled by the subject, who operates a
pair of knobs. The subject is challenged to maintain coincidence
between the target spot and the spot over which he has control. In
such an experiment the target can be given various sorts of motion,
characterized by second-, third-, or higher-order derivatives.
Stroud showed that, as there is a discontinuity in the orders of the
equations which a mathematician might use to describe the
movements of the target spot, so also there is a discontinuity in the
learning of the experimental subject. It is as if a new learning
process were involved with each step to a higher order of
complexity in the movement of the target.
It is to me fascinating to find that what one had supposed was a
pure artifact of mathematical description is also apparently an
inbuilt characteristic of the human brain, in spite of the fact that
this brain certainly does not operate by means of mathematical
equations in such a task.
There is also evidence of a more genera l nature which would
suppo rt the notion of discontinu ity between the order s of learning .
There is, for example, the curious fact that psychologis ts have not
habitually regarded what I call learning of the first order, the receip t
of a meaningfu l signal , as learning at all; and the other curious fact,
that psycholog ists have until recen tly shown very little apprec iation
of that third order of learning, in which the psychiatrist is predomi-
nantly interested. There is a formidable gulf between the think ing of
47 C. L. Hull, et al., Mathematico-d eductive Theory of Rote Learning: A Study
in Scientific Methodology , (Yale Universit y Institute of Human Relations) , New
Haven, Yale University Press, 1940; also H. F. Harlow, "The Formation of
Learnin g Sets," Psychol. Review, 1949, 56: 51-65.
257

the experimental psychologis t and the thinking of the psychiatrist or
anthropologi st. This gulf I believe to be due to the discont inuity in
the hierarch ic structure .
Learning , Genet ics, and Evo lution
Before we consider the impact of the double bind hypothesis
upon genet ics and e volut ionar y theory, it is nece ssary to exa mine th e
relationship between theories of learning and these two other bodies
of knowl edge. I referred earlier to the three subjects together as a
triad. The structure of this triad we m ust now cons ider.
Geneti cs, which covers the communicat ional pheno mena of
variation, differentiation, growth , and heredity, is commonly
recogn ized as the very stuff of which evolu tiona ry theory is made.
The Darwinian theory, when purged of Lamarckian ideas, consis ted
of a gene tics in which variation was presumed to be rando m,
combined with a theory of natural selection would impart adaptiv e
direction to the accumulation of changes. But the relation between
learning and this theory has been a matter of violen t controversy
which has raged over the so-called "inhe ritance of acquir ed
characteristics."
Darwin's position was acutely challenged by Samuel Butler, who
argued that heredi ty should be compared with—ev en identified with
—memory. Butle r proce eded from this premise to argue that the
proce sses of evolutionar y chang e, and especia lly adap tation, should
be regard ed as the achievements of a deep cunn ing in the ongo ing
flow of life, not as fortuitous bonuses confe rred by luck. He drew a
close analogy between the pheno mena of inven tion and the phe-
nomena of evolu tiona ry adap tation, and was perhaps the first to
point out the existen ce of residua l organs in machines. The curious
homology whereb y the engine is located in the front of an
automobile, wher e the horse used to be, would have delighted him.
He also argued very cogently that there is a proces s whereb y the
newer invention s of adap tive behavior are sunk deepe r into the
biologica l system of the organism. From planned and consciou s
actions they become habits, and the habits beco me less and less
258

conscious and less and less subject to volunt ary control. He
assumed, with-out evidence , that this habitualization, or sinking
process , could go so deep as to contr ibute to the body of memories,
which we would call the genotype, and which determine the
characteristics of the nex t gen eration.
The controve rsy about the inheritance of acquir ed characteristics
has two facets. On the one hand, it appears to be an argument which
could be settled by factual materia l. One good case of such
inheritance might settle the matter for the Lamarckian side. But the
case against such inher itance, being negative, can never be proved
by evidenc e and must rely upon an appea l to theory. Usual ly those
who take the nega tive view argue from the separation between germ
plasm and somatic tissue, urging that there can be no systematic
communica tion from the soma to the germ plasm in the light of
which the geno type might revise itself.
The difficulty looks like this: conceivab ly a biceps muscle
modified by use or disuse might secrete spec ific metaboli tes into the
circulation, and these might conce ivabl y serve as chemical
messengers from muscle to gonad . But (a) it is difficult to believe
that the chemistr y of biceps is so different from that of, say, triceps
that the message could be specific, and (b) it is difficult to believe
that the gonad tissue could be equipped to be appropr iately affected
by such messages. After all, the receiver of any message must know
the code of the sende r, so that if the germ cells are able to receive
the messages from the somatic tissue, they must alread y be carrying
some version of the somatic code . The directions which
evolutionar y change could take with the aid of such messages from
the so ma would have to be prefigured in the germ plasm.
The case against the inheritance of acquir ed chara cteristics thus
rests upon a separ ation , and the differenc e between the schoo ls of
thought crystallizes around philosophic reactions to such a
separa tion. Those who are willing to think of the world as organized
upon multipl e and separable principles will accept the notion that
somatic changes induced by environment may be cove red by an
explanation which could be totally separ ate from the explanation of
evolutionar y change . But those who prefer to see a unity in nature
will hope that these two bodies of explanat ion can someho w be
interrelated.
259

Moreover, the whole relationship between learning and
evolution has undergone a curious change since the days when
Butler maintained that evolution was a matter of cunning rather
than luck, and the change which has taken place is certainly one
which neither Darwin nor Butler could have foreseen. What. has
happened is that many theorists now assume learning to be
fundamentally a stochastic or probabilistic affair, and indeed, apart
from nonparsimonious theories which would postulate some
entelechy at the console of the mind, the stochastic approach is
perhaps the only organized theory of the nature of learning. The
notion is that random changes occur, in the brain or else-where,
and that the results of such random change are selected for survival
by processes of reinforcement and extinction. In basic theory,
creative thought has come to resemble the evolutionary process in
its fundamentally stochastic nature. Reinforcement is seen as
giving direction to the accumulation of random changes of the
neural system, just as natural selection is seen as giving direction
to the accumulation of random changes of variation.
In both the theory of evolution and the theory of learning,
however, the word "random" is conspicuously undefined, and the
word is not an easy one to define. In both fields, it is assumed that
while change may be dependent upon probabilistic phenomena, the
probability of a given change is determined by something different
from probability. Underlying both the stochastic theory of
evolution and that of learning, there are unstated theories regarding
the determinants of the probabilities in question.48 If, however, we
ask about change in these determinants, we shall again be given
stochastic answers, so that the word "random," up-on which all of
these explanations turn, appears to be a word whose meaning is
hierarchically structured, like the meaning of the word "learning,"
which was discussed in the first part of this lecture.
Lastly, the question of the evolutionar y funct ion of acquired
characteristics has been reopened by Wadding ton's work on
phenocopi es in Drosophila. At the very least, this work indicates
that the changes of pheno type which can be achiev ed by the
48 In this sense , of course, all the theories of change assume that the next change
is in some degree prefigu red in the system which is to undergo that change.
260

organis m under environmental stress are a very importan t part of the
machiner y by which the specie s or hereditary line maintains its
place in a stressfu l and competitive environment, pend ing the later
appearanc e of some mutation or other gene tic chang e which may
make the species or line better able to deal with the ongoing stress.
In this sense at least, the acquired characteristics have important
evolutionar y function. Howeve r, the actual experimental story
indica tes something more than t his and is worth r eproduc ing b riefly.
What Waddington works with is a pheno copy of the phenot ype
brought abou t by the gene bithor ax. This gene has very profound
effects upon the adult phenotype. In its presence the third segment
of the thorax is modified to resemble the second , and the little
balancing organs, or halteres , on this third segment beco me wing s.
The result is a four-winged fly. This four-wing ed chara cteristic can
be produced artificially in flies which do not carry the gene bitho rax
by subjec ting the pupae to a period of intoxication with ethyl ether.
Waddington works with large popula tions of Drosophila flies
derived from a wild strain believed to be free of the gene bithorax.
He subjec ts the pupae of this popula tion in success ive genera tions to
the ether treatment, and from the resulting adults selects for
breeding those which show the best approximation to bithorax . He
has continued th is exper iment over m any gene rations, and a lready in
the twen ty-seventh genera tion he finds that the bithor ax appe arance
is achieved by a limited number of flies whose pupae were
withdrawn from the experi mental treatment and not subjected to
ether. Upon breeding from these, it turns out that their bitho rax
appearanc e is not due to the presence of the specific gene, bithor ax,
but is due to a conste llation of genes which work together to give
this effect.
These very striking results can be read in various ways. We can
say that in selecting the best phenocop ies, Wadding -ton was in fact
selecting for a genetic potentiality for achieving this phenot ype. Or
we can say that he was selecting to reduce the thresho ld of ether
stress nec essar y to produce this result.
Let me sugg est a possible model for the description of these
pheno mena. Let us suppo se that the acquir ed characteristic is
achieved by some process of funda mentally stocha stic nature—
perhaps some sort of somatic learning—and the mere fact that
261

Wadding ton is able to select the "best" phenocopie s would lend
suppor t to this assumption. Now, it is evident that any such proces s
is, in the natu re of the case , waste ful. To achieve a result by trial and
error which could have been achieved in any more direct way
neces sarily consu mes time and effort in some sense of these words .
Insof ar as we think of adapt ability as achieved by stochas tic process,
we let in the no tion of an econo mics of adap tability.
In the field of mental process, we are very familiar with this
sort of economics, and in fact a major and necessary saving is
achieved by the familiar process of habit formation. We may, in the
first instance, solve a given problem by trial and error; but when
similar problems recur later, we tend to deal with them more and
more economically by taking them out of the range of stochastic
operation and handing over the solutions to a deeper and less
flexible mechanism, which we call "habit." It is, therefore,
perfectly conceivable that some analogous phenomenon may
obtain in regard to the production of bithorax characteristics. It
may be more economical to produce these by the rigid mechanism
of genetic determination rather than by the more wasteful, more
flexible (and perhaps less predictable) method of somatic change.
This would mean that in Waddington's population of flies there
would be a selective benefit for any hereditary line of flies which
might contain appropriate genes for the whole—or for some part—
of the bithorax phenotype. It is also possible that such flies would
have an extra advantage in that their somatic adaptive machinery
might then be available for dealing with stresses of other kinds. It
would appear that in learning, when the solution of the given
problem has been passed on to habit, the stochastic or exploratory
mechanisms are set free for the solution of other problems, and it
is quite conceivable that a similar advantage is achieved by passing
on the business of determining a somatic characteristic to the gene-
script49
49 These considerations alter somewhat the old problem of the evolutionar y effect
of use and disuse . Orthodox theor y could only suggest that a mutation reducin g the
(potential) size of a disused organ had survi val value in terms of the resulting
economy of tissue. The present theory would suggest that atrophy of an organ,
occurring at the somatic level, may constitute a drain upon the total available
adaptabilit y of the organism, and that this waste of adaptability might be saved if
262

It may be noted that such a model would be characterized by two
stochas tic mechanis ms: first, the more superficial mechanism by
which the change s are achieved at the somatic level, and, second ,
the stochast ic mechanis m of mutation (or the shuffling of gene
constellations) at the chromosomal level. These two stochast ic
systems will, in the long run under selective conditions, be
compelled to work together, even though no message can pass from
the more super ficial somatic system to the germ plasm. Samuel But-
ler's hunch that something like "habi t" might be crucial in evolution
was perhaps not too wide o f the m ark.
With this introduct ion we can now proceed to look at the
proble ms which a double bind theory of schizophreni a would pose
for the geneticist.
Genetic Prob lems Posed by D ouble Bind The ory
If schizoph renia be a modificat ion or distortion of the learning
process , then when we ask about the gene tics of schizoph renia, we
cannot be content just with gene alogi es upon which we discriminate
some indiv idual s who have been committed to hospi tals, and others
who have not. There is no a priori expec tation that these distortions
of the learning proces s, which are high ly formal and abstract in their
nature, will necess arily appear with that approp riate content which
would r esult in hosp ital commitment. Our t ask as gen eticists wil l not
be the simple one upon which the Mendel ians concentra ted,
assuming a one-t o-one relation between phenotype and genot ype.
We canno t simply assume that the hospitalized members carry a
gene for schizophrenia and that the others do not. Rather, we have
to expect that several genes or constellations of genes will alter
patterns and potentialities in the learning process, and that certain
of the resultant patterns, when confronted by appropriate forms of
environmental stress, will lead to overt schizophrenia.
In the most general terms, any learning, be it the absorption of
one bit of information or a basic change in the character structure
reduction of the organ could be achieved mo re directl y by genetic determinants.
263

of the whole organism, is, from the point of view of genetics, the
acquisition of an "acquired characteristic." It is a change in the
phenotype, of which that phenotype was capable thanks to a whole
chain of physiologic and embryologic processes which lead back
to the genotype. Every step in this backward leading series may
(conceivably) be modified or interrupted by environmental
impacts; but, of course, many of the steps will be rigid in the sense
that environmental impact at that point would destroy the
organism. We are concerned only with those points in the hierarchy
at which environment can take effect and the organism still be
viable. How many such points there may be we are far from
knowing. And ultimately, when we reach the genotype, we are
concerned to know whether the genotypic elements in which we
are interested are or are not variable. Do differences occur from
genotype to genotype which will affect the modifiability of the
processes leading to the phenotypic behaviors which we observe?
In the case of schizophrenia we deal evidently with a relatively
long and complex hierarchy; and the natural history of the disease
indicates that the hierarchy is not merely a chain of causes and
effects from gene-script to phenotype, which chain becomes at
certain points conditional upon environmental factors. Rather, it
seems that in schizophrenia the enviromental factors themselves
are likely to be modified by the subject's behavior whenever
behavior related to schizophrenia starts to appear.
To illustrate these complexit ies, it is perhap s worthwhil e to
consid er for a moment the genetic proble ms presen ted by other
forms of communica tiona l behav ior—hu mor, mathematical skill, or
musical composition. Perhaps in all these cases, there are
consid erable genetic differences between indiv iduals in thos e factor s
which make for an ability to acqui re the appropr iate skills. But the
skills themselves and their particular expression also depend largely
upon environmental circumstances and even upon specific training.
In addition, however , to these two components of the situation, there
is the fact that the individua l who shows ability, e.g., in musical
composit ion, is likely to mold his environment in a direction which
will favor his develop ing his ability, and that he will, in turn, create
an environment for others which will favor their develop ment in the
same direction.
264

In the case of humor, the situation may even be one degree
more complicated. It is not clear that in this case the relationship
between humorist and his human environment will necessarily be
symmetrical. Granted that in some cases the humorist promotes
humor in others, in many other cases there occurs the well-known
complementary relationship between humorist and "straight" man.
Indeed, the humorist, insofar as he hogs the center of the stage,
may reduce others to the position of receiving humor but not
themselves contributing.
These considerations can be applied unchanged to the problem
of schizophrenia. Anybody watching the trans-actions which occur
between the members of a family containing an identified
schizophrenic will perceive immediately that the symptomatic
behavior of the identified patient fits with this environment and,
indeed, promotes in the other members those characteristics
which.evoke the schizophrenic behavior. Thus, in addition to the
two stochastic mechanisms outlined in the previous section, we
now face a third, namely the mechanism of those changes whereby
the family, perhaps gradually, becomes organized (i.e., limits the
behaviors of the component individuals) in such a way as to fit the
schizophrenia.
A question which is frequent ly asked is this: "If this family is
schizophr enogenic , how does it happen that all of the siblings are
not diagnosabl e as schizophren ic patients? " Here it is necessa ry to
insist that the family, like any other organization, creates and
depends upon differen tiation among its members. As in many
organiza tions , there is room only for one boss, in spite of the fact
that the organization operates upon those premises which would
induce administrative skill and ambition in its members; so also in
the schizophrenogen ic family there may be room for only one
schizophr enic. The case of the humorist is quite comparabl e. The
organiza tion of the Marx family, which could create four
profess ional humorists, must have been quite exceptional. More
usually one such individua l would suffice to re-duc e the others to
more commonplace behavior al roles. Gene-ti cs may play a role in
deciding which of severa l siblings shall be the schizophren ic—or
which shall be the clown—but it is by no means clear that such
265

hered itary factors could complete ly determine the evolut ion or roles
within the family organiza tion.
A second question—to which we have no final answer—
concerns the degree of schizophrenia (genetic and/or acquired)
which must be assigned to the schizophrenogenic parent. Let me,
for purposes of the present inquiry, define two degrees of
schizophrenic symptomatology, and note that the so-called
"psychotic break" sometimes divides these two degrees.
The more serious and conspicuous degree of symptomatology is
what is conventionally called schizophrenia. I will call it "overt
schizophrenia." The persons so afflicted be-have in ways which are
grossly deviant from the cultural-environment. In particular, their
behavior seems characterized by conspicuous or exaggerated errors
and distortions regarding the nature and typing of their own
messages (internal and external), and of the messages which they
receive from others. Imagination is seemingly confused with
perception. The literal is confused with the metaphoric. Internal
messages are confused with external. The trivial is confused with
the vital. The originator of the message is confused with the
recipient and the perceiver with the thing perceived. And so on. In
general, these distortions boil down to this: that the patient behaves
in such a way that he shall be responsible for no
metacommunicative aspect of his messages. He does this,
moreover, in a manner which makes his condition conspicuous: in
some cases, flooding the environment with messages whose logical
typing is either totally obscure or misleading; in other cases,
overtly withdrawing to such a point that he commits himself to no
overt message.
In the "cover t" case the behav ior of the identified patient is
similar ly but less conspicuous ly characterized by a continual
changing of the logica l typing of his or her messages, and a
tenden cy to respond to the messages of others (especially to those of
other family members) as though these were of logical type,
differen t from that which the speake r intended. In this system of
behavio r the messages of the vis-a-vis are continual ly disqual ified,
either by indicating that they are inappropr iate replies to what the
cover t schizophrenic has said or by indicating that they are the
produc t of some fault in the character or motivat ion of the speaker .
266

Moreove r, this destructive behavio r is in genera l maintained in such
a way as to be undetected. So long as the cove rt schizophreni c can
succeed in putting the other in the wrong, his or her patho logy is
obscured and the blame falls else-where . There is some evidence
that these pe rsons fe ar col -lapse into overt s chizoph renia when f aced
by circumstances which would force them to recognize the pattern
of their opera tions . They will even use the threat, "You are driving
me crazy," as a defen se of their position.
What I am here calling covert schizophrenia is characteristic of
the parents of schizophrenics in the families which we have
studied. This behavior, when it occurs in the mother, has been
extensively caricatured; so I shall use here an example of which
the central figure is the father. Mr. and Mrs. P. had been married
some eighteen years and have a near-hebephrenic son of sixteen.
Their marriage is difficult and is characterized by almost continual
hostility. However, she is a keen gardener, and on a certain Sunday
afternoon they worked together planting roses in what was to be
her rose garden. She recalls that this was an unusually pleasant
occasion. On Monday morning, the husband went to work as usual,
and while he was gone Mrs. P. received a phone call from a
complete stranger inquiring, rather apologetically, when Mrs. P.
was going to leave the house. This came as somewhat of a
surprise. She did not know that from her husband's point of view
the messages of shared work on the rose garden were framed
within the larger context of his having agreed during the previous
week to sell the house.
In some cases, it almost looks as though the overt schizophrenic
were a caricature of the covert.
If we assume that both the gross ly schizophreni c symptoms of
the identified patient and the "cove rt schizoph renia" of the parents
are in part determined by genet ic factors, i.e., that, given the
appropri ate exper iential setting, genetics in some degre e renders the
patient more liable to deve lop these particular patterns of behav ior,
then we have to ask how these two degrees of patholog y might be
related in a gene tic theory.
Certainly, no answer to this quest ion is at presen t avail-able , but
it is clearly possible we here face two quite distinct probl ems. In the
case of the overt schizophreni c, the genet icist will have to identify
267

those formal characteristics of the patient which will render him
more likely to be driven to a psychotic break by the cover tly
incons istent behav ior of his parents (or by this in conjunct ion and
contrast with the more consis tent behav ior of peop le outside the
family). It is too early to make a specific guess at these
characteristics, but we may reason ably assume that they would
include some sort of rigidity. Perhap s the person prone to overt
schizophren ia would be characte rized by some extra streng th of
psycholog ical commitment to the status quo as he at the moment
sees it, which commitment would be hurt or frustrated by the
paren ts' rapid shifts of frame and context. Or perhaps this patient
might be characterized by the high value of some parameter
determining the relationship between problem solving and habit
formation. Perhaps it is the person who too readily hands over the
solutions to habit who is hurt by those changes in context which
invalidate his solutions just at the m oment when he has inco rpora ted
them into his habit structure.
In the case of cover t schizophr enia, the proble m for the geneticist
will be differen t. He will have to identify those formal
characteristics which we observe in the parents of the schizophr enic.
Here what is required would seem to be a flexibility rather than a
rigidity. But, having had some experience in dealing with these
people , 'I must confess to feeling that they are rigidly committed to
their patterns of incons isten cy.
Whether the two questions which the genet icist must answer can
simply be lumped togethe r by regarding the cover t patterns as
merely a milder version of the overt, or can be brought under a
single head by suggest ing that in some sense the same rigidity
opera tes at differen t levels in the two cases , I do not know .
Be that as it may, the difficulties which we here face are entirely
characteristic of any attempt to find a gene tic base for any
behavio ral charac teristic. Noto rious ly, the sign of any message or
behavio r is subjec t to reversa l, and this gene ralization is one of the
most importan t. contr ibutions of psycho-anal ysis, to our thinking. If
we find that a sexual exhibitionist is the child of a prudi sh paren t,
are we justified in going to the gene ticist to ask him to trace out the
genet ics of some basic characteristic which will find its phenotypic
expre ssion both in the prudishn ess of the parent and in the exhib i-
268

tionism of the offspring? The pheno mena of suppress ion and
overcompensat ion lead continu ally to the difficulty that an excess of
something at one level (e.g., in the geno type) may lead to a
deficiency of the direct expression of that something at , some more
superficial level (e.g,. in the ph enotype). And c onverse ly.
We are very far, then, from being able to pose specific questions
for the geneticist; but I believe that the wider implica tions of what I
have been saying modify somewhat the philosoph y of gene tics. Our
approach to the proble ms of schizoph renia by way of a theory of
levels or logical types has disclosed first that the proble ms of
adaptation and learning and their patholog ies must be considered in
terms of a hierarch ic system in which stoch astic change occurs at
the boundar y points between the segments of the hierarch y. We have
consider ed three such regions of stochas tic change —the level of
genetic mutation , the level of learning , and the level of chang e in
family organiz ation . We have disclosed the possibility of a
relationsh ip of these levels which orthodox gene tics would deny,
and we have disclosed that at least in human societies the
evolutionar y system consists not merely in the selective survival of
those persons who happen to select appropr iate environments but
also in the modification of family environ ment in a direction which
might enhance the phenot ypic and geno typic characteristics of the
individu al members.
What Is Man?
If I had been asked fifteen years ago what I unde rstood by the
word materia lism, I think I should have said that materialism is a
theory abou t the nature of the universe, and I would have accep ted
as a matter of course the notion that this theory is in some sense
nonmoral. I would have agreed that the scientist is an expert who
can prov ide himself and other s with insight s and te chniques , but that
science could have nothing to say about whethe r these techniques
should be used. In this, I would have been following the general
trend of scientific philosoph y assoc iated with such names at
269

Democritus , Gal ileo, Newton ,50 Lavois ier, and Darwin . I would hav e
been discard ing the less respec table views of such men as
Herac litus, the alchemists, William Blake , Lamarck, and Samuel
Butle r. For these, the motive for scien tific inquir y was the desire to
build a comprehens ive view of the unive rse which shou ld show
what Man is and how he is related to the rest of the univers e. The
picture which these men were trying to build was ethical and
aesthetic.
There is this much connect ion certainly between scientific truth,
on the one hand, and beauty and moralit y, on the other: that if a man
entertain false opinions regard ing his own natur e, he will be led
thereby to courses of action which will be in some profound sense
immoral or ug ly.
Today, if asked the same question regarding the meaning of
materialism, I would say that this word stands in my thinking for a
collection of rules about what questions shou ld be asked regarding
the nature of the universe . But I would not suppo se that this set of
rules ha s any claim to be un iquely right.
The mystic "sees the world in a grain of sand, " and the world
which he sees is either moral or aesthetic, or both. The Newtonian
scientist sees a regularity in the behav ior of falling bodies and
claims to draw from this regularity no normative conclusions
whatsoeve r. But his claim cease s to be consistent at the moment
when he preach es that this is the right way to view the unive rse. To
preach is poss ible only in terms of no rmative conc lusions .
I have touched upon several matters in the cours e of this lecture
which have been foci of controve rsy in the long battle between a
nonmoral materia lism and a more romantic view of the universe .
The battle between Darwin and Samuel Butler may have owed some
of its bitterness to what looked like personal affronts , but behind all
this the argument conce rned a question which had religious status.
50 The name of Newton certainly belongs in this list. But the man himself was of
a different kidney. His mystical preoccupation with alchemy and apocalyptic writings,
and his secret theological monis m indicate that he was not the first objecti ve
scientist but, rather, the "last of the magicians " (see J. M. Keynes, "Newton, the
Man," Tercentenar y Celebrations , London, Cambrid ge University Press, 1947, pp.
27-34). Newton and Blake were alike in devoting much time and thought to the
mystical works of Jacob Boeh me.
270

The battle was really abou t "vitalism." It was a question of how
much life and wha t order of life could be ass igned to organisms; and
Darwin' s victory amounted to this, that while he had not succeeded
in detracting from the mysterious liveliness of the indiv idual
organis m, he had at least demonstrated that the evolution ary picture
could be reduced t o natural "law."
It was, therefore, very important to demonstra te that the as yet
unconqu ered territory—the life of the individua l organis m—could
not contain anything which would recapture this evolution ary
territory. It was still mysteriou s that living organis ms could achieve
adaptiv e change during their individua l lives, and at all costs these
adaptiv e chang es, the famous acquir ed cha racteristics, must no t hav e
influen ce up-on the evolutionar y tree. The "inheri tance of acqui red
characteristics" threaten ed alwa ys to recapture the field of evolution
for the vitalist side. One part of biology must be separate from the
other. The objec tive scientists claimed, of course , to believe in a
unity in nature— that ultimately the whole of natural pheno mena
would prove suscep tible to their analysis, but for about a hundr ed
years it was convenien t to s et up an i mper meable sc reen be tween the
biology of the individua l and the theory of evolut ion. Samuel
Butler's "inheri ted memory" was an attack upon this screen .
The quest ion with which I am conce rned in this conc luding
section of the lecture could be put in various ways. Is the battle
between nonmoral materia lism and the more mystical view of the
universe affected by a chang e in the funct ion assigned to the
"acquir ed characteristics?" Does the older materialist thesis really
depend upon the premise that contex ts are isolable? Or is our view
of the world changed when we admit an infinite regress of contexts,
linked to each other in a complex network of metarelations? Does
the possibility that the separa te levels of stochas tic change (in
pheno type and geno type) may be conne cted in the larger contex t of
the ecologi cal system alter ou r allegiance in the battle?
In break ing away from the premise that contexts are al-ways
conceptua lly isolable, I have let in the notion of a universe much
more unified—and in that sense much more mystical—th an the
conv ention al universe of nonmoral materialism. Does the new
position so achieved give us new grounds for hope that science
might answer moral or aesthetic que stion s?
271

I believe that the position is significantly changed, and perhaps
I can best make this clear by considering a matter which you as
psychiatrists have thought about many times. I mean the matter of
"control" and the whole related complex suggested by such words
as manipulation, spontaneity, free will, and technique. I think you
will agree with me that there is no area in which false premises
regarding the nature of the self and its relation to others can be so
surely productive of destruction and ugliness as this area of ideas
about control. A human being in relation with another has very
limited control over what happens in that relationship. He is a
part of a two-person unit, and the. control which any part can
have over any whole is strictly limited.
The infinite regress of contex ts which I have talked abou t is only
anothe r example of the same pheno menon. What I have contributed
to this discuss ion is the notion. that the contrast between part and
whole, whenever this contrast appears in the realm of
communication , is simply a contrast in logica l typing. The whol e is
alwa ys in a metarelationship with its parts. As in logic the
propos ition can never determine the meta proposi tion, so also in
matters of con trol the smaller con text can never determine the larger.
I have remarked (e.g., when discussing the pheno mena of
phenot ypic compensa tion) that in hierarchie s of logical typing there
is often some sort of change of sign at each level, when the levels
are related to each other in such a way as to create a self-corrective
system. This appea rs in a simple diagra mmatic form in the initiatory
hierarch y which I studied in a New Guinea tribe. The initiators are
the natural enemies of the novices, because it is their task to bully
the novices into shape. The men who initiated the presen t initiators
now have a role of criticizing what is now being done in the initia-
tion ceremonies , and this makes them the natural allies of the
present novices. And so on. Something of the same sort also occurs
in American college fraternities, where juniors tend to be allied with
freshmen and s eniors wi th sophomores.
This gives us a view of the world which is still almost
unexplored. But some of its complexities may be suggested by a
very crude and imperfect analogy. I think that the functioning of
such hierarchies may be compared with the business of trying to
back a truck to which one or more trailers are attached. Each
272

segmentation of such a system denotes a reversal of sign, and each
added segment denotes a drastic decrease in the amount of control
that can be exerted by the driver of the truck. If the system is
parallel to the right-hand side of the road, and he wants the trailer
immediately behind him to approach the right-hand side, he must
turn his front wheels to the left. This will guide the rear of the
truck away from the right-hand side of the road so that the front of
the trailer is pulled over to its left. This will now cause the rear of
the trailer to point toward the right. And so on.
As anybody who has attempted this will know, the amount of
available control falls off rapidly. To back a truck with one trailer
is already difficult because there is only a limited range of angles
within which the control can be exerted. If the trailer is in line, or
almost in line, with the truck, the control is easy, but as the angle
between trailer and truck diminishes, a point is reached at which
control is lost and the attempt to exert it only results in jackknifing
of the system. When we consider the problem of controlling a
second trailer, the threshold for jackknifing is drastically reduced,
and control becomes, therefore, almost negligible.
As I see it, the world is made up of a very complex net-work
(rather than a chain) of entities which have this sort of relationship
to each other, but with this difference, that many of the entities
have their own supplies of energy and perhaps even their own
ideas of where they would like to go.
In such a world the problems of control become more akin to
art than to science, not merely because we tend to think of the
difficult and the unpredictable as contexts for art but also because
the results of error are likely to be ugliness.
Let me then conclude with a warn ing that we social scientists
would do well to hold back our eagernes s to control that world
which we so imperfectly under stand . The fact of our imperfect
unde rstand ing should not be allowed to feed our anxiety and so
increa se the need to control. Rathe r, our studies could be inspired by
a more ancien t, but toda y less honored , motive: a curio sity about the
world of which we are p art. The reward s of such work a re not power
but beau ty.
It is a strange fact that every great scien tific advan ce—not least
the adv ances which Newton achiev ed—has b een e legan t.
273

Addit ional References
W. R. Ashby, Design for a Brain, New York, John Wiley & Sons,
Inc., 1952.
, Introduct ion to Cyberne tics, New York and
London, John Wiley & Sons, Inc., 1956.
G. Bateson, D. D. Jackson, J. Haley, and J. H. Weakland, "Toward
a Theory of Schizophrenia," Behaviora l Science, 1956, 1:
251-64.
G. Bateson, "Cultural Problems Posed by a Study of Schizophrenic
Process," Sympos ium on Schizophr enia, an Integrated
Appr oach, by Alfred Auerback, M. D., ed., American
Psychiatric Association, Symposium of the Hawaiian Divi-
sional Meeting, 1958, New York, Ronald Press, 1959.
, "The New Conceptual Frames for Behavioral Re-search,"
Proceed ings of the Sixth Annual Psychiat ric Conference at
the New Jersey Neuro-Ps ychiat ric Institute, Princeton,
1958, pp. 54-71.
, "The Group Dynamics of Schizophrenia," Chronic
Schizoph renia , L. Appleby, J. M. Scher, and J. H.
Cummings, eds., Glencoe, Ill., The Free Press, 1960.
, "Social Planning and the Concept of Deutero-
Learning," Relation to the Democratic Way of Life, Con-
ference on Science, Philosophy and Religion, Second
Symposium, led by L. Bryson and L. Finkelstein, New
York, Harper & Bros., 1942.
, Naven, a Surv ey of Proble ms Suggested by a Composite
Picture of Culture of a New Guinea Tribe Drawn from
Three Points of View, Ed. 2, Stanford, Calif., Stanford
University Press, 1958.
S. Butler, Thought and Language, 1890, published in the
Shrewsbury Edition of the works of Samuel Butler, 1925,
vol. xix.
, Luck or Cunning as the Main Means of Organic
Modification, London, Trubner, 1887.
C. D. Darlington, "The Origins of Darwinism," Scientific
American, 1959, 200: 60-65.
274

C. Darwin, On the Origin of Species, by Means of Natural
Selection, London, Murray, 1859.
C. C. GiIlispie, "Lamarck and Darwin in the History of Science,"
American Scientist, 1958, 46: 388-409.
J. Stroud, "Psychological Moment in Perception-Discussion,"
Cybernetics: Circular Causal and Feedback Mechanisms in
Biological and Social Systems, Transactions of the Sixth
Conference, H. Von Foerster, et al:, eds., New York, Josiah
Macy; Jr. Foundation, 1949, pp. 27-63.
C. H. Waddington, The Strategy of the Genes, London, George
Allen & Unwin, Ltd., 1957.
, "The Integration of Gene-Controlled Processes and Its
Bearing on Evolution," Caryologia, Supplement, 1954, pp.
232-45.
, "Genetic Assimilation of an Acquired Character,"
Evolution, 1953, 7: 118-26.
A. Weismann, Essays upon Heredity, authorized translation, E. B.
Poulton, et al., eds., Oxford, Clarendon Press, 1889.
275

Doub le Bind, 1969*
Double bind theory was, for me, an exemplification of how to
think about such matters and, in this aspec t at least, the whole
busine ss is wor th so me re-examination.
Sometimes—of ten in science and alwa ys in art—one does not
know what the proble ms were till after they have been solved. So
perhap s it will be useful to state retrospec tively what problems were
solved for me by double bind theor y.
First there was t he proble m of reification.
Clear ly there are in the mind no objects or even ts—no pigs, no
coconut palms, and no mothers. The mind contains only transfo rms,
percepts, images, etc., and rules for making these transfo rms,
percepts, etc. In what form these rules exist we do not know , but
presu mabl y they are embodied in the very machine ry which creates
the transfor ms. The rules are certainly not commonly explicit as
consc ious " thoughts ."
In any case, it is nonsense to say that a man was frightened by a
lion, because a lion is not an idea. The man makes an idea of the
lion.
The explanatory world of substance can invoke no differences
and no ideas but only forces and impacts . And, per cont ra, the world
of form and communicat ion invokes no things , forces, or impacts
but only differences and ideas. (A difference which makes a
difference is an idea. It is a "bit," a unit of information.)
But these things I learned only later—was enabled to learn them
by doubl e bind theory. And yet, of course, they are implic it in the
theory which could hardly have been created without them.
Our original paper on the double bind conta ins numerous errors
due simply to our having not yet articulately examined the
* This paper was given in August, 1969, at a Symp osium on the Double Bind;
Chair man, Dr. Robert Ryder; sponsored by the American Psychological Association.
It was prepar ed under Career Develop ment Award (MH -21,931) of the National
Institute of Mental He alth.
276

reification problem. We talk in that paper as though a doubl e bind
were a something and a s though su ch so me-things could be counted .
Of course that's all nons ense. You canno t count the bats in an
inkblot because there are none . And yet a man—if he be "bat-
minded"— may "see" s evera l.
But are there double binds in the mind? The question is not
trivial. As there are in the mind no coconuts but only percep ts and
transforms of coconu ts, so also, when I perceive (consciou sly or
uncon scious ly) a doubl e bind in my boss' behavior , I acqui re in my
mind no doub le bind but only a percept or transform of a doub le
bind. And that is not what the theory is abou t.
We are talking then abou t some sort of tangle in the rules for
making the transfo rms and about the acquis ition or cultivation of
such tangles . Doubl e bind theory asserts that there is an expe riential
component in the determination or etiology of schizophren ic
symptoms and related behav ioral patterns, such as humor, art,
poetry, etc. Notabl y the theory does not distinguish between these
subspecie s. Within its terms there is nothing to determine whethe r a
given individua l shall beco me a clown, a poet, a schizophren ic, or
some combination of these. We deal not with a single syndrome but
with a genus of syndromes, most of which are not convent ional ly
regarded as patho logica l.
Let me coin the word "transcontextual" as a general term for
this genus of syndromes.
It seems that both those whose life is enriched by trans-
contextu al gifts and those who are impoverish ed by transcontextual
confusions are alike in one respec t: for them there is alwa ys or often
a "doubl e take." A falling leaf, the greeting of a friend , or a
"primrose by the river's brim" is not "just that and nothing more."
Exogenous experi ence may be framed in the contexts of dream, and
internal thought may be proje cted into the contexts of the external
world. And so on. For all this, we seek a partial explana tion in
learning and expe rience .
There must, o f cou rse, also be gene tic components in the etiolog y
of the transcon textua l syndromes. These would expectably operate
at levels more abstract than the expe riential. For example, genetic
component s might determine skill in learning to be transcon textua l
or (more abstractly) the poten tialities for acqu iring this skill. Or,
277

conver sely, the geno me might determine skills in resisting
transcont extua l pathwa ys, or the potentiality for acqui ring this latter
skill. (Gene ticists have paid very little attention to the necess ity of
defining t he logical typing of messages carried by DNA.)
In any case, the meeting point wher e the gene tic determination
meets the exper iential is surely quite abstract, and this must be true
even though the embodiment of the gene tic message be a single
gene . (A single bit of information—a single difference— may be the
yes-or-no answer to a question of any degree of complexit y, at any
level of abstraction. )
Curren t theori es which propos e (for "schi zophreni a") a single
dominant gene of "low pene trance " seem to leave the field open for
any experien tial theory which would indicate what class of
exper iences might cause the latent poten tiality to appea r in the
phenot ype.
I must confess however that these theories seem to me of little
interes t until the propon ents try to specify what components of the
complex process of dete rmining "schi zophrenia" are provided b y the
hypothetical gene . To identify these components must be a
subtractive proce ss. Where the contribution of environment is
large, the gene tics cannot be investigated until the environ mental
effect has b een i dentified and can be con trolled.
But sauce for the goose is also sauce for the gande r, and what is
said above about geneticists places an obligation upon me to make
clear what component s of transcont extua l proces s could be provided
by doubl e bind expe rienc e. It is approp riate therefore to re-examine
the theory of deuterolearning upon which double bind theory is
based .
All biolog ical systems (organis ms and social or ecologica l
organizations of organis ms) are capab le of adaptiv e chang e. But
adapt ive change takes many forms, such as response, learning,
ecological succession, biological evolution, cultural evolution, etc.,
according to the size and complexity of the system which we
choose to consider.
Whatever the system, adaptive change depends upon feedback
loops, be it those provided by natural selection or those of
individual reinforcement. In all cases, then, there must be a process
of trial and error and a mechanism of comparison.
278

But trial and error must always involve error, and error is
always biologically and/or psychically expensive. It follows
therefore that adaptive change must always be hierarchic.
There is needed not only that first-order change which suits the
immediate environmental (or physiological) demand but also
second-order changes which will reduce the amount of trial and
error needed to achieve the first-order change. And so on. By
superposing and interconnecting many feedback loops, we (and all
other biological systems) not only solve particular problems but
also form habits which we apply to the solution of classes of
problems.
We act as though a whole class of problems could be solved in
terms of assumptions or premises, fewer in number than the
members of the class of problems. In other words, we (organisms)
learn to learn, or in the more technical phrase, we deutero-learn.
But habits are notoriously rigid and their rigidity follows as a
necessary corollary of their status in the hierarchy of adaptation.
The very economy of trial and error which is achieved by habit
formation is only possible because habits are comparatively "hard
programmed," in the engineers' phrase. The economy consists
precisely in not re-examining or rediscovering the premises of
habit every time the habit is used. We may say that these premises
are partly "unconscious", or—if you please—that a habit of not
examining them is developed.
Moreove r, it is important to note that the premises of habit are
almost necessa rily abstract. Ever y proble m is in some degree
different fro m every other and its desc ription or repres entation in the
mind will therefo re contain uniqu e proposi tions . Clearly to sink
these unique propos itions to the level of premises of habit would be
an error. Habi t can deal succe ssful ly only with proposi tions which
have gene ral or repetitive truth, and these are commonly of a
relatively high order of abs traction.51
51 What is important, however, is that the proposition be constantl y true, rather than
that it be abstract . It just so happens—coincidentally—that abstractions, if well cho-
sen, have a constanc y of truth. For human beings it is rather constantly true that air is
present around the nose; the refle xes which control respiration can th erefore be hard-
programm ed in the medulla. For the porpoise, the proposition "air around the
blowhole " is only intermittentl y true, and therefore respiration must be con-trolled
279

Now the particular propos itions which I believe to be important
in the determination of the transcon textua l syndromes are those
formal abstractions which describe and determine interpe rsonal
relationship .
I say "describe and determine," but even this is inadequate .
Better would be to say that the relationship is the exchang e of these
messages ; or that the relation ship is immanent i n these messages .
Psycholog ists commonly speak as if the abstractions of re-
lationsh ip ("depend ency," "hostility," "love, " etc.) were real things
which are to be described or "expr essed" by messages. This is
epistemology backwards : in truth, the messages constitute the
relationship , and words like `.`dependenc y" are verbally coded
descriptions of patterns immanent in the combination of exchanged
messages .
As has already been mentioned, there are no "things" in the mind
—not even " dependenc y."
We are so befuddled by language that we cannot think straigh t,
and it is conv enien t, sometimes, to remember that we are really
mammals. The epistemology of the "heart" is that of any nonhu man
mammal. The cat does not say "milk"; she simply acts out (or is ) her
end of an interchange , the pattern of which we in language would
call "dependenc y."
But to act or be one end of a pattern of interaction is to propose
the other end . A context is set for a certain c lass of respon se.
This weaving of contexts and of messages which propo se context
—but which , like all messages what soever , have "meaning" only by
virtue of context— is the subjec t matter of the so-called double bind
theory.
The matter may be illustrated by a famous and formally correct52
botan ical analog y. Goethe pointed out 150 years ' ago that there is a
sort of syntax or grammar in the anatomy of flowering plants . A
"stem" is that which bears "leaves"; a "leaf" is that which has a bud
in its axil; a bud is a stem which originates in the axil of a leaf; etc.
The formal (i.e., the communicationa l) natur e of each organ is
in a more fle xible manner fro m some higher center .52 Formally correct because morpho genesis, like behavior, is surel y a matter of
messag es in contexts. (See G. Bateson, "A Re-examination of 'Bateson's Rule,'"
Journal of Genetics, in press.)
280

determined by its contextu al status— the context in which it occurs
and the context which i t sets for other parts.
I said abov e that double bind theory is conce rned with the
experien tial component in the gene sis of tangles in the rules or
premises of habit. I now go on to assert that exper ienced brea ches in
the weave of contextua l structu re are in fact "doubl e binds" and
must neces sarily (if they contribut e at all to the hierarch ic process es
of learning and adapt ation ) promote what I am calling
transcon textu al syndromes.
Consid er a very simple paradig m: a female porpois e (Steno
bredanensis) is trained to accep t the sound of the trainer's whistle
as a "seconda ry reinfo rcement." The whist le is expectab ly followed
by food, and if she later repeats what she was doing when the
whis tle blew, she will expectab ly again hear the whist le and receive
food.
This porpo ise is now used by the trainers to demonstra te "op erant
cond itioning" to the publi c. When she enters the exhibition tank, she
raises her head above surface, hears the whis tle and is fed. She then
raises her head again and is again reinfo rced. Three repetitions of
this sequenc e is enough for the demonstration and the porpo ise is
then sent off-stage to wait for the next performance two hours later.
She has learned some simple rules which relate her actions, the
whis tle, the exhibi tion tank, and the trainer into a pattern—a
contextu al structu re, a set of rules for how to put the in-formation
togethe r.
But this pattern is fitted only for a single episode in the
exhibition tank. She must break that pattern to deal with the class of
such episodes . There is a larger context of contexts which will put
her in the wrong .
At the next perfor mance, the trainer again wants to demonstrate
"operant condi tioning," but to do this she must pick on a different
piece of conspicuous behav ior.
When the porpo ise comes on stage, she again raises her head.
But she gets no whistle. The trainer waits for the next piece of
conspicuous behavior— likely a tail flap, which is a common
expression of anno yance. This behavior is then rein-forced and
repeat ed.
281

But the tail flap was, of course, not rewarded in the third
perfo rmance.
Finally the porpoise learned to deal with the context of contexts
—by offering a different or new piece of consp icuous behavior
whenever she c ame on stage.
All this had happ ened in the free natural history of the
relationship between porpoi se and trainer and audience . The
sequenc e was then repeated experimental ly with a new porpo ise and
carefully recorded .53
Two points from this expe rimental repea t of the sequenc e must
be added:
First, tha t it was necessar y (in the trainer's judg ment) to bre ak the
rules of the exper iment many times. The experience of being in the
wrong was so disturbing to the porpois e that in order to preserv e the
relationship between porpoi se and trainer (i.e., the con text of contex t
of contex t) it was necessa ry to give many reinfor cements to which
the porpo ise was not entitled.
Second, that each of the first fourte en sessions was characterized
by many futile repetitions of whatever behavior had been reinforced
in the immediate ly previous session. Seemingly only by "acciden t"
did the animal provide a piece of differen t behavio r. In the time-out
between the fourteenth and fifteen th sessions, the porpo ise appeared
to be much excited, and when she came on stage for the fifteenth
session she put on an elabo rate perfo rmance including eight
conspi cuous pieces of behavio r of which four were entirely new—
never before obs erved in this spe cies of animal.
The story illustrates, I believe, two aspects of the genesis of a
transcont extua l syndrome:
First, that severe pain and maladjust ment can be induc ed by
putting a mammal in the wrong regard ing its rules for making sense
of an important relationship with anoth er mammal.
And second, that if this patholog y can be ward ed off or resisted,
the total exper ience may promote creativity.
53 K. Pryor, R. Haag , and J. O'Riel ly, "Deute ro-Learning in a
Roughtooth Porpoise (Steno bredanensis)," U. S. Naval Ordinance
Test Station, China Lake, NOTS TP 4270
282

Bibliography
G. Bateson, "Social Planning and the Concept of Deutero-
Learning," Science, Philosophy and Religion; Second
Symposium, L. Bryson and L. Finkelstein, eds., New York,
Conference on Science, Philosophy and Religion in their
Relation to the Democratic Way of Life, Inc., 1942.
"Minimal Requirements for a Theory of Schizophrenia,"
A.M.A. Archives of General Psychiatry, 1960, 2: 477-91.
, Perceval's Narrative, A Patient's Account of his Psychos is,
1830-1832, edited and with an introduction by Gregory
Bateson, Stanford, Calif., Stanford University Press, 1961.
, "Exchange of Information about Patterns of Hu-man
Behavior," Information Storage and Neural Control; Tenth
Annual Scientific Meeting of the Houston Neurological
Society, W. S. Fields and W. Abbott, eds., Springfield, Ill.,
Charles C. Thomas, 1963.
, "The Role of Somatic Change in Evolution," Evolution,
1963, 17: 529-39.
283

The L ogical Categories o f Learning and
Communi cation*
All species of behavio ral scientists are concerned with "learning"
in one sense or another of that word. Moreover , since "learn ing" is a
communication al pheno menon, all are affected by that cyberne tic
revolution in though t which has occurred in the last twent y-five
years. This revolu tion was triggered by the engine ers and
communication theoris ts but has older roots in the physiologi cal
work of Cannon and Claude Bernard , in the physics of Clarke
Maxwel l, and in the mathematica l philosoph y of Russel l and
Whitehead . Insof ar as behavior al scientists still ignore the problems
of F incipia Mathematica,54 they can claim approximatel y sixty
years of obsole scence .
It appears , however , that the barriers of misunder stand ing which
divide the various specie s of behav ioral scien tists can be illuminated
(but not el iminated) by an application of Russel l's Theory of Logic al
Types to the concept of "learning" with which all are concerned. To
attempt this illumination will be a purpose of the present essay.
The Theory of Logica l Types
First, it is approp riate to indica te the subjec t matter of the Theory
of Logical Types: the theory asserts that no class can, in formal
logical or mathematical discours e, be a member of itself; that a class
of classes cannot be one of the classes which are its members; that a
name is not the thing named; that "John Bateson" is the class of
* This essay was written in 1964 while the author was emplo yed by the
Communications Research Institute, under a Career Develop ment Award (K3-NH-
21, 931) from the National Institute of Mental Health. It was submitted as a position
paper to the "Conference on World Views" sponsored by the Wenner-Gren Foundation,
August 2- 11, 1968. The section on "Learnin g III" was added in 1971 .54 A. N. Whitehead and B. Russell , Principia Mathematica, 3 vols., 2nd ed.,
Cambridg e, Cambridg e Universit y Press, 1910 -13.
284

which that boy is the unique member; and so forth. These assertions
may seem trivial and even obvious, but we shall see later that it is
not at all unusual for the theori sts of behavio ral science to commit
errors which are precisely analogous to the error of classifying the
name with the thing named—or eating the menu card instead of the
dinner—an error o f logica l typing .
Somewhat less obvious is the further assertion of the theory: that
a class cannot be one of thos e items which are correctly classified as
its nonmembers. If we classify chairs together to const itute the class
of chairs, we can go on to note that tables and lamp shade s are
members of a large class of "noncha irs," but we shall commit an
error in formal discou rse if we coun t the class of chair s among the
items within the class o f noncha irs.
Inasmuch as no class can be a member of itself, the class of
nonch airs clearly canno t be a nonchai r. Simple conside rations of
symmetry may suffice to conv ince the nonmathematical reader: (a)
that the class of chairs is of the same order of abstraction (i.e., the
same logica l type) as the class of noncha irs; and further, (b) that if
the class of chairs is not a chair, then, corresponding ly, the class of
nonch airs is not a nonchai r.
Lastly, the theory asserts that if these simple rules of formal
discours e are contravened, paradox will be gene rated and the
discours e vitiated.
The theory, then, deals with highl y abstract matters and was first
derived within the abstract world of logic. In that world , when a
train of propos itions can be shown to genera te a paradox, the entire
structure of axioms, theore ms, etc., involved in generating that
paradox is thereby negated and reduced to nothing. It is as if it had
never been. But in the real world (or at least in our descriptions of
it), there is always time, and nothing which has been can ever be
totally negated in this way. The computer which encounters a para-
dox (due to faulty programming) does not vanish away.
The "if … then …" of logic cont ains no time. But in the computer,
cause and effect are used to simulate the "if .. . then . . ." of logic;
and all sequences of cause and effect necessarily involve time.
(Conve rsely, we may say that in scientific explana tions the "if …
then …" of logic is used to simulate the "if … then …" of cause and
effect.)
285

The computer never truly encoun ters logical paradox, but only
the simulation of paradox in trains of cause and effect. The
computer therefor e does not fade away. It merely oscillates.
In fact, there are importan t difference s between the world of
logic and the world of pheno mena, and these differenc es must be
allowed for whenever we base our arguments upon the partial but
important a nalog y which ex ists between th em.
It is the thesis of the present essay that this partial analogy can
provid e an important guide for behavio ral scientists in their
classification of pheno mena related to l earning. Prec isely in the f ield
of animal and mechanic al communication something like the theory
of types must app ly.
Question s of this sort, howeve r, are not often discuss ed in
zoolog ical laboratories, anthropo logic al field camps, or psychiat ric
convent ions, and it is necessar y therefor e to demonstrate that these
abstract cons idera tions are important to behav ioral scientists.
Consider the following syllogism:
(a)Changes in frequency of items of mammalian behavior
can be described and predicted in terms of various
"laws" of reinforcement.
(b)"Exploration" as observed in rats is a category, or
class, of mammalian behavior.
(c)Therefore, changes in frequency of "exploration"
should be describable in terms of the same "laws" of
reinforcement.
Be it said at once: first, that empirica l data show that the
conclu sion (c) is untrue; and second , that if the conclusion (c) were
demonstrab ly true, then e ither (a) or (b) would be untrue.55
55 is conceivable that the same words might be used in describ ing
both a class and its members and be true in both cases. The word
"wave" is the name of a class of movements of particles. We can
also say that the wave itself "moves," but we shall be referring to a
movement of a class of movements. Under friction, this metamove-
ment will not lose velocity as would the moveme nt of a particle.
286

Logi c and natural history would be better served by an expanded
and corrected ve rsion of the conclus ion (c) some-what as follows :
(c) If, as asserted in (b), "exploration" is not an item of
mammalian behavio r but is a category of such items,
then no descriptive statement which is true of items of
behavior can be true of "explor ation ." If, howeve r,
descrip tive statements which are true of items of
behavior are also true of "explor ation ," then
"exploration" is an item and no t a categor y of items.
The whole matter turns on wheth er the distinction between a
class and its members is an ordering principle in the behavio ral
pheno mena which we study.
In less formal languag e: you can reinforce a rat (positively
or nega tively) when he investigates a particular strange object,
and he will approp riately learn to approach or avoid it. But the very
purpose of exploration is to get information abou t which objects
shou ld be appro ached and which avoided. The discove ry that a
given objec t is dangerous is therefore a success in the busine ss of
getting information. The success will not discourage the rat from
future explo ration of other strange objects.
A priori it can be argued that all percep tion and all respon se, all
behavior and all classes of behavio r, all learning and all gene tics, all
neurophysiolog y and endocr inolog y, all organization and all
evolution—one entire subject matter must be regarded as
communica tiona l in nature, and there-fo re subject to the great
generalizations or "laws" which apply to communicat ive
pheno mena. We therefore are warn ed to expe ct to find in our data
those principles of order which funda mental communication theory
would pro-po se. The Theory of Logica l Types, Information Theory,
and so forth, are expec tably to be our guides.
The "Learning" of Computer s, Rats, and M en
The word "learn ing" undoubted ly denote s change of some kind.
To say what kind of change is a delicate matter.
287

Howeve r, from the gross co mmon denominato r, "change," we can
deduce that our descriptions of "learning" will have to make the
same sort of allowance for the varieties of logical type which has
been routine in physical science since the days of Newton . The
simplest and most familiar form of change is motion, and even if
we work at that very simple physical level we must structure our
descriptions in terms of "position or zero motion," "const ant
veloc ity," "acceleration," "rate of chang e of acceleration ," and so on
.56
Change deno tes proces s. But processe s are themselves subject to "change ." The proce ss may accelerate, it may slow down, or it may
under go other types of change such that we shall say that it is now a "differen t" process .
These conside rations suggest that we should begin the ordering
of ou r ideas about "learning" at the very simplest level.
Let us consider the case of specificity of respon se, or zero
learning. This is the case in which an entity shows minimal change
in its response to a repeated item of senso ry input . Pheno mena
which a pproach this deg ree of simplicity occur in various con texts:
(a)In experimental settings, when "learning" is complete
and the animal gives approximately 100 per cent correct
responses to the repeated stimulus.
(b)In cases of habituation, where the animal has ceased to
give overt response to what was formerly a disturbing
stimulus.
(c)In cases where the pattern of the response is minimally
determined by experience and maximally determined by
genetic factors.
(d)In cases where the response is now highly stereo-typed.
(e)In simple electronic circuits, where the circuit
structure is not itself subject to change resulting
from the passage of impulses within the circuit—i.e.,
56 The Newtonian equations which describe the motions of a "particle " stop at the
level of "acceleration. " Change of acceleration can only happen with deforma-
tion of the moving body, but the Newtonian "particle " was not made up of "parts"
and was therefore (logically) incapable of deformation or any other internal change. It
was therefore not subje ct to rate of chan ge of acceleration.
288

where the causal links between "stimulus" and
"response" are as the engineers say "soldered in."
In ordina ry, nontechni cal parlanc e, the word "learn" is often
applied to what is here called "zero learning," i.e., to the simple
receip t of information from an extern al event , in such a way that a
similar event at a later (and appropri ate) time will conve y the same
information: I "learn" from the factory whist le that it is twelve
o'clock .
It is also interes ting to note that within the frame of our
definition many very simple mechanic al device s show at least the
pheno menon of zero learning. The question is not, "Can machines
learn?" but what level or order of learning does a given machine
achieve ? It is wor th looking at an ex treme, i f hypotheti cal, case:
The "player" of a Von Neumannian game is a mathematical
fiction, comparable to the Euclidean straight line in geometry or the
Newtoni an particle in physics. By definition, the "player" is capab le
of all computat ions neces sary to solve whatev er prob lems the even ts
of the game may presen t; he is incapab le of not performing these
computations whenever they are approp riate; he alwa ys obeys the
findings of his computat ions. Such a "player" receives information
from the even ts of the game and acts appropr iately upon that infor-
mation. But his learning is limited to what is here called zero
learning.
An examination of this formal fiction will contribute to our
definition of zero learning .
The "player" may receive, from the events of the game,
information of higher or lower logica l type, and he may use this
information to make decisions of highe r or lower type. That is, his
decisions may be either strategic or tactical, and he can identify and
respond to indica tions of both the tactics and the strategy of his
opponen t. It is, how-ever , true that in Von Neumann's formal
definition of a "ga me," a ll probl ems which t he game may present are
conceived a s computable , i.e., whil e the ga me may conta in probl ems
and information of many differen t logica l types, the hierarch y of
these types is strictly finite.
It appea rs then that a definition of zero learning will not depend
upon the logical typing of the information received by the organis m
289

nor upon the logica l typing of the adaptiv e decision s which the
organism may make. A very high (but finite) order of complexi ty
may characterize adap tive behavior based on nothing higher than
zero learning .
(1) The "player" may compute the value of information which
would benefit him and may compute that it will pay him to acqu ire
this information by engaging in "explo ratory" moves. Alternat ively,
he may make delaying or tentative moves while he waits for needed
information .
It follows that a rat engag ing in exploratory behavior m ight do so
upon a basis of zero learning.
(2) The "player" may compute that it will pay him to make
rando m moves. In the game of matching penn ies, he will compute
that if he selects "heads" or "tails" at rando m, he will have an even
chance of winning . If he uses any plan or pattern, this will appear as
a pattern or redundanc y in the sequence of his moves and his
opponent will thereb y receive information . The "player" will
therefor e elect to play in a rando m manner.
(3) The "player" is in capabl e of "er ror." He may, for good rea son,
elect to make rando m moves or explor atory moves, but he is by
definition incapab le of "learn ing b y trial and e rror."
If we assume that, in the name of this learning process , the word
"error" means what we meant it to mean when we said that the
"player" is incapab le of error, then "trial and error" is exclud ed from
the repertoire of the Von Neumannian player. In fact, the Von
Neumannian "player" forces us to a very careful examination of
what we mean by "trial and error" learning , and indeed what is
meant by "learning" of any kind. The assumption regard ing the
meaning of the word "error" is not trivial and must now be
examined.
There is a sense in which the "player" can be wrong . For
example, he may base a decision upon probabi listic considerations
and then make that move which , in the light of the limited available
information , was most probabl y right. When more information
beco mes available, he may discover that that move was wrong. But
this disco very can contribu te nothing to his future skill. By
definition, the player used correctly all the available information.
He estimated the probabi lities correctly and made the move which
290

was most probabl y correct. The discover y that he was wrong in the
particula r instanc e can have no bearing upon future in-stances .
When the same problem returns at a later time, he will correctly go
through the same computat ions and reach the same decision.
Moreove r, the set of alternatives among which he makes his choice
will be the same set—and co rrectly so.
In contrast, an organis m is capabl e of being wrong in a number
of ways of which the "pla yer" is incapab le. These wrong choices are
appropri ately called "error" when they are of such a kind that they
would provid e information to the organism which might contribut e
to his future skill. These will all be cases in which some of the
available i nformation was ei ther igno red or i ncorrectly used. Various
species of such profitable error can be classified.
Suppose that the external event system conta ins details which
might tell the organism: (a) from what set of alternatives he should
choo se his next move; and (b) which member of that set he should
choo se. Such a situation p ermits two orders of error:
The organism may use correctly the information which tells
him from what set of alternatives he should choose, but choose the
wrong alternative within this set; or
He may choose from the wrong set of alternatives. (There is
also an interesting class of cases in which the sets of alternatives
contain common members. It is then possible for the organism to
be "right" but for the wrong reasons. This form of error is
inevitably self-reinforcing.)
If now we accept the overa ll notion that all learning (other than
zero learning ) is in some degree stochas tic (i.e., contains
component s of "trial and error") , it follows that an ordering of the
processe s of learning can be built upon an hierarchic classification
of the types of error which a re to be corrected in the va rious learning
processe s. Zero learning will then be the label for the immediate
base of all those acts (simple and complex) which are not subjec t to
correction by trial and error. Learn ing I will be an appropriate label
for the revision of choice within an unchang ed set of alterna tives ;
Learning II will be the label for the revision of the set from which
the cho ice is to be m ade; and so on .
291

Learning I
Following the formal analogy provided by the "laws" of motion
(i.e., the "rules" for descr ibing motion), we now look for the class of
pheno mena which are appropr iately described as changes in zero
learning (as "motion" describes change of position) . These are the
cases in which an entity gives at Time 2 a differen t response from
what it gave at Time 1, and again we encounte r a variety of cases
variousl y related to expe rience, physiolog y, genet ics, and
mechanica l process :
(a)There is the phenomenon of habituation—the change from
responding to each occurrence of a repeated event to not overtly
responding. There is also the extinction or loss of habituation,
which may occur as a result of a more or less long gap or other
interruption in the sequence of repetitions of the stimulus event.
(Habituation is of especial interest. Specificity of response, which
we are calling zero learning, is characteristic of all protoplasm,
but it is interesting to note that "habituation" is perhaps the only
form of Learning I which living things can achieve without a
neural circuit.)
(b)The most familiar and perhaps most studied case is that of
the classical Pavlovian conditioning. At Time 2 the dog salivates
in response to the buzzer; he did not do this at Time 1.
(c)There is the "learning" which occurs in contexts of
instrumental reward and instrumental avoidance.
(d)There is the phenomenon of rote learning, in which an
item in the behavior of the organism becomes a stimulus for
another item of behavior.
(e)There is the disruption, extinction, or inhibition of
"completed" learning which may follow change or absence of
reinforcement.
In a word , the list of Learn ing I contains those items which are
most co mmonly called "learn ing" i n the ps ycho-log ical laboratory.
292

Note that in all cases of Learn ing I, there is in our description an
assumption about the "cont ext." This assumption must be made
explicit. The definition of Learning I assumes that the buzzer (the
stimulus) is someho w the "same" at Time 1 and at Time 2. And this
assumption of "sameness" must also delimit the "cont ext," which
must (theore tically) be the same at both times. It follows that the
events which oc curred at Time 1 are not, in our des cription; included
in our definition of the context at Time 2, becau se to include them
would at once create a gross d ifference between "con-t ext at Time 1"
and "contex t at Time 2." (To paraphr ase Herac litus: "No man can go
to bed with the same girl for the first time twice.")
The conv ention al assumption that context can be repeated, at
least in some cases, is one which the writer adopts in this essay as a
corners tone of the thesis that the study of behavior must be ordered
according to the Theory of Log ical Types. Without the assumption of
repeat able contex t (and the hypothesi s that for the organisms which
we study the sequence of exper ience is really somehow punctuated
in this manner) , it would follow that all "learn ing" would be of one
type: namely, all would be zero learning . Of the Pavlov ian
experiment, we would simply say that the dog's neural circuits
contain "solde red in" from the beginn ing such characteristics that in
Contex t A at Time 1 he will not salivate, and that in the totally
different Context B at Time 2 he will salivate. What previous ly we
called "learn ing" we would now describe as "discrimination"
between the events of Time 1 and the even ts of Time 1 plus Time 2.
It would then follow logica lly that all questions of the type, "Is this
behavior `learned' or `innate'?" shou ld be answer ed in favor of
genetics.
We would argue that withou t the assumption of repea t-able
context, our thesis falls to the ground , togethe r with the whole
general concep t of "learn ing." If, on the other hand, the assumption
of repeatab le contex t is accepted as somehow true of the organisms
which we study, then the case for logical typing of the pheno mena
of learning necessarily stands, because the notion "contex t" is itself
subject to logical typing.
Either we must discard the notion of "contex t," or we retain this
notion and, with it, accept the hierarchic series—s timulus, con text of
293

stimulus, context of contex t of stimulus, etc. This series can be
spelled out in the form of a hierarchy of logica l types a s follows:
Stimulus is an elementar y signal, internal or external . Context of
stimulus i s a metamessage which classifies the elementary signal.
Context of contex t of stimulus is a meta-metamessage which
classifies the metamessage .
And so oil.
The same hierarch y could have been built up from the notion of
"respon se" o r the no tion of "reinforce ment."
Alternatively, following up the hierarchic classification of errors
to be corrected by stochast ic process or "trial and error," we may
regard "cont ext" as a collective term for all those event s which tell
the organism among what set of alternatives he must make his next
choic e.
At this point it is convenien t to introdu ce the term "con-text
marker." An organism responds to the "same" stimulus differently in
differing contexts , and we must therefore ask abou t the source of the
organisms's information . From what percep t does he know that
Context A is different from Con-t ext B?
In many instances , there may be no specific signal or label
which will classify and different iate the two contex ts, and the
organism will be forced to get his information from the actual
conger ies of events that make up the contex t in each case. But,
certainly in human life and probably in that of many other
organisms, there occu r signal s whose major function is to classify
contex ts. It is not unreasonab le to sup-pose that when the harness is
placed upon the dog, who has had prolonged training in the
psycholog ical laborato ry, he knows from this that he is now
embarking upon a series of contexts of a certain sort. Such a source
of information we shall call a "context marker," and note
immediately that, at least at the human level, there are also
"markers of contexts of contexts." For example: an audience is
watching Hamlet on the stage, and hears the hero discuss suicide
in the con-text of his relationship with his dead father, Ophelia,
and the rest. The audience members do not immediately telephone
for the police because they have received information about the
context of Hamlet's context. They know that it is a "play" and have
received this information from many "markers of context of
294

context"—the playbills, the seating arrangements, the curtain, etc.,
etc. The "King," on the other hand, when he lets his conscience be
pricked by the play within the play, is ignoring many "markers of
context of context."
At the human level, a very diverse set of events falls within the
categor y of "contex t markers. " A few examples are here listed:
(a)The Pope's throne from which he makes announcements
ex cathedra, which announcements are there.. by
endowed with a special order of validity.
(b)The placebo, by which the doctor sets the stage for a
change in the patient's subjective experience.
(c)The shining object used by some hypnotists in "inducing
trance."
(d)The air raid siren and the "all clear."
(e)The handshake of boxers before the fight.
(f) The observances of etiquette.
These, however , are examples from the social life of a highly
complex organism, and i t is more pro fitable at this stage to ask about
the an alogous pheno mena a t the pre-ve rbal level.
A dog may see the leash in his master's hand and act as if he
knows that this indicates a walk; or he may get in-formation from
the sound of the word "walk" that this type of contex t or sequence is
coming.
When a rat starts a sequen ce of exploratory activities, does he do
so in respon se to a "stimulus?" Or in response to a context? Or in
response to a con text marker?
These quest ions bring to the surface formal proble ms abou t the
Theory of Logica l Types which must be discussed. The theory in its
origin al form deals only with rigorousl y digital communication, and
it is doubtfu l how far it may be applied to analogue or iconic
systems. What we are here calling "contex t markers" may be digital
(e.g., the word "walk " mentioned abov e) ; or they may be analogu e
signal s —a briskness in the master's movements may indicate that a
walk is pend ing; or some part of the coming context may serv e as a
marker (the leash as a part of the walk) ; or in the extreme case, the
295

walk itself in all its complexit y may stand for itself, with no label or
marker between the dog and the expe rienc e. The perceiv ed event
itself may communicat e its own occu rrenc e. In this case, of course,
there can be no error of the " menu card" type. Moreov er, no paradox
can be generated because in purely analogu e or iconic communica-
tion th ere is no signal for "not."
There is, in fact, almost no formal theor y dealing with analogue
communication and, in particular, no equiva lent of Information
Theory or Logi cal Type Theory. This gap in formal knowledge is
inconven ient when we leave the rarified world of logic and
mathematics and come face to face with the pheno mena of natural
history. In the natural world , communication is rarely either purely
digital or purely analogic. Often discre te digital pips are combined
together to make analog ic pictures as in the printer's halftone block;
and sometimes, as in the matter of context markers , there is a
continuous gradat ion from the ostensiv e through the iconic to the
purely digital. At the digital end of this scale all the theore ms of
information theory have their full force, but at the osten sive and
analog ic end t hey are meaningless .
It seems also that while much of the behavio ral communication
of even higher mammals remains ostensive or analog ic, the intern al
mechanis m of these creatures has beco me digitalized at least at the
neurona l level. It would seem that analogi c communicat ion is in
some sense more primitive than digital and that there is a broad
evolu tiona ry trend toward the substitution of digital for analog ic
mechanis ms. This trend seems to operate faster in the evolution of
interna l mechanisms than in the evolu tion of extern al behavio r.
Recapi tulating and ex tending what was sa id abov e:
(a)The notion of repeatable context is a necessary premise for
any theory which defines "learning" as change.
(b)This notion is not a mere tool of our description but contains
the implicit hypothesis that for the organisms which we study, the
sequence of life experience, action, etc., is somehow segmented or
punctuated into subsequences or "contexts" which may be equated
or differentiated by the organism.
296

(c)The distinction which is commonly drawn between
perception and action, afferent and efferent, input and out-put, is
for higher organisms in complex situations not valid. On the one
hand, almost every item of action may be re-ported either by
external sense or endoceptive mechanism to the C.N.S., and in this
case the report of this item be-comes an input. And, on the other
hand, in higher organisms, perception is not by any means a
process of mere passive receptivity but is at least partly determined
by efferent control from higher centers. Perception, notoriously,
can be changed by experience. In principle, we must allow both for
the possibility that every item of action or output may create an
item of input; and that percepts may in some cases par-take of the
nature of output. It is no accident that almost all sense organs are
used for the emission of signals between organisms. Ants
communicate by their antennae; dogs by the pricking of their ears;
and so on.
(d)In principle, even in zero learning, any item of experience or
behavior may be regarded as either "stimulus" or "response" or as
both, according to how the total sequence is punctuated. When the
scientist says that the buzzer is the "stimulus" in a given sequence,
his utterance implies an hypothesis about how the organism
punctuates that sequence. In Learning I, every item of perception
or behavior may be stimulus or response or reinforcement
according to how the total sequence of interaction is punctuated.
Learning II
What has been said above has cleared the ground for the
consider ation of the next level or logical type of "learn ing" which
we shall here call Learn ing II. Various terms have been propo sed in
the literatur e for various pheno mena of this order . "Deute ro-
297

learning ,"4 57"set learning,"58 "learning to learn," and "transfer of
learning" may be m entioned .
We recapitulate and e xtend the definitions s o far given:
Zero learning is characterized by specificity of response , which
—righ t or wrong— is no t sub ject to correction.
Learning I is change in specificity of response by correction of
errors of cho ice within a set of alterna tives.
Learning II is change in the process of Learn ing I, e.g., a
corrective change in the set of alternat ives from which choice is
made, or it is a change in how the sequence of exper ience is
punctua ted.
Learning III is change in the process of Learn ing II, e.g., a
corrective change in the system of sets of alternatives from which
choic e is made. (We shall see later that to demand this level of
perfo rmance of some men and some mammals is sometimes
pathogen ic.)
Learning IV would be chang e in Learn ing III, but probably does
not occur in any adult living organism on this earth. Evolut ionar y
proce ss has, however , created organisms whose ontogen y brings
them to Leve l III. The combinat ion of phylogenesis with
ontogene sis, in fact, ach ieves Level IV.
Our immediate task is to give subst ance to the definition of
Learn ing II as "change in Learn ing I," and it is for this that the
ground has been prepa red. Briefly, I believe that the phenomena of
Learn ing II can all be included under the rubric of changes in the
manner in which the stream of action and experi ence is segmented
or punc tuated into contexts together with changes in the use of
contex t markers .
The list of phenomena classified under Learning I includ es a
consid erable (but not exhaus tive) set of differen tly structu red
contex ts. In classical Pavlovi an contexts , the contingency pattern
which describes the relation between "stimulus" (CS), animal's
action (CR), and reinforce ment. (UCS ) is profoundl y different from
57 'G. Bateson , "Social Plannin g and the Concept of Deutero- Learnin g,"
Conference on Science, Philosophy and Religion, Second Symposium,
New York, Harper , 1942 .58 H. E. Harlo w, "The Form ation of Learning Sets," Psycho!. Review, 1949,
56: 51- 65.
298

the contingen cy pattern characteristic of instrumental contexts of
learning.
In the Pavlovian case: If stimulus and a certain lapse of time:
then reinfo rcement.
In the Instrumental Re ward cas e: If stimulus and a pa rticular item
of behav ior: then reinforc ement.
In the Pavlovian case, the reinforce ment is not contingent upon
the animal's behav ior, wher eas in the instrumental case, it is. Using
this contrast as an example, we sa y that Lea rning II has occur red if it
can be shown that experience of one or more contexts of the
Pavlovian type results in the animal's acting in some later context as
though this, too, had the Pavlov ian contingenc y pattern. Similarl y, if
past expe rience of instrumental sequen ces leads an animal to act in
some later contex t as though expec ting this also to be an
instrumental contex t, we shall again say that Learning II has
occurred .
When so defined, Learn ing II is adaptive only if the animal
happ ens to be right in its expec tation of a given contingenc y pattern,
and in such a case we shall expect to see a measurable learning to
learn. It shou ld requi re fewer trials in the new context to establish
"correct" behavio r. If, on the other hand , the animal is wrong in his
identification of the later cont ingenc y pattern, then we shall expect a
delay of Learning I in the new contex t. The animal who has had
prolonged exper ience of Pavlovian contexts might never get around
to the particula r sort of trial-and-er ror behavio r necessar y to
discover a correct instrumental response .
There are at least four fields of exper imentat ion where Learn ing
II has be en ca refully recorded :
(a) In human rote learning . Hull59 carried out very careful
quan titative studie s which revealed this pheno menon, and
construc ted a mathematica l model which would simulate or expla in
the curves of Learning I which he recorded . He also observed a
second-ord er pheno menon which we may call "learn ing to rote
learn" and publi shed the curves for this pheno menon in the
Appendix to his book . These curves were separa ted from the main
59 E. L. Hull, et al., Mathematico-deductive Theory of Rote Learning,
New Haven, Yale University, Institute of Human Relations, 1940
299

body of the book because , as he states, his mathematical model (of
Rote Le arning I) did not cove r this aspec t of the data.
It is a corollary of the theoretical position which we here take
that no amount of rigorous discourse of a given logica l type can
"expl ain" pheno mena of a higher type. Hull's model acts as a
touchs tone of logica l typing, automatica lly exclud ing from
explan ation phenomena beyond its logical scope. That this was so—
and that Hull perc eived it—is testimonial both to his rigor and to his
persp icacity.
What the data show is that for any given subje ct, there is an
improve ment in rote learning with success ive sessions ,
asymptotically approaching a degree of skill which varied from
subje ct to subjec t.
The context for this rote learning was quite complex and no
doubt appear ed subjectively differen t to each learne r. Some may
have been more motivated by fear of being wrong, while others
looked rather for the satisfactions of being right. Some would be
more influenced to put up a good record as compared with the other
subje cts; others would be fascinated to compete in each session with
their own previous showing , and so on. All must have had ideas
(correct or in corre ct) about the na ture of the experi mental set ting, all
must have had "lev els of aspiration," and all must have had prev ious
exper ience of memorizing various sorts of material. Not one of
Hull's subjects could have come into the learning contex t
uninf luenced by previous Learn ing II.
In spite of all this previou s Learning II, and in spite of genet ic
differenc es which might operate at this level, all showed
improve ment over several sessions. This improve ment canno t have
been due t o Lea rning I because any recall of the sp ecific sequence of
syllables learned in the previous session would not be of use in
dealing with the new sequence . Such recall would more probably be
a hindrance . I submit, therefo re, that the improve ment from session
to session can only be accoun ted for by some sort of adapta tion to
the conte xt which Hul l provided for rote learning.
It is also worth noting that educa tors have strong opinion s about
the value (positive or negative) of training in rote learning.
"Progre ssive" educato rs insist on training in "insight," while the
more con serva tive insist on rote and d rilled recall.
300

(b) The second type of Learning II which has been ex-
perimentally studied is called "set learning ." The conc ept and term
are derived from Harlow and apply to a rather special case of
Learning II. Broad ly, what Harlow did was to present rhesus
monkeys with more or less complex gestalten or "problems."
These the monkey had to solve to get a food reward. Harlow
showed that if these problems were of similar "set," i.e., contained
similar types of logical complexity, there was a carry-over of
learning from one problem to the next. There were, in fact, two
orders of contingency patterns involved in Harlow's experiments:
first the overall pattern of instrumentalism (if the monkey solves
the problem, then reinforcement); and second, the contingency
patterns of logic within the specific problems.
(c) Bitterman and others have recently set a fashion in
experimentation with "reversal learning ." Typicall y in these
experiments the subject is first taught a binary discrimination. When
this has been learned to criterion , the meaning of the stimuli is
revers ed. If X initially "meant" R1, and Y initially meant R2, then
after reversa l X comes to mean R2, and, Y comes to mean R1. Again
the trials are run to criterion when again the meanings are reversed .
In these experiments, the crucial question is: Does the subje ct learn
abou t the reversal? I.e., after a series of reversals, does the subject
reach criterion in fewer trials than he did at the beginning of the
series?
In these experiments, it is conspicuously clear that the question
asked is of logical type higher than that of questions about simple
learning. If simple learning is based upon a set of trials, then
reversal learning is based upon a set of such sets. The parallelism
between this relation and Russell's relation between "class" and
"class of classes" is direct.
(d) Learn ing II is also exemplified in the well-known pheno mena
of "exper imental neurosis." Typically an animal is trained, either in
a Pavlovian or instrumental learning con-text, to discriminate
between some X and some Y; e.g., between an ellipse and a circle.
When this discrimination has been learned, the task is made more
difficult: the ellipse is made progr essive ly fatter and the circle is
flattened . Finally a stage is reached at which discrimination is
301

impossib le. At this stage the animal starts to show symptoms of
severe disturbance .
Notabl y, (a) a naive animal, presen ted with a situation in which
some X may (on some rando m basis) mean either A or B, does not
show disturbance ; and (b) the disturbanc e does not occur in
absenc e of the many context markers characteristic of the labora tory
situation.60
It appea rs, then, that Learning II is a necessa ry prepara tion for
the behav ioral disturban ce. The information, "This is a contex t for
discrimination ," is communica ted at the beginning of the sequenc e
and underlined in the series of stages in which discrimination is
made progressively more difficult. But when discrimination
beco mes impossible, the structure of the context is totally changed.
The context markers (e.g., the smell of the laborator y and the
exper imental harness) now beco me mislead ing bec ause t he animal is
in a situation which demands guesswork or gambling, not
discrimination . The en-tire expe rimental sequence is, in fact, a
procedu re for putting the animal in the wrong at the level of
Learn ing 1 1.
In my phrase, the animal is placed in a typical "double bind,"
which i s expect ably schizoph renogenic .61
In the strange world outside the psychologica l laboratory,
pheno mena which belong to the categor y Learning II are a major
preoc cupat ion of anthropolog ists, educa tors, psychiatrists, animal
trainers, human parents, and children . All who think about the
proce sses which determine the character of the individu al or the
proce sses of change in human (or animal) relationsh ip must use in
their think ing a variety of assu mptions abou t Learning II. From time
to time, these people call in the laboratory psycholog ist as a
consul tant, and then are confron ted with a linguistic barrier. Such
barriers must alwa ys result when, for example, the psychiatr ist is
talking abou t Learn ing II, the psychologi st is talking abou t Learning
I, and neither recogn izes the logical structu re of the difference .
60 H. S. Liddell, "Reflex Method and Experi mental Neurosis," Personality
and Behavior Disorders, New York, Ronald Press, 194461 G. Bateson , et al., "Toward a Theor y of Schizophrenia ," Behavioral
Science, 1956, 1: 251-64.
302

Of the multitudinous ways in which Learn ing II emerges in
human affairs, only three wi ll be discus sed in this essay:
(a) In describing individua l human beings , both the scientist and
the layman commonly reso rt to adjec tives descriptive of "chara cter."
It is said that Mr. Jones is dependen t, hostile, fey, finicky, anxious,
exhibitionistic, narcissistic, passive, competitive, energetic, bold,
cowardly, fatalistic, humorous, playful, canny, optimistic,
perfectionist, careless, careful, casual, etc. In the light of what has
already been said, the reader will be able to assign all these
adjectives to their appropriate logical type. All are descriptive of
(possible) results of Learning II, and if we would define these
words more carefully, our definition will consist in laying down
the contingency pattern of that context of Learning I which would
expectably bring about that Learning II which would make the
adjective applicable.
We might say of the "fatalistic" man that the pattern of his
transactions with the environment is such as he might have
acquired by prolonged or repeated experience as subject of
Pavlovian experiment; and note that this definition of "fatalism" is
specific and precise. There are many other forms of "fatalism"
besides that which is defined in terms of this particular context of
learning. There is, for example, the more complex type
characteristic of classical Greek tragedy where a man's own action
is felt to aid the inevitable working of fate.
(b) In the punctuation of human interaction. The critical reader
will have observed that the adjectives above which purport to
describe individual character are really not strictly applicable to
the individual but rather describe transactions between the
individual and his material and human environment. No man is
"resourceful" or "dependent" or "fatalistic" in a vacuum. His
characteristic, whatever it be, is not his but is rather a
characteristic of what goes on between him and something (or
somebody) else.
This being so, it is natural to look into what goes on between
peop le, there to find contex ts of Learning I which are likely to lend
their shape to process es of Learning II. In such systems, involving
two or more persons , where most of the important events are
postures , actions, or utterance s of the living creature s, we note
303

immediate ly that the stream of even ts is commonly punc tuated into
contex ts of learning by a tacit agreement between the persons
regarding the nature of their relationship—o r by contex t markers
and tacit agreement that these context markers shall "mean" the
same for both parties. It is instructive to attempt analysis of an
ongoing interchange between A and B. We ask abou t any particular
item of A's behavio r: Is this item a stimulus for B? Or is it a
respon se of A to something B said earlier? Or is it a reinfor cement
of some item provided by B? Or is A, in this item, consu mmating a
reinforc ement fo r himself? Etc.
Such question s will reveal at once that for many items of A's
behavio r the answer is often quite unclear. Or if there be a clear
answe r, the clarity is due only to a tacit (rarely fully explicit)
agreement between A and B as to the nature of their mutual roles,
i.e., as to the nature of the contextua l structu re which they will
expec t of each other.
If we look at such an exchange in the abstract:
a1b1a2b2a3b3a4b4a5b5 where the a's refer to items
of A's behav ior, and the b's to items of B's behavio r, we can take any
ai and constru ct around it three simple contex ts of learning. These
will be:
(ai bi a,+ 1) , in which ai is the stimulus for bi.
(b1_1 ai bi) , in which ai is the response to b.-1, which response B
reinforc es wi th bi.
(a1_1 bi _1 ai) , in which ai is now A's reinforce ment of B's bi-1,
which was respon se to ai_1.
It follows that ai may be a stimulus for B or it may be A's
respon se to B, or it may be A's reinfo rcement of B.
Beyond this, however , if we conside r the ambiguit y of the
notions "stimulus" and "response ," "afferen t" and "efferent"—as
discu ssed above—we note that any ai may also be a stimulus for A;
it may be A's reinfo rcement of self; or it may be A's respons e to
some previous behavio r of his own, as is the case in sequenc es of
rote behav ior.
This genera l ambiguit y means in fact that the ongo ing sequence
of interch ange between two persons is structu red only by the
person 's own perception of the sequence as a series of cont exts, each
contex t leading into the next. The particular manner in which the
304

sequence is structur ed by any particular person will be determined
by that person's previous Learn ing II (or possib ly by his genet ics).
In such a system, words like "dominant " and "submissive,"
"succoring" and "dependen t" will take on definab le meaning as
descrip tions of segments of interchange . We shall say that "A
dominates B" if A and B show by their behav ior that they see their
relationsh ip as characterized by sequence s of the type a1b1a2, where
a1 is seen (by A and B) as a signal defining condi tions of
instrumental reward or punish ment; b1 as a signa l or act obeying
these condi tions ; and a2 as a signa l reinforcing b1.
Similar ly we shall say that "A is dependent on B" if their
relationsh ip is characterized by sequenc es a1b1a2,, where al is seen as
a signal of weakness ; b1 as a helping act; and a2 as an
acknowledge ment o f b1.
But it is up to A and B to distingui sh (consciou sly or un-
conscious ly or not at all) between "dominance" and " dependence ." A
"command" c an closel y resemble a cry for "help ."
(c) In psychother apy, Learning II is exemplified most con-
spicuous ly by the pheno mena of "transfe rence ." Orthodox Freud ian
theory asserts that the patient will inevitably bring to the therap y
room inapprop riate notions abou t his relation-s hip to the therapist.
These notions (consc ious or unconsciou s) will be such that he will
act and talk in a way which would press the therapist to respond in
ways which would resemble the patient's picture of how some
important other person (usua lly a parent) treated the patient in the
near or distan t past. In the language of the present pape r, the patient
will try to shape his interchange with the therapi st according to the
premises of his (the patient's) former Learn ing II.
It is commonly observed that much of the Learning II which
determines a patient's transfer ence patterns and, in-deed , determines
much of the relationa l life of all human beings , (a) dates from ea rly
infancy , and (b) is uncons cious . Both of these genera lizations seem
to be correct and both n eed so me explan ation .
It seems probable that these two gene ralizations are true because
of the very nature of the pheno mena which we are discussing . We
sugg est that what is learned in Learning II is a way of punc tuating
events . But a way of punc tuating is not true or false. There is
nothing contained in the proposi tions of this learning that can be
305

tested agains t reality. It is like a picture seen in an inkblo t; it has
neither correctness nor incorrectness . It is only a way of seeing the
inkblot.
Consider the instrumental view of life. An organis m with this
view o f life in a new situation wi ll engage in trial-and-e rror behavio r
in order to make the situation prov ide a positive reinforc ement. If he
fails to get this reinfo rcement, his purpo sive philosoph y is not
thereby negat ed. His trial-and -error behavior will simply continue.
The premises of "purpose" are simply not of the same logical type
as the material facts of life, and therefore cannot easily be
contradicted by them.
The practitioner of magic does not unlearn his magical view of
events when the magic does not work. In fact, the propositions
which govern punctuation have the general characteristic of being
self-validating.62 What we term "con-text" includes the subject's
behavior as well as the external events. But this behavior is
controlled by former Learning II and therefore it will be of such a
kind as to mold the total context to fit the expected punctuation. In
sum, this self-validating characteristic of the content of Learning II
has the effect that such learning is almost ineradicable. It follows
that Learning II acquired in infancy is likely to persist through life.
Conversely, we must expect many of the important characteristics
of an adult's punctuation to have their roots in early infancy.
In regard to the unconsciousness of these habits of punctuation,
we observe that the "unconscious" includes not only repressed
material but also most of the processes and habits of gestalt
perception. Subjectively we are aware of our "dependency" but
unable to say clearly how this pattern was constructed nor what
cues were used in our creation of it.
62 J. Ruesch and G. Bateson, Communica tion: The Social Matrix of
Psychiatr y, New York, Norton, 1951.
306

Learning III
What has been said above abou t the self-validating character of
premises acqui red by Learn ing II indica tes that Learning III is likely
to be difficult and rare even in human beings . Expec tably, it will
also be difficult for scientists, who are only human, to imagine or
describ e this proces s. But it is claimed that something of the sort
does from time to time occu r in ps ychotherap y, religious conver sion,
and in other sequenc es in which there is profound reorganiza tion of
character.
Zen Buddhists, Occidental mystics, and some psychiatrists
assert that these matters are totally beyond the reach of language.
But, in spite of this warning, let me begin to speculate about what
must (logically) be the case.
First a distinction must be drawn: it was noted above that the
experiments in reversal learning demonstrate Learning II whenever
there is measurable learning abou t the fact of reversal. It is
possible to learn (Learning I) a given premise at a given time and
to learn the converse premise at a later time without acquiring the
knack of reversal learning. In such a case, there will be no
improvement from one reversal to the next. One item of Learning I
has simply re-placed another item of Learning I without any
achievement of Learning II. If, on the other hand, improvement
occurs with successive reversals, this is evidence for Learning II.
If we apply the same sort of logic to the relation between
Learning II and Learning III, we are led to expect that there might
be replacement of premises at the level of Learning II without the
achievement of any Learning III.
Preliminary to any discussion of Learning III, it is there-fore
necessary to discriminate between mere replacement without
Learning III and that facilitation of replacement which would be
truly Learning III.
That psychotherapists should be able to aid their patients even
in a mere replacement of premises acquired by Learning II is
already no mean feat when we consider the self-validating
307

character of such premises and their more or less unconscious
nature. But that this much can be done there is no doubt.
Within the controlled and protected setting of the therapeutic
relationship, the therapist may attempt one or more of the
following maneuvers:
(a)to achieve a confrontation between the premises of the
patient and those of the therapist—who is carefully trained
not to fall into the trap of validating the old premises;
(b)to get the patient to act, either in the therapy room or
outside, in ways which will confront his own premises;
(c)to demonstrate contradiction among the premises which
currently control the patient's behavior;
(d)to induce in the patient some exaggera tion or caricature
(e.g., in dream or hypnosis) of experience based on his old
premises.
As William Blake noted , long ago, "Withou t Contra ries is no
progr ession ." (Elsewher e I have called these contradictions at level
II "doubl e bind s.")
But there are alwa ys loopho les by which the impact of
contradic tion can be reduc ed. It is a commonplac e of learning
psycholog y that while the subject will learn (Learning I) more
rapidly if he is reinfo rced every time he responds correctly, such
learning will disappear rather rapidly if reinforcement ceases . If, on
the other hand , reinfor cement is only occas ional , the subjec t will
learn more slowl y but the resulting learning will not easily be
extinguished when reinforcement ceases altoge ther. In other words ,
the subje ct may learn (Learning 11) that the contex t is such that
absenc e of reinforc ement does not indicate that his respon se was
wrong or inappropr iate. His view of the contex t was, in fact, correct
until the exper imenter changed his tactics.
The the rapis t must certainly so suppor t or hedg e the contraries b y
which the patient is driven that loophol es of this and other kinds are
blocked . The Zen cand idate who has been assigned a paradox
(koan ) must labor at his task "like a mosquito biting on an iron ba r."
I have argued elsewhere ("Style, Grace , and Information in
Primitive Art," see p. 128) that an essen tial and necessary function
of all habit formation and Learning I1 is an economy of the though t
proce sses (or neura l pathwa ys) which are used for proble m-solving
308

or Learning I. The premises of what is commonly called
"character"— the definitions of the "self" —sav e the individu al from
having to examine the abstract, philosophi cal, aesthe tic, and ethical
aspects of many sequen ces of life. "I don't know whether it's good
music; I only know whethe r I like it."
But Learning III will throw these unex amined premises open to
question and change .
Let us, as was done above for Learn ing I and II, list some of the
changes whi ch we sh all be willing to call Learn ing III.
(a)The individual might learn to form more readily those habits
the forming of which we call Learning II.
(b)He might learn to close for himself the "loopholes" which
would allow him to avoid Learning III.
(c)He might learn to change the habits acquired by Learning II.
(d)He might learn that he is a creature which can and does
unconsciously achieve Learning II.
(e)He might learn to limit or direct his Learning II.
(f)If Learning II is a learning of the contexts of Learning I,
then Learning III should be a learning of the contexts of those
contexts.
But the above list proposes a paradox . Learn ing III (i.e., learning
about Learn ing II) may lead either to an increas e in Learn ing II or
to a limitation and perhaps a reduct ion of that phenomenon.
Certainly it must lead to a greater flexibility in the premises
acquired by the process of Learn ing II —a freedom from their
bondag e.
I once heard a Zen master state categori cally: "To become
accusto med to anything is a terrible thing."
But any freedo m from the bond age of habit must also denot e a
profound redefinition of the self. If I stop at the level of Learn ing II,
"I" am the aggregate of those charac teristics which I call my
"character." "I" am my habits of acting in context and shaping and
perceiv ing the contex ts in which I act. Selfhood is a product or
aggregat e of Learn ing II. To the degree that a man achieves
309

Learn ing III, and learns to perceive and act in terms of the contexts
of contexts , his "self" will take on a sort of irrelevan ce. The concept
of "self" will no longer function as a nodal argument in the
punctua tion of exper ience .
This matter needs to be examined. In the discussion of Learning
II, it was asserted that all words like "dependency," "pride,"
"fatalism," refer to characteristics of the self which are learned
(Learning II) in sequences of relationship. These words are, in fact,
terms for "roles" in relationships and refer to something artificially
chopped out of interactive sequences. It was also suggested that
the correct way to assign rigorous meaning to any such words is to
spell out the formal structure of the sequence in which the named
characteristic might have been learned. Thus the interactive
sequence of Pavlovian learning was proposed as a paradigm for a
certain sort of "fatalism," etc.
But now we are asking about the contexts of these con-texts of
learning, i.e., about the larger sequences within which such
paradigms are embedded.
Consider the small item of Learning II which was mentioned
above as provid ing a "loopho le" for escape from Learning III. A
certain characte ristic of the self—cal l it "pers istence"—is gener ated
by exper ience in multiple sequen ces among which reinforc ement is
sporad ic. We must now ask about the larger context of such
sequenc es. How are such s equences gene rated ?
The question is explosive . The simple stylized expe rimental
sequenc e of interac tion in the laboratory is gene rated by and partly
determines a network of contingenc ies which goes out in a hundred
directions leading out of the labora tory into the processe s by which
psycholog ical research is designed, the interactions between
psycholog ists, the econo mics of re-search money, etc., etc.
Or consider the same formal sequence in a more "natural"
setting. An organism is searching for a needed or missing object. A
pig is rooting for acorns , a gambler is feeding a slot machine hoping
for a jackpo t, or a man must find the key to his car. There are
thousand s of situations where living thing s must persist in certain
sorts of behav ior precisely because reinforc ement is sporadic or
improbab le. Learn ing II will simplif y the universe by hand ling these
instances as a single categor y. But if Learning III be concern ed with
310

the contexts of these instan ces, then the categorie s of Learn ing II
will be bu rst open.
Or consider what the word "reinforcement" means at the
various levels. A porpoise gets a fish from the trainer when he does
what the trainer wants. At level I, the fact of the fish is linked with
the "rightness" of the particular action. At level II, the fact of the
fish confirms the porpoise's under-standing of his (possibly
instrumental or dependent) relationship with the trainer. And note
that at this level, if the porpoise hates or fears the trainer, pain
received from the latter may be a positive reinforcement
confirming that hate. ("If it's not the way I want it, I'll prove it.")
But what of "reinforcement" at level III (for porpoise or for
man)?
If, as I have sugg ested a bove, t he cre ature is dr iven to l evel III by
"contraries" genera ted at level II, then we may expec t that it is the
resolv ing of these contra ries that will constitute positive
reinfo rcement at level III. Such resolut ion c an take m any forms.
Even the attempt at level III can be dange rous, and some fall by
the wayside. These are often labeled by psychiatry as psychotic , and
many of them find themselves inhibited from using the first person
pronoun.
For others, more success ful, the resolu tion of the contraries may
be a collapsing of much that was learned at level II, revealing a
simplicity in which hunger leads directly to eating, and the
identified self is no longer in charge of organizing the behavio r.
These are the incorruptib le innocen ts of the world .
For others , more creative, the resolut ion of contraries reveals a
world in which persona l identity merges into all the processe s of
relationsh ip in some vast ecolog y or aesthetics of cos mic in teraction.
That any of these can survive seems almost miraculous , but some
are perhaps saved from being swep t away on ocean ic feeling by
their ability to focus in on the minutiae of life. Every detail of the
universe is seen as propos ing a view of the whole . These are the
peop le for who m Blake wrote the famous advice in the "Augur ies of
Innocence :"
To see the World in a Grain of Sand, And a
Heaven in a Wild Flower,
311

Hold Infinity in the palm of your hand, And
Eternity in an hour.
The Rol e of Genet ics in Psych ology
Whatever can be said about an animal's learning or in-ability to
learn has bearing upon the genetic make-up of the animal. And what
has been said here about the levels of learning has bearing upon the
whole interplay between gene tic make-up and the changes which
that individu al can a nd m ust achieve.
For any given organism, there is an upper limit beyond which all
is determined by genetics. Planar ians can probabl y not go beyond
Learn ing I. Mammals other than man are probab ly capable of
Learn ing II but incapab le of Learn ing III. Man may sometimes
achieve Lea rning III.
This upper limit for any organism is (logic ally and presumably)
set by genetic pheno mena, not perhaps by individual gene s or
combinat ions of genes , but by whatever factors control the
development of basic phylar characteristics.
For every change of which an organism is capable, there is the
fact of that capability. This fact may be genetically determined; or
the capability may have been learned. If the latter, then genetics
may have determined the capability of learning the capability. And
so on.
This is in general true of all somatic changes as well as of those
behavioral changes which we call learning. A man's skin tans in the
sun. But where does genetics enter this picture? Does genetics
completely determine his ability to tan? Or can some men increase
their ability to tan? In the latter case, the genetic factors evidently
have effect at a higher logical level.
The problem in regard to any behavior is clearly not "Is it
learned or is it innate?" but "Up to what logical level is learning
effective and down to what level does genetics play a
determinative or partly effective role?"
312

The broad history of the evolution of learning seems to have
been a slow pushing back of genetic determinism to levels of
higher logical type.
A Note on H ierarchi es
The model discuss ed in this paper assumes, tacitly, that the
logica l types can be ordered in the form of a simple, unbranch ing
ladder . I believe that it was wise to deal first with the proble ms
raised by such a simple model.
But the world of action, expe rience, organiza tion, and learning
cannot be completel y mapped onto a model which excludes
proposi tions abou t the relation between classe s of differen t logical
type.
If CI is a class of proposi tions , and C2 is a class of proposi tions
abou t the m embers of C1; C3 then be ing a c lass of p roposi tions abou t
the members of C2; how then shall we classify propos itions abou t
the relation between these classes? For e xample, the propos ition "As
members of C1 are to members of C2, so members of C2 are to
members of C3" canno t be classified within the unbranch ing ladde r
of types.
The whole of this essa y is built upon the premise that the relation
between C2 and C3 can be compared with the relation between C1
and C2. I have again and again taken a stance to the side of my
ladder of logical types to discuss the structure of this ladder. The
essay is therefore itself an example of the fact that the ladder is not
unbranching.
It follows that a next task will be to look for examples of
learning which cannot be classified in terms of my hierarchy of
learning but which fall to the side of this hierarchy as learning
about the relation between steps of the hierarchy. I have suggested
elsewhere ("Style, Grace, and Information in Primitive Art") that
art is commonly concerned with learning of this sort, i.e., with
bridging the gap between the more or less unconscious premises
acquired by Learning II and the more episodic content of
consciousness and immediate action.
313

It shou ld also be not ed that the st ructur e of th is essay is induc tive
in the sense that the hierarchy of order s of learning is presented to
the reader from the bottom upward , from level zero to level III. But
it is not intended that the explanations of the pheno menal world
which the model affords shall be unidirectiona l. In expla ining the
model to the reader, a unidirection al appro ach was neces sary, but
within the model it is assumed that higher levels are explan atory of
lower levels and vice versa. It is also assumed that a similar
reflexiv e relation—both inductive and deduct ive—obtains among
ideas and items of learning as these exist in the lives of the crea tures
which we study.
Finally, the model remains ambiguous in the sense tha t while it is
asserted that there are explanato ry or determinative relations
between ideas of adjacen t levels both upward and down ward , it is
not clear whether direct explana tory relation s exist between
separated levels, e.g., between level III and level I or between level
zero and level II.
This question and that of the status of propos itions and ideas
collateral to the hierarch y of types remains unexamined.
314

The Cybern etics o f "Self": A Theory of
Alcohol ism*
The "logic" of alcohol ic addiction has puzz led psychiatrists no
less than the "logic" of the strenuou s spiritual regime whereb y the
organiza tion Alcoholi cs Anon ymous is able to coun teract the
addiction. In the present essay it is sugges ted: (1) that an entirely
new epistemology must come out of cybernetics and systems theory,
involving a new understanding of mind, self, human relationship ,
and power; (2) that the addic ted alcoho lic is opera ting, when sober ,
in terms of an epistemology which is convent ional in Occiden tal
culture but which is not accept able to systems theory; (3) that
surrende r to alcohol ic intoxica tion provides a partial and subjective
short cut to a more correct st ate of mind; and (4 ) that the theo logy of
Alcoho lics Anon ymous coinc ides closely with an epistemolog y of
cybernet ics.
The present essay is based upon ideas which are, perhaps all of
them, familiar either to psychiatrists who have had dealings with
alcohol ics, or to philosoph ers who have though t about the
implica tions of cybernetics and systems theory. The only novel ty
which can be claimed for the thesis here offered deriv es from
treating these ideas seriously as premises of argument and from the
bringing together of commonplace ideas from two too separ ate
fields of though t.
In its first concep tion, this essay was plann ed to be a systems-
theore tic study of alcohol ic addic tion, in which I would use data
from the publications of Alcoholi cs Anon ymous, which has the only
outstand ing record of success in dealing with alcohol ics. It soon
became evident, howeve r, that the religious views and the
organiza tiona l structu re of AA presen ted points of great interest to
systems theory, and that the correct scope of my study should
include not only the premises of alcoholi sm but also the premises of
the AA system of treating it and th e premises of AA organiza tion.
* This article appeared in Psychiatry, Vol. 34, No. 1, pp. 1-18, 1971. Copyright
© 1971 by the William Alanson White Psychiatric Foundation. Reprinted by
permission of Psychiatry
315

My debt to AA will be evident throughout— also, I hope , my
respect for that organization and especially for the extraord inary
wisdo m of its cofounders , Bill W. and Dr. Bob.
In addition, I have to acknowledge a debt to a small sample of
alcoho lic patients with whom I worked intens ively for about two
years in 1949-52 , in the Veterans Administ ration Hospi tal, Palo Alto,
Californi a. These men, it should be mentioned, carried other
diagnos es—mostly of "schizoph renia"—in addition to the pains of
alcoho lism. Sever al were members of AA. I fear that I helped them
not at all.
The Prob lem
It is rather gene rally believed that "causes" or "reasons" for
alcoho lism are to be looked for in the sober life of the alcohol ic.
Alcohol ics, in their sober manifestations , are commonly dubbed
"immature ," "materna lly fixated," "oral," "homosexual ," "pass ive-
aggre ssive ," "fearful of success ," "oversens itive," "proud ," "affable,"
or simply "weak ." But the logical implica tions of this belief are
usual ly not exa mined:
(1)If the sober life of the alcohol ic somehow drives him to drink
or proposes the first step toward intoxication, it is not to be
expec ted that any procedure which reinforce s his particula r style of
sobriety will reduce or con trol his alcoholism.
(2)If his style of sobriety drives him to drink, then that style must
conta in error or p atholog y; and in toxica tion m ust provide so me—at
least subjec tive— corre ction of this error. In other words, compared
with his sobriety, which is in some way "wrong," his intoxication
must be in some way "right." The old tag In vino veritas may
contain a truth more profound than is usually attributed to it.
(3)An alternative hypothesis would suggest that when sober,
the alcoholic is somehow more sane than the people around him,
and that this situation is intolerable. I have heard alcoholics argue
316

in favor of this possibility, but I shall ignore it in this essay. I
think that Bernard Smith, the non-alcoholic legal representative of
AA, came close to the mark when he said, "the [AA] member was
never enslaved by alcohol. Alcohol simply served as an escape
from personal enslavement to the false ideals of a materialistic
society."63 It is not a matter of revolt against insane ideals around
him but of escaping from his own insane premises, which are
continually reinforced by the surrounding society. It is possible,
however , that the alcoho lic is in some way more vulnerable or
sensitive than the normal to the fact that his insane (but
conventional) premises lead to unsatisfying results.
(4)The present theory of alcoholism, therefore, will pro-vide a
converse matching between the sobriety and the intoxication, such
that the latter may be seen as an appropriate subjective correction
for the former.
(5)There are, of course, many instances in which people resort
to alcohol and even to extreme intoxication as an anesthetic
giving release from ordinary grief, resentment, or physical pain.
It might be argued that the anesthetic action of alcohol provides a
sufficient converse matching for our theoretical purposes. I shall,
however, specifically exclude these cases from consideration as
being not relevant to the problem of addictive or repetitive
alcoholism; and this in spite of the undoubted fact that "grief,"
"resentment," and "frustration" are commonly used by addicted
alcoholics as excuses for drinking.
I shall demand, therefore , a conve rse matching between sobriety
and intoxication more specific than that provided by mere
anesthes ia.
63 [Alcoholics Anon ymous], Alcoholics Anony mous Comes of Age, New
York, Harper, 1957, p. 279. (Italics added.)
317

Sobriety
The friends and relatives of the alcohol ic commonly urge him to
be "strong," and to "resist temptation ." What they mean by this is
not very clear, but it is significan t that the alcoho lic himself—whi le
sober— commonly agrees with their view of his "prob lem." He
believes that he could be, or, at least, ough t to be "the captain of his
soul."64 But it is a cliche of alcohol ism that after "that first drink ,"
the motivat ion to stop drink ing is zero. Typicall y the whole matter is
phras ed overt ly as a battle between "self" and "John Barleycorn."
Covert ly the alcohol ic may be planning or even secretly laying in
suppl ies for the next binge , but it is almost impossible (in the
hospi tal setting) to get the sober alcoho lic to plan his next binge in
an overt manner. He cannot , seemingly, be the "capt ain" of his soul
and overtly will or command his own drunkenness . The "captain"
can only command sob riety —and then not be obe yed.
Bill W., the cofounder of Alcohol ics Anon ymous, himself an
alcoho lic, cut through all this mythology of conflict in the very first
of the fa mous "Twelve Steps" of AA. The first step de mands that the
alcoho lic agree that he is power less over alcoho l. This step is.
usual ly regarded as a "surrender" and many alcoholi cs are either
unable to achieve it or achieve it only briefly during the period of
remorse following a binge . AA does not regard these cases as
promising: they have not yet "hit bottom"; their despa ir is
inadequ ate and after a more or less brief spell of sobriety they will
again attempt to use "self-cont rol" to fight the "temptation." They
will not or canno t accept the premise that, drunk or sober , the total
person ality of an alcoho lic is an alcoho lic person ality which canno t
conce ivabl y fight alcohol ism. As an AA leaflet puts it, "trying to use
will power is like trying to lift yourself by your boo tstraps. "
The first two steps of AA are as follows:
64 ' This phrase is used by AA in derision of the alcoholic who tries to use will power
against the bottle. The quotation, along with the line, "My head is blood y but un-
bowed," comes from the poem "Invictus " by William Ernest Henley, who was a
cripple but not an alcoholic. The use of the will to conquer pain and physical dis-
abilit y is probabl y not comparable to the alcoholic 's use of will.
318

1.We admitted we were powerl ess over alcoho l—tha t our
lives had beco me unmanageab le.
2.Came to believe that a Power greater than our-selves
could restore us to sanity.65
Implici t in the combination of these two steps is an extraord inary
—and I believe correct— idea: the expe rience of defea t not only
serves to convinc e the alcohol ic that change is necessary; it is the
first step in that change . To be defea ted by the bottle and to know it
is the first "spiritual experi ence. " The myth of self-power is thereby
broken b y the demonstra tion of a greater power .
In sum, I shall argue that the "sobriety" of the alcoholi c is
characterized by an unusua lly disastrous variant of the Cartesian
dualism, the division between Mind and Matter, or, in this case,
between conscious will, or "self," and the remainder of the
personal ity. Bill W.'s stroke of genius was to break up with the first
"step" the structu ring o f this dualism.
Philosophi cally viewed, this first step is not a surrender ; it is
simply a change in epistemology, a change in how to know abou t
the persona lity-in-the-wor ld. And, notably, the change is from an
incorr ect to a more correct epistemolog y.
Epist emology and Ontology
Philosophe rs have recognized and separ ated two sorts of
proble m. There are first the proble ms of how things are, what is a
person, and what sort of a world this is. These are the problems of
ontolog y. Second , there are the problems of how we know anything,
or more speci fically, how we know what sort of a world it is and
what sort of creatures we are that can know something (or perhaps
nothing) of this matter. These are the problems of epistemology. To
these questions, both ontological and epistemological, philos-
ophers try to find true answers.
65 '[Alcoholics Anon ymous], Alcoholics Anonymous , New York, Works
Publishing, 1939
319

But the natura list, observing human behavio r, will ask rather
differen t quest ions. If he be a cultural relativist, he may agree with
those philosoph ers who hold that a "true" ontolog y is conce ivable ,
but he wi ll not ask whethe r the on tolog y of th e people h e observe s is
"true." He will expect their epistemology to be culturally determined
or even idiosyncrat ic, and he will expect the culture as a whole to
make sen se in terms of their particular epistemology and ontology.
If, on the other hand , it is clear that the local epistemolog y is
wrong, then the naturalist should be alert to the possib ility that the
culture a s a whole wil l never r eally make "sense ," or wil l make sense
only under restricted circumstances, which contac t with other
cultures and new technologi es might di srupt.
In the natural history of the living human being, ontolog y and
epistemology cannot be separa ted. His (commonly unconsc ious)
beliefs abou t what sort of world it is will determine how he sees it
and acts within it, and his ways of perceiving and acting will
determine his beliefs about its nature. The living man is thus bound
within a net of epistemologica l and ontologic al premises which—
regardles s of ultimate truth or falsity—beco me partially self-
validating fo r him66
It is awkward to refer constan tly to both epistemolog y and
ontology and incor rect to sugges t that they are separable in human
natural history. There seems to be no conven ient word to cove r the
combinat ion of these two concep ts. The neare st approx imation s are
"cogni tive structure" or "character structure ," but these terms fail to
sugges t that what is importan t is a body of habitu al assumptions or
premises implici t in the relationsh ip between man and environment,
and that these premises may be true or false. I shall there-fore use
the single t erm "epistemolog y" in this essay to cove r both a spects o f
the net of premises which govern adapt ation (or maladapta tion) to
the human and physical environment. In Geor ge Kelly's vocabu lary,
these are the rules by which an indiv idual "cons trues " his
exper ience .
I am conce rned espec ially with that group of premises upon
which Occiden tal concepts of the "self" are built, and conve rsely,
66 J. Ruesch and G. Bateson, Communications: The Social Matrix of Psychiatry ,
New York, Norton, 1951.
320

with premises which are corrective to some of the more gross
Occidenta l errors associa ted wi th that concept.
The Epist emology o f Cyberne tics
What is new and surprising is that we now have partial answers
to some of these quest ions. In the last twent y-five years
extraordina ry advances have been made in our knowledge of what
sort of thing the environment 'is, what sort of thing an organism is,
and, especi ally, what sort of thing a mind is. These advan ces have
come out of cybernetics, systems theory, information theory, and
related scienc es.
We now know , with consid erabl e certainty, that the ancient
proble m of whethe r the mind is immanent or transcendent can be
answered in favor of immanence , and that this answer is more
econo mical of explana tory entities than any transcenden t answer : it
has at least the nega tive sup-po rt of Occa m's Razor .
On the positive side, we can assert that any ongo ing ensemble of
events and objects which has the appropr iate complexit y of causal
circuits and the appropr iate energy relations will sure ly show mental
characteristics. It will compare, that is, be responsive to difference
(in addition to being affected by the ordina ry physical "cause s" such
as impact or force ). It will "proc ess information " and will inevitably
be self-corrective either toward homeosta tic optima or toward the
maximization of certain variable s.
A "bit" of information is definab le as a difference which makes a
difference . Such a difference , as it travels and undergoes succes sive
transformation in a circuit, is an elementary idea.
But, most relevan t in th e presen t cont ext, we know tha t no par t of
such an in terna lly interactive system can have un ilateral con trol ove r
the remainder or over any other part. The mental characteristics are
inheren t or immanent in the ensemble as a whole.
Even in very simple self-cor rective systems, this holistic
character is eviden t. In the steam engine with a "governo r," the very
word "governor" is a misnomer if it be taken to mean that this part
of the system has unilateral control. The governor is, essentially, a
321

sense organ or transducer which receives a transform of the
difference between the actual running speed of the engine and
some ideal or preferred speed. This sense organ transforms these
differences into differences in some efferent message, for example,
to fuel supply or to a brake. The behavior of the governor is de-
termined, in other words, by the behavior of the other parts of the
system, and indirectly by its own behavior at a previous time.
The holistic and mental character of the system is most clearly
demonstrated by this last fact, that the behavior of the governor
(and, indeed, of every part of the causal circuit) is partially
determined by its own previous behavior. Message material (i.e.,
successive transforms of difference) must pass around the total
circuit, and the time required for the message material to return to
the place from which it started is a basic characteristic of the total
system. The behavior of the governor (or any other part of the
circuit) is thus in some degree determined not only by its
immediate past, but by what it did at a time which precedes the
present by the interval necessary for the message to complete the
circuit. There is thus a sort of determinative memory in even the
simplest cybernetic circuit.
The stability of the system (i.e., whether it will act self-
correctively or oscillate or go into runaway) depends upon the
relation between the operational product of all the transformations
of difference around the circuit and upon this characteristic time.
The "governor" has no control over these factors. Even a human
governor in a social system is bound by the same limitations. He is
controlled by information from the system and must adapt his own
actions to its time characteristics and to the effects of his own past
action.
Thus, in no system which shows mental characteristics can any
part have unilateral contro l over the whole . In other words, the
mental charac teristics of the system are immanent, not in some part,
but in the system as a whole .
The significance of this conclusion appear s when we ask, "Can a
computer think?" or, "Is the mind in the brain?" And the answer to
both quest ions will be negative unless the question is focused upon
one of the few mental characte ristics which are contained within the
computer or the brain. A computer is self-cor rective in regard to
322

some of its internal variables. It may, for example, include
thermometers or other sense organs which are affected by
differences in its working temperatu re, and the response of the sense
organ to these differences may affect the action of a fan which in
turn corrects the temperature . We may therefo re say that the system
shows mental characteristics in regard to its internal temperatu re.
But it would be incorrect to say that the main busines s of the
computer—the transformation of input differenc es into output
differences— is "a mental proce ss." The computer is only an are of a
larger circuit which always includes a m an and an environment fro m
which information is received and upon which efferent messages
from the computer have effect. This total system, or ensemble, may
legitimatel y be said to show mental characteristics. It opera tes by
trial and error and h as creative character.
Similar ly, we may say that "mind" is immanent in those circuits
of the brain which are complete within the brain. Or that mind is
immanent in circuits which are complete within the system, brain
plus body. Or, finally, that mind is immanent in the larger system—
man plus environment.
In princip le, if we desire to explain or unders tand the mental
aspect of any biolog ical event , we must take into account the system
—that is, the network of closed circuits, within which that
biologi cal event is determined. But when we seek to expla in the
behavior of a man or any other organism, this "system" will usually
not have the same limits as the "self"—as this term is commonly
(and va rious ly) understood .
Consid er a man felling a tree with an axe. Each stroke of the axe
is modified or corrected, according to the shape of the cut face of the
tree left by the previous strok e. This self-cor rective (i.e., mental)
process is brough t about by a total system, tree-eyes-brain-muscles –
axe-stroke-t ree; and it is this total system that has the characteristics
of immanent mind.
More correctly, we shou ld spell the matter out as: (differences in
tree) – (differences in retina) -(differences in brain) – (differenc es in
muscles) -(differen ces in movement of axe) -(differenc es in tree),
etc. What is transmitted around the circuit is transforms of
differences. And, as noted above, a difference which makes a
difference is an idea or unit of information.
323

But this is not how the average Occidental sees the event
sequence of tree felling. He says, "I cut down the tree" and he even
believes that there is a delimited agent, the "self," which performed
a delimited "purposive" action upon a de-limited object.
It is all very well to say that "Billiard ball A hit billiard ball B
and sent it into the pocket"; and it would perhaps be all right (if we
could do it) to give a complete hard-science account of the events
all around the circuit containing the man and the tree. But popular
parlance includes mind in its utterance by invoking the personal
pronoun, and then achieves a mixture of mentalism and
physicalism by restricting mind within the man and reifying the
tree. Finally the mind itself becomes reified by the notion that,
since the "self" acted upon the axe which acted upon the tree, the
"self" must also be a "thing." The parallelism of syntax between "I
hit the billiard ball" and "The ball hit another ball" is totally
misleading.
If you ask anybody abou t the localization and bound aries of the
self, these confusions are immediately displayed. Or conside r a
blind man with a stick. Where does the blind man's self begin? At
the tip of the stick? At the hand le of the stick? Or at some point
halfway up the stick? These questions are nonsense , because the
stick is a pathwa y along which difference s are transmitted unde r
transfor mation, so tha t to draw a delimiting line across this pathwa y
is to cut off a part of the systemic circuit which determines the blind
man's locomotion .
Similarly, his sense organs are transdu cers or pathwa ys for
information , as also are his axons, etc. From a systems-theore tic
point of view, it is a misleading metaphor to say that what travels in
an axon is an "impulse." It would be more correct to say that what
travels is a differenc e, or a transfo rm of a difference . The metaphor
of "impulse" suggests a hard-science line of thought which will
ramify only too easily into nonsens e about "psychic energy," and
those who talk this kind of nonsense will disrega rd the information
conten t of quiescence. The quies cence of an axon differs as much
from activity as its activity does from quies cence . Therefore
quies cence and activity have equal informationa l relevance. The
message of activity can only be accepted as valid if the message of
quies cence can a lso be trusted.
324

It is even incorrect to speak of the "message of activity" and the
"message of quiescence ." Alwa ys the fact that in-formation is a
transform of differenc e should be remembered, and we might better
call the one message "activity —not quiescence" and the other "quiescenc e—not activity."
Similar considerations apply to the repentan t alcoholi c. He
cannot simply elect "sobriety." At best he could only elect "sobriety
—no t drunk enness ," and his universe remains polarized, carrying
always both alternatives.
The total self-cor rective unit which processes information, or, as
I say, "thinks " and "acts" and "decides," is a system whose
bounda ries do not at all coincide with the boundaries either of the
body or of what is popula rly called the "self" or "cons ciousnes s";
and it is important to notice that there are multiple differenc es
between t he thinking s ystem and t he "self" as popul arly conce ived:
The system is not a transc endent entity as the "self" is commonly
suppo sed to be.
The ideas are immanent in a network of causal path-wa ys along
which transforms of difference are conducted . The "ideas" of the
system are in all cases at least binary in structure . They are not
"impulses" but "info rmation ."
This network of pathwa ys is not bounded with consciousnes s but
extends to include the pathwa ys of all unconsc ious mentation—bo th
autonomic and repress ed, neura l and hormonal.
The network is not bounded by the skin but include s all external
pathwa ys along which information can travel. It also includes those
effective differences which are immanent in the "objects" of such
information. It includes the path-ways of sound and light along
which travel transfo rms of difference s originally immanent in things
and other peop le —and e specia lly in our own actions.
It is importan t to note that the basic—and I believe erroneous—
tenets of popula r epistemology are mutually rein-f orcing. If, for
example, the popula r premise of transcendence is discarded , the
immediate substitute is a premise of immanence in the body. But
this alterna tive will be unacc eptab le because large parts of the
thinking network are located outside the body. The so-called "Bod y-
Mind" problem is wrongl y posed in terms which force the argument
toward paradox : if mind be supposed immanent in the body, then it
325

must be transcend ent. If transcend ent, it must be immanent. And so
on.67
Similarly, if we exclude the unconsc ious proce sses from the
"self" and call them "ego-al ien," then these proces ses take on the
subje ctive coloring of "urges" and "forc es"; and this pseudod ynamic
quality is then extend ed to the consciou s "self" which attempts to
"resist" the "forces" of the unconsc ious. The "self" thereby becomes
itself an organization of seeming "forces ." The popul ar notion which
would equat e "self" with consciousne ss thus leads into the notion
that ideas are "forces"; and this fallacy is in turn suppo rted by saying
that the axon ca rries "impulses ." To find a wa y out of th is mess is by
no m eans easy.
We shall proceed by first examining the structure of the
alcoho lic's polarization. In the epistemologically unsound res-
olution, "I will fight the bottle," what is suppos edly lined up agains t
what?
Alcoho lic "Pride"
Alcohol ics are philosophers in that unive rsal sense in which all
human being s (and all mammals) are guided by highly abstract
principles of which they are either quite unconsciou s, or unaware
that the principle govern ing their perception and action is
philosophic . A common misnomer for such principles is
"feelings ."68
This misnomer arises naturally from the Anglo-Saxon
epistemological tenden cy to reify or attribute to the body all mental
pheno mena which are peripher al to consciou sness . And the
misnomer is, no doubt , suppor ted by the fact that the exercise and/or
frustration of these princip les is often accompanied by visceral and
other bodily sensat ions. I believe, howeve r, that Pascal was correct
67 " R. G. Collingwood, The Idea of Nature, Oxford, Ox-ford Universit y Press,
1945.68 " G. Bateson, "A Social Scientist Views the Emotions, " Expr ession of the
Emotions in Man, P. Knapp, ed., International Universit y Press, 1963 .
326

when he said, "The heart has its reasons which the reason does not
at all perceive.
But the reader must not expect the alcohol ic to present a
consistent picture. When the unde rlying epistemology is full of
error, derivations from it are inevitably either self-con tradictory or
extremely restricted in scope . A consis tent corpu s of theorems
cannot be derived from an incons isten t body of axioms. In such
cases, the attempt to be consisten t leads either to the great
proliferation of complexit y characteristic of psychoanal ytic theory
and Christian theolog y or to the extremely narrow view
characteristic of contemporary behavio rism.
I shall therefor e proceed to examine the "pride" which is
characteristic of alcohol ics to show that this princ iple of their
behavior is derived from the strange dualistic epistemology
characteristic of Occiden tal civilization.
A conven ient way of describing such princip les as "pride ,"
"dependenc y," "fatalism," and s o forth, is to exa mine the princip le as
if it were a result of deutero-l earning69 and to ask what contexts of
learning m ight unde rstandab ly inculc ate this princip le.
(1) It is clear that the princip le of alcohol ic life which AA calls
"pride" is not cont extua lly structured around pas t achievement. They
do not use the word to mean pride in something accomplished. The
emphasis is not upon "I succeeded," but rather upon "I can. . . ." It is
an obsessiv e accep tance of a challenge , a repud iation of the
proposi tion " I cannot ."
(2)After the alcoholic has begun to suffer from—or be blamed
for—alcoholism, this principle of "pride" is mobilized behind the
proposition, "I can stay sober." But, noticeably, success in this
achievement destroys the "challenge." The alcoholic becomes
69 This use of formal contextual structure as a descriptive device does not
necess arily assume that the principle discuss ed is wholly or in part actuall y
learned in contexts havin g the appropriate formal structure. The principle could
have been geneticall y deter mined, and it might still follo w that the principle is best
described by the formal delin eation of the conte xts in which it is exemplified. It is
precisely this fitting of behavior to context that makes it difficult or impossible to
determine whether a principle of behavior was geneticall y deter mined or learned in
that context; see G. Bateson , "Social Planning and the Concept of Deutero-
Learnin g," Conference on Science, Philosophy and Religion, Second
Symposium, New York, Harper , 1942.
327

"cocksure," as AA says. He relaxes his determination, risks a
drink, and finds himself on a binge. We may say that the
contextual structure of sobriety changes with its achievement.
Sobriety, at this point, is no longer the appropriate contextual
setting for "pride." It is the risk of the drink that now is
challenging and calls out the fatal "I can….
(3)AA does its best to insist that this change in con-textual
structure shall never occur. They restructure the whole context by
asserting over and over again that "Once an alcoho lic, always an
alcoho lic." They try to have the alcoholic place alcoholism within
the self, much as a Jungian analyst tries to have the patient
discover his "psychological type" and to learn to live with the
strengths and weaknesses of that type. In contrast, the contextual
structure of alcoholic "pride" places the alcoholism outside the
self: "I can resist drinking."
(4)The challenge component of alcoholic "pride" is linked with
risk-taking. The principle might be put in words: "I can do
something where success is improbable and failure would be
disastrous." Clearly this principle will never serve to maintain
continued sobriety. As success begins to appear probable, the
alcoholic must challenge the risk of a drink. The element of "bad
luck" or "probability" of failure places failure beyond the limits of
the self. "If failure occurs, it is not mine." Alcoholic "pride"
progressively narrows the concept of "self," placing what happens
outside its scope.
(5)The principle of pride-in-risk is ultimately almost suicidal.
It is all very well to test once whether the universe is on your side,
but to do so again and again, with increasing stringency of proof,
is to set out on a project which can only prove that the universe
hates you. But, still and all, the AA narratives show repeatedly
that, at the very bottom of despair, pride sometimes prevents
suicide. The final quietus must not be delivered by the "self."70
70 See Bill's Story, Alcoholics Anon ymous, op . cit.
328

Pride and Sy mmet ry
The so-called pride of the alcoholi c alwa ys presumes a real or
fictitious "other," and its complete contex tual defin ition therefore
demands that we charac terize the real or imagined relationsh ip to
this "othe r." A first step in this task is to classify the relationsh ip as
either "symmetrical" or "complementar y."71 To do this is not entirely
simple when the "other" is a creation of the unconsciou s, but we
shall see that the indication s for such a classification are clear.
An explanator y digres sion is, howeve r, necessa ry. The primary
criterion is simple:
If, in a binary relationship , the behaviors of A and B are regarded
(by A and B) as similar and are linked so that more of the given
behavior by A stimulates more of it in B, and vice versa, then the
relationsh ip is "symmetrical" in regard to these behav iors.
If, conv ersely, the behav iors of A and B are dissimilar but
mutuall y fit together (as, for example, spectatorsh ip fits ex-
hibitionis m), and the behavio rs are linked so that more of A's
behavior stimulates more of B's fitting behavior, then the
relationsh ip is "complementar y" in regard to these b ehavio rs.
Common examples of simple symmetrical relationship are
armaments races, keep ing up with the Jones es, athletic emulation,
boxing matches, and the like. Common examples of complementar y
relationsh ip are dominance-sub mission, sadism-masochis m,
nurturan ce-depend ency, spectatorship -exhibitionism, and the like.
More complex consideration s arise when higher logical typing is
present . For example: A and B may compete in gift-giving , thus
superposing a larger symmetrica l frame upon primaril y
comple mentar y behav iors. Or, converse ly, a therapist might engage
in competition with a patient in some sort of play therap y, placing a
comple mentar y nurturan t frame around the primarily symmetrica l
transa ction s of the game.
Various sorts of "doubl e binds" are genera ted when A and B
perceiv e the premises of their relationship in differen t terms—A
71 G. Bateson, Naven, Cambridge, Cambridge University Press, 1936.
329

may regard B's behavio r as competitive when B though t he was
helping A. And so on.
With these complexit ies we are not here conce rned, be-cause the
imagina ry "othe r" or coun terpa rt in the "pride" of the alcoholi c does
not, I believe , play the complex games which are charac teristic of
the "voic es" o f sch izophr enics .
Both comple mentar y and symmetrical relationships are liable to
progr essive change s of the sort which I have called
"schismogenesis ."72 Symmetrical struggle s and armaments races
may, in the curren t phrase, "escala te"; and the normal pattern of
succo ring-dep endenc y between parent and child may beco me
monstrous . These potentially patholog ical develop ments are due to
unda mped or uncor rected positive feedback in the system, and may
—as stated—o ccur in either comple mentary or symmetrical
systems. However , in mixed systems schismogenesis is neces sarily
reduc ed. The armaments race between two nations will be slowed
down by acceptan ce of comple mentary themes such as dominance ,
de-pend ency, admiration, and so forth, between them. It will be
speeded up by the repud iation of these themes.
This antithetical relationsh ip between comple mentar y and
symmetrica l themes is, no doub t, due to the fact that each is the
logical oppos ite of the other. In a merely symmetrica l armaments
race, nation A is motivated to greater efforts by its estimate of the
greater strength of B. When it estimates that B is weake r, nation A
will relax its efforts. But the exact opposi te will happen if A's
structu ring of the relationship is complementary. Observ ing that B is
weaker than they, A will go ahead wi th hop es of conques t.73
This antithes is between complementar y and symmetrical patterns
may be more than simply logica l. Notably, in psychoanal ytic
theory,74 the patterns which are called "libidinal" and which are
modalities of the erogenous zone s are all complementary. Intrus ion,
72 Ibid.73 G. Bateson, "The Pattern of an Armaments Race–Part I: An Anthropological
Approach," Bulletin of Atomic
Scientists, 1946, 2(5): 10–11: also L. F. Rich ardson, "Generalized Forei gn
Politics, " British Journ al of Psychology, Monograph Supple ments, 1939.74 E. H. Erikson, "Configurations in Play—Clinical
Notes," Psychoanalytic Quarterly, 1937, 6: 139–214.
330

inclus ion, exclusion , reception, retention, and the like—all of these
are classed as "libidinal." Whereas rivalry, competition, and the
like fall under the rubric of "ego" and "defense."
It is also possible that the two antithetical codes—s ymmetrica l
and complementar y—may be physiologica lly represented by
contrasting states of the central nervous system. The progr essive
changes of schismogenesi s may reach climactic discont inuities and
sudd en reversa ls. Symmetrica l rage may suddenl y turn to grief; the
retreating animal with tail between his legs may suddenl y "turn at
bay" in a despe rate battle of symmetry to the death. The bully may
sudd enly beco me the coward when he is challeng ed, and the wolf
who is beaten in a symmetrical conflict may sudd enly give "sur-
render" signals which p revent further attack.
The last example is o f spec ial interest. If the struggle b etween th e
wolve s is symmetrical— that is, if wolf A is stimulated to more
aggressive behavior by the aggressive behavior of B—th en if B
sudd enly exhib its what we may call "nega tive aggress ion," A will
not be able to continue to fight unles s he can quickly switch over to
that complementary state of mind in which B's weakness would be a
stimulus for his aggression . Within the hypothes is of symmetrical
and complemetar y modes, it becomes unnece ssary to postulate a
specifically "inhib itory" effect for the surrender signal.
Human beings who possess language can apply the label
"aggression" to all attempts to damage the other, regardless of
whether the attempt is prompted by the other's strength or
weakness; but at the prelinguistic mammalian level these two sorts
of "aggression" must appear totally different. We are told that from
the lion's point of view, an "attack" on a zebra is totally different
from an "attack" on another lion.75
Enough has now been said so that the question can be posed: Is
alcoholic pride contextually structured in symmetrical or
complementary form?
First, there is a very strong tendency toward symmetry in the
normal drinking habits of Occidental culture. Quite apart from
addictive alcoholism, two men drinking together are impelled by
convention to match each other, drink for drink. At this stage, the
75 13 K. Z. Lorenz, On Aggression, New York, Harcourt, Brace & World, 1966.
331

"other" is still real and the symmetry, or rivalry, between the pair is
friendly.
As the alcohol ic beco mes addicted and tries to resist drinking, he
begins to find it difficult to resist the social contex t in which he
should match his friends in their drinking . The AA says, "Heaven
knows, we have tried hard enough and long enough to drink like
other peop le!"
As things get worse, the alcoholi c is likely to beco me a solitary
drink er and to exhibit the whole spectrum of respon se to challeng e.
His wife and friends begin to sugges t that his drinking is a
weakness , and he may respond , with symmetry, both by resenting
them and by asserting his streng th to resist the bottle. But, as is
characteristic of symmetrical responses , a brief period of success ful
struggle weakens his motivation and he falls off the wagon. Sym-
metrical effort requires continual opposi tion from the opponent.
Gradual ly the focus of the battle changes, and the alcoholic finds
himself committed to a new and more deadl y type of symmetrical
conflict. He must now prove that the bottle canno t kill him. His
"head is blood y but unbowed ." He is still the "captain of his soul"—
for what i t's worth .
Meanwhi le, his re lationship s with wife and boss and friends have
been deteriorating. He neve r did like the complementary status of
his boss as an authori ty; and now as he deteriorates his wife is more
and more forced to take a comple mentar y role. She may try to exert
autho rity, or she beco mes protective, or she shows forbearan ce, but
all those provoke either rage or shame. His symmetrical "pride" can
tolerate no complementary role.
In sum, the relationsh ip between the alcoho lic and his real or
fictitious "other" is clearly symmetrica l and clearly schismogenic . It
escalates. We shall see that the religiou s conve rsion of the alcoho lic
when saved by AA can be de-sc ribed as a dramatic shift from this
symmetrica l habit, or epistemology, to an almost purely
complementary view of his relationsh ip to others and to the univer se
or God.
332

Pride or Invert ed Proo f?
Alcoho lics may appe ar to be st iff-necked , but they are not stupid .
The part of the mind in which their policy is decided certainly lies
too deep for the word "stupid ity" to be applicable . These levels of
the mind are prelinguis tic and the computation which goes on there
is cod ed in primary pr ocess.
Both in dream and in mammalian interaction, the only way to
achieve a proposi tion which contains its own nega tion ("I will not
bite you," or "I am not afraid of him") is by an elabora te imagining
or acting out of the proposi tion to be negated, leading to a reduc tio
ad absur dum. "I will not bite you" is achieved between two
mammals by an expe rimental combat which is a "not combat,"
sometimes called "play." It is for this reason that "agon istic"
behavior commonly evolves into friend ly greeting.76
In this sense , the so-called pride of the alcoholi c is in some
degree ironic . It is a determined effort to test some-thing like "self-
control" with an ulterior but unsta teable purpose of proving that
"self-cont rol" is ineffectu al and absurd. "It simply won't work." This
ultimate proposi tion, since it conta ins a simple negation, is not to be
expressed in primary process. Its final expre ssion is in an action—
the taking of a drink. The heroic battle with the bottle, that fictitious
"other," ends up in a "kiss and make friends ."
In favor of this hypothes is, there is the undoub ted fact that the
testing of self-contro l leads back into drinking. And, as I have
argued above, the whole epistemology of self-con trol which his
friend s urge upon the alcoho lic is monstrous . If this be so, then the
alcohol ic is right in rejecting it. He has achiev ed a reduc tio ad
absu rdum of the conventional epistemology.
But this description of achieving a reduc tio ad absu rdum verges
upon teleology. If the proposition "It won't work" can-not be
entertained within the coding of primary process, how then can the
computations of primary process direct the organism to try out
76 G. Bateson, "Metalog ue: What Is an Instinct ?," Aproaches to Animal
Communication, T. Sebeok, Hague, Mouton , 1969 .
333

those courses of action which will demonstrate that "It won't
work"?
Proble ms of this genera l type are frequ ent in psychiatry and can
perhap s only be resolved by a model in which, under certain
circumstance s, the organism's disco mfort activates a positive
feedba ck loop to increase the behav ior which preceded the
disco mfort. Such positive feedback would provide a veri fication tha t
it was really that particular behavio r which brough t about the
disco mfort, and might in-crease the discomfort to some threshold
level at whi ch chang e would beco me poss ible.
In psychothe rapy such a positive feedba ck loop is commonly
provid ed by the therapist who push es the patient in the direction of
his symptoms—a technique which has been called the "therapeu tic
double bind." An example of this techn ique is quoted later in this
essay, where the AA member challenges the alcohol ic to go and do
some "cont rolled drinking" i n order t hat h e may discover f or hi mself
that he has no con trol.
It is also usual that the symptoms and hallucin ation s of the
schizophren ic—l ike dreams—consti tute a corrective experienc e, so
that the whole s chizoph renic e pisode t akes on t he cha racte r of a s elf-
initiation . Barbara O'Brien's accoun t of her own psychosis77 is
perhap s the most striking example of this pheno menon, which has
been discus sed e lsewhere .78
It will be noted that the possib le existenc e of such a positive
feedba ck loop, which will cause a runawa y in the direction of
increasing discomfort up to some threshold (which might be on the
other side of death), is not included in conven tiona l theories of
learning . But a tendenc y to verify the unplea sant by seeking
repeated exper ience of it is a common human trait. It is perhaps
what F reud c alled th e "de ath instinct."
77 B. O'Brien, Operators and Things: The Inner Life of a Schizophrenic,
Cambridg e, Mass. , Arlington Books, 1958. 78 G. Bateson , ed., Perceval's Narrative, Stanford, Calif., Stanford University
Press, 1961, Introduction
334

The Drunken State
What has been said above abou t the treadmill of symmetrica l
pride is only one half of the picture. It is the picture of the state of
mind of the alcohol ic battling with the bottle. Clearly this state is
very unpleasan t and clearly it is also unrealistic. His "othe rs" are
either totally imaginar y or are gross distortions of persons on whom
he is dependen t and who m he may love. He has an al terna tive to th is
unco mfortabl e state—h e can get drunk . Or, "at least," have a drink.
With this complementar y surrende r, which the alcoho lic will
often see as an act of spite—a Barth ian dart in a symmetrica l
strugg le—his entire epistemolog y change s. His anxieties and
resent ments and panic vanish as if by magic. His self-contro l is
lessened , but his need to compare himself with others is reduced
even further. He feels the physiologica l warmth of alcoho l in his
veins and, in many cases , a corresponding psycholog ical warmth
toward o thers . He may be either maudlin or ang ry, but h e has at least
become again a part of the human sc ene.
Direct data bearing upon the thesis that the step from sobriety
into intox ication is also a step from symmetrical challenge into
comple mentari ty are scarce, and always confu sed both by the
distortion s of recall and by the complex toxicity of the alcohol. But
there is strong evidence from song and story to indicate that the step
is of this kind. In ritual, partaking of wine has alwa ys stood for the
social aggrega tion of persons united in religiou s "communion" or
secula r Gemütlichkeit. In a very literal sense, alcohol supposed ly
makes the individua l see himself as and act as a part of the group .
That is, it enab les complementar ity in the relationships which
surround him.
335

Hitting Bottom
AA attaches great importance to this pheno menon and regards
the alcoholi c who has not hit bottom as a poor prospec t for their
help. Converse ly, they are inclined to explain their failure by saying
that the individual who goes back to his alcoho lism has not yet "hit
bottom."
Certa inly many sorts of disaste r may cause an alcohol ic to hit
bottom. Various sorts of accidents , an attack of delirium tremens, a
patch of drunken time of which he has no memory, rejection by
wife, loss of job, hopeless diagnosis , and so on—an y of these may
have the required effect. AA says that "bottom" is different for
differen t men and so me may be dead before they reach it.79
It is possible, howeve r, that "botto m" is reach ed many times by
any given individual ; that "botto m" is a spell of panic which
provid es a favo rable moment for change , but not a moment at which
change is inevitable. Friends and relatives and even therap ists may
pull the alcohol ic out of his panic, either with drugs or reassur ance,
so that he "re-c overs" and goes back to his "pride" and alcoho lism—
only to hit a more disastrous "bottom" at some later time, when he
will again be ripe for a chang e. The attempt to change the alcoho lic
in a period between such moments of panic is unlikely to succeed .
The nature of the panic is made clear by the following
description o f a "test."
We do not like to pronounce any individua l as alcohol ic, but you
can quick ly diagno se yourself . Step over to the neares t barroom and
try some controlled drink ing. Try to drink and stop abrup tly. Try it
more than once. It will not take long for you to decide, if you are
honest with yourself about it. It may be worth a bad case of jitters if
you get a full knowl edge of your condi tion.80
We might compare the test quoted above to commanding a drive r
to brake sudden ly when traveling on a slippe ry road: he will
discov er fast that his contro l is limited. (The metaphor "skid row"
for the alcoholi c section of town i s not inappropr iate.)
79 Personal communication fro m a me mber .
336

The panic of the alcoholi c who has hit bottom is the panic of the
man who thought he had control over a vehic le but suddenl y finds
that the vehicle can run away with him. Sudden ly, pressur e on what
he knows is the brake seems to make the vehicle go faster. It is the
panic of discover ing that it (the system, self plus vehicle) is bigger
than he is.
In terms of the theory here presen ted, we may say that hitting
bottom exemplifi es systems theory at three levels :
(1)The alcoholic works on the discomforts of sobriety to a
threshold point at which he has bankrupted the epistemology of "self-control." He then gets drunk—because the "system" is bigger
than he is—and he may as well surrender to it.
(2)He works repeatedly at getting drunk until he proves that
there is a still larger system. He then encounters the panic of
"hitting bottom."
(3)If friends and therapists reassure him, he may achieve a
further unstable adjustment—becoming addicted to their help—
until he demonstrates that this system won't work, and "hits
bottom" again but at a lower level. In this, as in all cybernetic
systems, the sign (plus or minus) of the effect of any intrusion
upon the system depends upon timing.
(4)Lastly, the phenomenon of hitting bottom is complexly
related to the experience of double bind.81 Bill W. narrates that he
hit bottom when diagnosed as a hopeless alcoholic by Dr. William
D. Silkworth in 1939, and this event is regarded as the beginning
of AA history.82 Dr. Silkworth also "supplied us with the tools with
which to puncture the toughest alcoholic ego, those shattering
phrases by which he described our illness: the obsession of the mind
that compels us to drink and the allergy of the body that condemns us
to go mad or die."83 This is a double bind correctly founded upon
the alcoholic's dichotomous epistemology of mind versus body. He
is forced by these words back and back to the point at which only
80 Alcoholics Anon ymous, op. cit., p. 43.
81 Bateson, et al., "Toward a Theory of Schizophrenia ," Behavioral Science,
1956, 1: 251-64. 82 AA Comes of Age, op. cit., p. vii83 Ibid., p. 13. (Italics in the original)
337

an involuntary change in deep unconscious epistemology—a
spiritual experience—will make the lethal description irrelevant.
The Theo logy o f Alcohol ics Anonymous
Some outstanding points of the theology of AA are:
(1) There is a Power greater than the self. Cybernetic s would go
somewhat further and recogni ze that the "self" as ordinarily
unders tood is only a small part of a much larger trial-and-e rror
system which does the thinking , acting , and deciding . This system
includes all the informational path-ways which are relevant at any
given moment to any given decision . The "self" is a false reification
of an imprope rly de-limited part of this much larger field of
interlocking processes. C ybernet ics also recogn izes tha t two or m ore
person s —an y group of persons— may togeth er form such a
thinkingand-a cting system.
(2) This Power is felt to be personal and to be intimatel y linked
with each person . It is "God as you unde rstand him to be."
Cybernet ically speaking , "my" relation to any larger system
around me and including other things and persons will be different
from "your" relation to some similar system around you. The
relation "part of" must necessar ily and logica lly al-ways be
complementary but the meaning of the phrase "part of" will be
differen t for every person .84 This difference will be especially
important in systems containing more than one person. The system
or "power" must necessarily appear different from where each
person sits. Moreover , it is expe ct-ab le that such systems, when they
encount er each other, will recogniz e each other as systems in this
sense . The "beau ty" of the woods through which I walk is my
recogn ition both of the individua l trees and of the total ecolog y of
the woods as systems. A similar esthetic recogni tion is still more
striking when I talk with a nothe r person.
84 This diversity in styles of integration could account for the fact that some
persons become alcoholic while others do not .
338

(3) A favor able relationship with this Power is discove red
through " hitting bottom" and " surrend er."
(4) By resisting this Powe r, men and espec ially alcohol ics bring
disaster upon themselves . The materialistic philosoph y which sees
"man" as pitted against his environment is rapidly breaking down as
technolog ical man becomes more and more able to oppose the
largest systems. Ever y battle that he wins brings a threat of disaster.
The unit of surviva l—ei ther in ethics or in evolu tion— is not the
organis m or the species but the largest system or "power" within
which the creature lives. If the creature destroys its environ ment, it
destroys it-self.
(5) But—and this is important— the Power does not re-ward and
punish. It does not have "power" in that sense . In the biblical phra se,
"All things work together for good to them that love God." And,
conv ersely, to them that do not. The idea of power in the sense of
unilateral control is foreign to AA. Their organiza tion is strictly
"democrati c" (their word ), and even their deity is still bound by
what we might call a systemic determinism. The same limitation
applies both to the relationship between the AA sponsor and the
drunk whom he hopes to help and to the relationship between AA
central office and every local group.
(6) The first two " steps " of Alcohol ics Anon ymous taken together
identify the add iction a s a manifestation o f this Power .
(7) The healthy relation between each person and this Power is
comple mentar y. It is in precise contrast to the "pride" of the
alcohol ic, which is predica ted upon a symmetrica l relationship to an
imagined "other." The schis mogenes is is always more powerful than
the pa rticipan ts in it.
(8) The qual ity and content of each person's relation to the Power
is indicated or reflected in the social structure of AA. The secula r
aspect of this system—its governan ce—is delineated in "Twelv e
Traditions"85 which supple ment the "Twelve Steps ," the latter
developing man's relationsh ip to the Power . The two documents
overlap in the Twelfth Step, which enjoins aid to other alcoho lics as
a necessary spiritual exercise without which the member is likely to
relaps e. The total system is a Durkhei mian religion in the sense that
85 AA Comes of Age, op. cit. 24 Ibid., p. 288. 25 Ibid., pp. 286-94.
339

the relationship between man and his community parallels the
relationship between man and God. "AA is a power greater than any
of us."86
In sum, the relationship of each individua l to the "Power" is best
defined i n the words is part of."
(9) Anon ymity. It must be unders tood that anon ymity means
much more in AA thinking and theology than the mere protection of
members from exposure and shame. With increasing fame and
succe ss of the organization as a whole , it has beco me a temptation
for members to use the fact of their membership as a positive asset
in publi c relations , politics, education, and many other fields. Bill
W., the co-founde r of the organization, was himself caugh t by this
temptation in early days and has discussed the matter in a published
article.87 He sees first that any grabb ing of the spotlight must be a
person al and spiritual danger to the member, who cannot affort such
self-seeking; and beyond this that it would be fatal for the
organization as a whole to beco me involved in politics, religious
controvers y, and social reform. He states clearly that the errors of
the alcohol ic are the same as the "forces which are today ripping the
world apa rt at its sea ms," but that it is no t the bus iness o f AA to save
the world . Their single purpo se is "to carry the AA message to the
sick alcoho lic who wants it."88 He conc ludes that anon ymity is "the
greatest symbol of self-sacr ifice that we know ." Elsewhe re the
twelfth of the "Twelve Traditions" states that "anon ymity is the
spiritual foundat ion of our traditions , ever reminding us to place
principles be fore person alities."
To this we may add that anon ymity is also a profound statement
of the systemic relation, part-to-whole. Some systems theoris ts
would go even further, because a major temptation for systems
theory lies in the reification of theoretical concepts . Anatol Holt says
he wants a bumper sticker which would (paradoxica lly) say, "Stamp
out nouns."89
86 Ibid, p. 288 .87 Ibid, pp.286-29488 Ibid.89 M. C. Bateson , ed., Our Own Metaphor, Wenner-Gren Foundation,
Conference on the Effects of Conscious Purpose on Human Adaptation, 1968; New
York, Knopf, in press.
340

(10) Prayer. The AA use of prayer similar ly affirms the
comple mentari ty of part-whole relationship by the very simple
techniqu e of asking for that relationship . They ask for those-
personal charac teristics, such as humilit y, which are in fact-
exercised in the very act of prayer. If the act of prayer be sincere
(which is not so easy), God cannot but grant the reques t. And this is
peculiarly true of "God, as you unde rstand h im." This self-affirming
tautology, which contains its own beau ty, is precisely the balm
requir ed after the angu ish of the doub le binds which went with
hitting bo ttom.
Somewhat more complex is the famous "Serenity Prayer": "God
grant us the serenity to accept the things we cannot change, courag e
to chang e the things we can, and wisdo m to know the difference ."90
If double binds cause anguish and despai r and destroy personal
epistemologica l premises at some deep level, then it follows ,
conv ersely, that for the healing of these wounds and the growth of a
new epistemology, some conve rse of the doub le bind will be
appropri ate. The double bind leads The Seren ity Prayer explicitly
frees the worshipp er from these m addening bonds .
to the con clusion of despa ir, "There are no alternatives."
In this connect ion it is worth mention ing that the great
schizophr enic, John Perceval, observed a change in his "voice s." In
the beginn ing of his psychosis they bullied him with "cont radic tory
commands" (or as I would say, doubl e binds) , but later he began to
recover when they offered him choice of clearly defined
alternatives.91
(11) In one characteristic, AA differs profoundl y from such
natural mental systems as the family or the redwood forest. It has a
single purpose—"to carry the AA message to the sick alcohol ic who
want s it"— and the organiza tion is dedicated to the maximizat ion of
that purpose. In this respect, AA is no more sophis ticated than
Gener al Motors or an Occiden tal nation. But biologica l systems,
other than those premised upon Occidenta l ideas (and especially
money), are multipurpo sed. There is no single variable in the red-
90 This was not originally an AA document and its authorship is unknown. Small
variations in the text occur. I have quoted the form which I personally prefer from AA
Comes of Age, op. cit., p. 196.91 Bateson, Perceval . . . , op. cit.
341

wood forest of which we can say that the whol e system is oriented
to maximizing that variable and all other variables are subsid iary to
it; and, indeed , the redwood fores t works toward optima, not
maxima. Its needs are satiable, and too much of anything is toxic.
There is, howeve r, this: that the single purpose of AA is directed
outward and is aimed at a nonco mpetitve relationship to the larger
world . The variable to be maximized is a complementarity and is of
the nature of "servic e" rather than do minance.
The Epist emologic al Status of Compl ementa ry and
Symm etrical Prem ises
It was noted above that in human interaction, symmetry and
complementarity may be complexl y combined. It is therefo re
reasonable to ask how it is possible to regard these themes as so
fundamental that they shall be called "epistemological," even in a
natural history study of cultural and interpersonal premises.
The answer seems to hang upon what is meant by "fundamental"
in such a study of m an's natural history; and the word see ms to carry
two so rts of meaning.
First, I call more fundamental those premises which are the more
deepl y embedded in the mind, which are the more "hard
progr ammed" and the less suscep tible to change . In this sense, the
symmetrica l pride or hubris of the alcoholi c is fund amental .
Second, I shall call more funda mental those premises of mind
which refer to the larger rather than the smaller systems or gestalten
of the univer se. The proposi tion "Grass is green" is le ss funda mental
than the propos ition "Colo r differences make a differenc e."
But, if we ask abou t what happens when premises are changed , it
beco mes clear that these two defin itions of the "fund amental"
overlap to a very grea t extent. If a man achieves or suffers change in
premises which are deep ly embedded in his mind, h e will surely find
that the results of that change will ramify throughout his whole
unive rse. Such c hanges we may well call "epis temological."
The question then remains regard ing what is epistemological ly
"right" and what is epistemological ly "wrong." Is the change from
alcoho lic symmetrica l "prid e" to the AA species of complementar ity
342

a correction of his epistemology? And is comple mentar ity always
someho w be tter than s ymmetry?
For the AA member, it may well be true that complementarity is
always to be preferred to symmetry and that even the trivial rivalry
of a game of tennis or chess may be dangerous . The superficial
episode may touch off the deeply embedded symmetrical premise.
But this does not mean that tennis and chess propose
epistemologica l error for everybody.
The ethical and philosoph ic probl em really concerns only the
wide st universe and the deepe st psycholog ical levels. If we deeply
and even unconsc iousl y believe that our relation to the largest
system which concern s us— the "Power greater than self"— is
symmetrical and e mulat ive, then we are in error.
Limitations of the Hypo thesi s
Finally, the above analysis is subject to the following limitat ions
and implications:
(1)It is not asserted that all alcohol ics operate according to the
logic which is here outlined. It is very possible that other types of
alcohol ics exist and almost certain that alcoholi c addiction in other
cultures wi ll follow other lines.
(2)It is not asserted th at the way of Alcohol ics Anon ymous is the
only way to live correctly or that their theolog y is the only correct
derivation from the epistemology of cybernetics and systems
theory.
(3)It is not asserted that all transac tions between human beings
ough t to be comple mentar y, though it is clear that the relation
between the individual and the larger system of which he is a part
must necessa rily be so. Relation s between persons will (I hope)
always be complex.
(4)It is, howeve r, asserted that the nonalcoho lic world has many
lessons which it might learn from the epistemolog y of systems
theory and from the ways of AA. If we cont inue to operate in terms
of a Cartesian dualism of mind versus matter, we shall probably
also continu e to see the world in terms of God versus man; elite
343

versus peop le; chosen race versus others ; nation versus nation; and
man versus environment. It is doub tful wheth er a species having
both an advanced techno logy and this strange way of looking at its
world can endur e.
344

Comment on Par t III
In the essays collected in Part III, I speak of an action or
utterance as occurr ing "in" a context, and this convent ional way of
talking sugges ts that the particula r action is a "dependent" variab le,
while the contex t is the "independen t" or determining variable. But
this view o f how an action is related to its con text is likely to distract
the reader—as it has distracted me—from perceiving the ecology of
the ideas which togeth er constitute the small subsystem which I call "context."
This heuristic error—copi ed like so many others from the ways
of though t of the physicist and chemist—requ ires correction.
It is importan t to see the particula r utteranc e or action as part of
the ecologi cal subsystem called contex t and not as the produc t or
effect of what remains of the contex t after the piece which we want
to exp lain has been c ut out from it.
The mistake in question is the same formal error as that
mentioned in the comment on Part II where I discuss the evolu tion
of the horse. We should not think of this process just as a set of
changes in the animal's adap tation to life on the grassy plains but .as
a constancy in the relationship between animals and
environment. It is the ecolog y which s urvives and s lowly evolves . In
this evolution, the relata— the animals and the grass—und ergo
changes which are indeed adaptive from moment to moment. But if
the process of adapta tion were the whole story, there could be no
systemic patholog y. Trouble arises precisely because the "logic" of
adaptation is a different "logic" from that of the surviva l and
evolution of the ecologica l system.
In Warren Brode y's phrase, the "time-grain" of the adapta tion is
different from that of the ecolog y.
"Surv ival" means that certain descrip tive statements abou t some
living system continue to be true through some period of time; and,
conv ersely, "evolu tion" refers to changes in the truth of certain
descrip tive statements abou t some living system. The trick is to
define which statements abou t which systems remain true or
unde rgo ch ange.
345

The paradox es (and the patho logie s) of systemic process arise
precisely because the constancy and survival of some larger system
is maintained by change s in the cons tituent subsystems.
The relative constancy—the surviva l—of the relationship
between animals and grass is maintained by changes in both relata.
But any adap tive chang e in either of the relata, if unco rrected by
some chang e in the other, will alwa ys jeopa rdize the relationsh ip
between the m. These ar guments propos e a new concep tual fr ame for
the "doub le bind" hypothesi s, a new concep tual frame for thinking
about "schizophren ia," and a new way of looking at contex t and
levels of learning.
In a word , schizophreni a, deutero-learning , and the doub le bind
cease to be matters of individua l psycholog y and be-come part of
the ecolog y of ideas in systems or "minds" whose bound aries no
longe r coincide with the skins of the participan t individua ls.
346

Part IV: Biology and Evo lution

On Empty-Head edness Among Biologis ts
and
State Boards o f Educ ation*
My father, the gene ticist William Bateson, used to read us
passag es of the Bible at breakfa st—l est we grow up to be empty-
headed atheists; and so I find it natural to wonder what broadening
of the mind may come from the strange anti-evo lution ary ruling of
the State Board of Educat ion in Cal iforn ia.1
Evolut ion has long been badly taught . In particular, students—
and even profess ional biologis ts—acqu ire theor ies of evolut ion
withou t any deep understanding of what problem these theories
attempt to solve. They learn but little of the evolut ion of
evolu tiona ry theory.
The extraordina ry achievement of the writers of the first chapter
of Gene sis was their perception of the problem: Where does order
come from? They observed that the land and the water were, in
fact, separate and that species were separa te; they saw that such
separation and sorting in the univer se presen ted a funda mental
probl em. In modern terms, we may say that this is the probl em
implicit in the Second Law of Thermodynamics: If rando m events
lead to things getting mixed up, by what nonrando m even ts did
things co me to be sorted? And what is a "rando m" event?
This problem has been centra l to biolog y and to many other
sciences for the last 5000 years, and the problem is not trivial.
With what Word shou ld we designa te the princ iple of order
which s eems to be immanent i n the un ivers e?
The Californ ia ruling sugg ests that student s be told of other
attempts to solve this ancient problem. I myself collected one of
these among the Stone Age head-hun ters of the Iatmul tribe in New
Guinea . They, too, note that the land and the wate r are separa te even
in their swampy region. They say that in the beginning there was a
* This item in BioScience, Vol. 20, 1970, is reproduced by permission from
that journal .1 See "California's Anti-Evolution Ruling," BioScience, March 1, 1970 .
348

vast crocod ile, Kavwok mali, who padd led with his front legs and
padd led with his back legs, and thereby kept the mud in suspens ion.
The culture hero, Keve mbuangga, speared the crocod ile, who then
ceased to paddle , causing the mud and the water to separ ate. The
result was dry land upon which Keve mbuangga stamped his foot in
triumph. We might s ay he verified that "it was good ."
Our students might have their minds broadened somewhat if they
would look at other theori es of evolution and consid er how a man's
spirit must take a different shape if he believes that all sorting in the
universe is due to an external agent, or if, like the Iatmul and
modern scientists, he sees that the poten tiality for order and pattern
is immanent throughout this world .
And then the student may be forced by the new system to look at
the "Grea t Chain of Being ," with Supr eme Mind at the top and the
protozoa at the bottom. He will see how Mind was invoked as an
explanatory principle all through the Middle Ages and how Mind
later became the problem. Mind became that which needed
explanation when Lamarck showed that the Great Chain of Being
shou ld be inverted to give an evolu tiona ry sequence from the
protozoa upward. The proble m then was to expla in Mind in terms of
what could be known of this sequen ce.
And when the student reaches the mid-ninete enth century, he
might be given as a textbook Philip Henr y Gosse's Creation
(Omphalos): An Attempt to Untie the Geological Knot. He will learn
from this extraordina ry book things about the structur e of animals
and plants which are today scarce ly mentioned in many courses of
biolog y; notab ly, that all animals and plants show a time structure,
of which the rings of growth in trees are an elementary example and
the cycles of life history, a more complex one. Every plant and
animal is constructed upon the premise of its cyclic nature.
After all, there can be no harm in Goss e, who was a devout
fundamental ist—a Plymouth Brother— as well as a distingu ished
marine biologis t. His book was published in 1857, two years before
the Orig in a f Speci es. He wrote it t o show that th e facts of the foss il
record as well as those of biolog ical homolog y could be made to fit
with the principles of funda mentalism. It was to him inconce ivable
that God could have created a world in which Adam had no navel ;
the trees in the Garden of Eden, no rings of growth; and the rocks,
349

no strata. Therefore, God must have created the world as though it
had a past.
It will do the student no harm to wrestle with the paradox es of
Gosse's "Law of Proch ronis m"; if he listens care-fu lly to Gosse 's
groping genera lizations about the biological world, he will hear an
early version of the "steady state" hypothes is.
Of course , everybody knows that biologic al pheno mena are
cyclic-f rom egg, to hen, to egg, to hen, etc. But not all biologis ts
have examined the implica tions of this cyclic characteristic for
evolu tiona ry and ecolog ical theor y. Gosse's view of the biolog ical
world might b roaden their minds.
It is silly and vulgar to approa ch the rich spectrum of
evolu tiona ry thought with quest ions only abou t who was right and
who was wrong . We might as well assert that the amphibia and
reptiles were "wrong" and the mammals and birds "right" in their
solutions to the problems of how to live.
By fighting the funda mentalists, we are led into an empty-
headedne ss analogou s to th eirs. The truth of t he m atter i s that "Other
men have labou red and ye are entered into their labours" (John
1:38), and this text is not only a reminder of the need for humility, it
is also an epitome of the vast evolut ionar y process into which we
organisms are willy-nilly entered.
350

The R ole of Somatic Chang e in
Evolution*
All theories of biolog ical evolu tion depend upon at least three
sorts of change: (a) change of geno type, either by mutation or by
redistribu tion of genes; (b) somatic change under pressure of
environment; and (c) changes in environmental condi tions . The
proble m for the evolu tionist is to build a theory combining these
types of chang e into an ongoing process which , under natural
selection, will account for the pheno mena of adaptation and
phylogen y.
Certain conven tiona l premises may be selected to gove rn su ch
theory building:
(a) The theory shall not depend upon Lamarckian inheritance.
August Weismann's argument for this premise still stands . There is
no reason to believe that either somatic chang e or change s in
environment can, in principle, call (by physiologi cal
communica tion) for approp riate genot ypic change . Indeed, the little
that we know abou t communication within the multicellular2
individu al indicates that such communica tion from soma to gene
script is likely to be rare and unlikely to be adap tive in effect.
However , it is appropriate to attempt to spell out in this essay what
this premise implies:
Whenever some characteristic of an organism is modifiab le
unde r measurab le envi ronmental impact or under measurab le impact
of internal physiolog y, it is possible to write an equa tion in which
the value of the charac teristic in question is expres sed as some
function of the value of the impacting circumstance . "Human skin
color is some function of exposure to sunlight," "respi ration rate is
some function of atmospher ic pressure ," etc. Such equa tions are
* This essay appeared in the journal Evolution, Vol 17, 1963, and is reprinted
with the editor 's permission.2 The problems of bacterial genetics are here deliberatel y excluded.
351

const ructed to be true for a variety of particula r observations , and
necessarily contain subsidiary proposi tions which are stable (i.e.,
continue to be true) over a wide range of values of impacting
circumstance and somatic characte ristic. These subsidiary
propos itions are of different logical type from the origin al
observ ation s in the labor atory and are, in fact, descriptive not of the
data but of our equation s. They are statements abou t the form of the
particular equa tion and about the value s of the parameters
mentioned within it.
It would be simple, at this point, to draw the line between
genotype and phenotype by saying that the forms and parameters of
such equations are provided by genes, while the impacts of
environment, etc. determine the actual event within this frame.
This would amount to saying, e.g., that the ability to tan is
genotypically determined, while the amount of tanning in a
particular case depends upon exposure to sun-light.
In terms of this oversimplified approach to the overlapping
roles of genotype and environment, the proposition excluding
Lamarckian inheritance would read somewhat as follows: In the
attempt to explain evolutionary process, there shall be no
assumption that the achievement of a particular value of some
variable under particular circumstances will affect, in the gametes
produced by that individual, the form or parameters of the
functional equation governing the relationship between that
variable and its environmental circumstances.
Such a view is overs implified, and paren theses must be added to
deal with more complex and extreme cases. First, it is importan t to
recogn ize that the organism, considered as a communication al
system, may itself operate at multiple levels of logical typing; i.e.,
that there will be instances in which what were above called
"para meters" are subject to change . The indiv idual organism might
as a result of "training" change its ability to develop a tan unde r
sunlight. And this type of change is certainly of very great
importance in the field of animal behav ior, where "learn ing to learn"
can never be ignored .
Second, the oversimplified view must be elabo rated to cove r
negat ive effects. An environ mental circumstance may have such
352

impact upon an organism unable to adap t to it, that the individual in
question will in fact produce no gametes.
Third, it is expec table that some of the parameters in one
equation may be subjec t to change unde r impact from some
environmental or physiologi c circumstance other than the
circumstance m entioned in that equa tion.
Be all that as it may, both Weismann's objec tion to Lamarckian
theory and my own attempt to spell the matter out share a certain
parsimony: an assumption that the princip les which order
pheno mena shall not themselves be suppos ed changed by those
pheno mena which they order. William of Occa m's razor might be
reformulated: in any explanation, logical types shall not be
multipl ied b eyond necessity.
(b) Somat ic change is absolut ely neces sary for survival. Any
change of environment which requir es adapt ive chang e in the
species will be lethal unless, by somatic change , the organis ms (or
some of them) are able to weather out a period of unpredictable
duration, until either approp riate genot ypic change occurs (whe ther
by mutation or by redistribution of genes already available in the
popu lation), or becaus e the environment returns to the previous
normal. The premise is truistical, regard less of the magnitude of the
time span invo lved.
(c) Somat ic change i s also ne cessary t o cope wi th any changes o f
geno type which migh t aid the organism in its external strugg le with
the environment. The individual organism is a complex
organization of interdependent parts. A mutational or other
genotypic change in any one of these (however externally valuable
in terms of survival) is certain to require change in many others—
which changes will probably not be specified or implicit in the
single mutational change of the genes. A hypothetical pregiraffe,
which had the luck to carry a mutant gene "long neck," would have
to adjust to this change by complex modifications of the heart and
circulatory system. These collateral adjustments would have to be
achieved at the somatic level. Only those pregiraffes which are
(genotypically) capable of these somatic modifications would
survive.
(d) In this essay, it is assumed that the corpus of genotypic
message s is preponderant ly digital in nature. In contrast, the soma is
353

seen as a working system in which the genotypic recipes are tried
out. Should it transp ire that the geno typic corpus is also in some
degre e analog ic—a working model of the soma—pre mise c (above )
would be negated to that degree . It would then be conceivab le .that
the mutant gene "long neck" might modify the message of those
genes which affect the develop ment of the heart. It is, of course,
known that genes may have pleiotropic effect, but these pheno mena
are relevan t in the present connection only if it can be shown, e.g.,
that the effect of gene A upon the phenot ype and its effect upon the
phenot ypic expression of gene B are mutually appropr iate in the
overa ll integra tion and ad aptation o f the organis m.
These considera tions lead to a classifying of both genotypic and
environmental changes in terms of the price which they exact of the
flexibi lity of the somatic system. A lethal change in either
environment or geno type is simply one which demands somatic
modification s which the organism canno t achieve .
But the somatic price of a given change must depend, not
absolu tely upon the change in question , but upon the range of
somatic flexib ility available to the organism at the given time. This
range , in turn, will depend upon how much of the organism's
somatic flexibility is already being used up in adjusting to other
mutations or environmental changes . We face an economics of
flexibi lity which , like any other econo mics, will become
determinative for the course of evolu tion if and only if the organis m
is opera ting close to the limits set by this econo mics.
Howeve r, this econo mics of somatic flexib ility will differ in one
important respect from the more familia r econo mics of money or
available ener gy. In these latter, each new expend iture can simply be
added to the preceding expend itures and the econo mics becomes
coercive when the additive total appro aches the limit of the budge t.
In contr ast, the combined effect of multiple changes , each of which
exacts a price in the soma, will be multiplicative. This point may
be stated as follows: Let S be the finite set of all possible living
states of the organism. Within S, let s1 be the smalle r set of all states
compatib le with a given mutation (ml), and let s2 be the set of states
compatib le with a second mutation (m2). It follows that the two
mutations in combination will limit the organism to the logical
354

product of s1 and s2, i.e., to that usually smaller subset of states
which is composed only of members common to both s1 and s2. In
this way each success ive mutation (or other genotypic change) will
fractiona te the possibilities for the somatic adjustment of the
organis m. And, should the one mutation require some somatic
change, the exact opposi te of a change required by the other, the
possibilities for somatic adjustment may immediately be reduced to
zero.
The same argument must surely apply to multiple environmental
changes which demand somatic adjus tments; and this will be true
even of those changes in environment which might seem to benefit
the organis m. An improve ment in diet, for example, will exclude
from the organis m's range of somatic adjustments those patterns of
growth which we would call "stunted" and which might be required
to meet some other exigen cy of the environment.
From these considera tions it follows that if evolu tion proceeded
in accordance with conven tiona l theory, its proces s would be
blocked. The finite nature of somatic change indicates that no
ongo ing process of evolut ion can result only from success ive
externa lly adap tive genotypic changes since these must, in
combination , beco me lethal, demanding combinations of internal
somatic adjus tments o f which the soma is incapab le.
We turn therefo re to a considera tion of other classes of genot ypic
change. What is requi red to give a balanced "theor y of evolution is
the occur rence of genot ypic changes which shall increase the
available r ange of so matic fl exibility. When the i nternal or ganiza tion
of the organisms of a species has been limited by environmental or
mutationa l pressure to some narrow subse t of the total range of
living states, further evolut ionar y progr ess will require some sort of
geno typic change which wi ll compensate for this limitat ion.
We note first that while the results of genot ypic chang e are
irreversible wi thin th e life of th e indiv idual o rganism, the oppos ite is
usually true of change s which are achieved at the somatic level.
When the latter are produced in respon se to special environmental
cond itions, a return of the environment to the previou s norm is
usually followed by a diminution or loss of the chara cteristic. (We
may reasonab ly expec t that the same would be true of those somatic
adjustments which must accompany an extern ally adap tive mutation
355

but, of course, it is impossib le in this case to remove from the
individua l the impact o f the mutationa l chang e.)
A further point regard ing these reversible somatic changes is of
special interest. Among higher organisms it is not unusual to find
that there is what we may call a "defense in depth" agains t
environmental demands. If a man is moved from sea level to 10,000
feet, he may begin to pant and his heart may race. But these first
changes are swif tly reversible: if he descends the same day, they will
disappear immediate ly. If, howeve r, he remains at the high altitude,
a second line of defens e appear s. He will beco me slowl y acclimated
as a result of complex physiologi cal changes. His heart will cease to
race, and he will no longe r pant unless he unde rtakes some speci al
exertion. If now he returns to sea level, the characteristics of the
second line of defense will disappea r rather slowly and he may even
exper ience some discomfort.
From the point of view of an econo mics of somatic flexibility,
the first effect of high altitude is to reduce the organism to a limited
set of states (si) chara cterized by the racing of the heart and the
panting. The man can still survive , but only as a comparatively
inflexib le creature. The later acclimation has precisely this value : it
corrects for the loss of flexibility. After the man is acclimated he can
use his panting mechanis ms to adjust to other emergenc ies which
might otherwi se be lethal.
A similar "defen se in dep th" is clearly recogn izable in the field of
behavio r. When we encoun ter a new probl em for the first time, we
deal with it e ither by trial and error or poss ibly by insight. Later, and
more or less gradu ally, we form the "habit" of acting in the way
which e arlier expe rience rewarded. To continue to use insigh t or trial
and error upon this class of proble m would be wasteful. These
mechanisms can now be saved fo r other proble ms.3
Both in acclimation and in habit formation the econo my of
flexibi lity is achieved by substituting a deeper and more endu ring
change for a more superficial and more reversible one. In the terms
used above in discussing the anti-Lamarckian premise, a change has
occur red in the parameters of the functional equat ion linking rate of
3 Bateson, "Minimal Require ments for a Theor y of Schizophrenia, " A.M.A.
Archives of General Psychiatry, 1960, 2: 447.
356

respiration to externa l atmospheri c pressure . Here it seems that the
organis m is behaving as we may expec t any ultrastable system to
behave. Ashb y4 has shown that it is a genera l formal characteristic
of such systems that those circuits con-trolling the more rapidl y
fluctu ating variables act as balancing mechanis ms to protec t the
ongo ing constancy of those variables in which change is normally
slow and of small amplitud e; and that any interferen ce which fixes
the values of the changefu l variables must have a disturbing effect
upon the constancy of the normally steady components of the
system. For the man who must constantly pant at high altitudes , the
respiration rate can no long er be used as a changeab le quan tity in the
maintain ing of physiologic al balance. Converse ly, if the respira tion
rate is to become avail-able again as a rapidly fluctua ting variable,
some change must occu r among the more stable components of the
system. Such a change will, in the nature of the case, be achieved
comparativ ely slowly and be comparat ively irrever sible.
Even acclimation and habit formation are, howeve r, still
revers ible within the life of the individua l, and this very reve rsibility
indica tes a lack of communica tiona l econo my in these adapt ive
mechanis ms. Reve rsibility implies that the changed value of some
variable is achiev ed by means of homeostat ic, error-activated
circuits. There must be a means of detecting an undes irable or
threatening change in some variable, and there must be a train of
cause and effect whereby corrective action is initiated. Moreover,
this entire circuit must, in some degree, be available for this
purpose for the entire time during which the reversible change is
maintained—a considerable using up of available message
pathways.
The matter of communicational economics becomes still more
serious when we note that the homeostatic circuits of an organism
are not separate but complexly interlocked, e.g., hormonal
messengers which play a part in the homeostatic control of organ
A will also affect the states of organs B, C, and D. Any special
ongoing loading of the circuit controlling A will therefore
diminish the organism's freedom to control B, C, and D.
4 W. R. Ashb y, "The Effect of Controls on Stabilit y," Nature, 1945, 155: 242;
also Ashb y, Design for a Brain, New York, John Wiley & Co., 1952.
357

In contr ast, the changes brought about by mutation or other
genot ypic change are presumably of a totally different nature. Ever y
cell contains a copy of the new genotypic corpu s and therefor e will
(when approp riate) behav e in the changed manner, without any
change in the messages which it receives from surround ing tissue s
or organs. If the hypothetical pregiraffes carrying the mutant gene
"long neck" could also get the gene "big heart," their hearts would
be enlarged without the necess ity of using the homeostat ic pathwa ys
of the body to achieve and maintain this enlargement. Such a
mutation will have surviva l value not be-caus e it enable s the
pregiraffe to supp ly its elevated head with sufficient blood, since
this was already achieved by somatic change but because it
increases the overall flexibility of the organism, enabling it to
survive other demands which may be placed upon it either by
environmental or, genotypic chang e.
It appea rs, then, that the process of biolog ical evolu tion could be
continuous if there were a class of mutations or other genot ypic
changes which would simulate Lamarcki an inheritance . The
function of these chang es would be to achieve by geno typic flat
those characteristics which the organism at the given time is already
achieving by the unecono mical method of somatic change . Such a
hypothes is, I believe, confl icts in no way with conven tiona l theories
of genetics and natural selection. It does, however , somewhat alter
the current convent ional picture of evolu tion as a whole , though
related ideas were put forward over sixty years ago. Baldwin5
sugges ted that we consider not only the opera tion of the extern al
environment in natural selection but also what he called "organic
selection" in which the fate of a given variation would depend upon
its physiologic viability. In the same article, Baldwin attribu tes to
Lloyd Morgan the sugge stion that there might exist "coinciden t
variations" which would simulate Lamarckian inheri tance (the so-
called "Baldwin e ffect").
According to such a hypothesis , genotypic change in an organism
beco mes comparable to legislative change in a socie ty. The wise
legislator will only rarely initiate a new rule of behavio r; more
usual ly he will confine himself to affirming in law that which has
5 J. M. Baldwin, "Organic Selection," Science, 1897, 5: 634.
358

alread y beco me the custom of the people. An innovat ive rule can be
introdu ced only at the price of activating and perhap s overloading a
large number o f homeostatic circuits in the socie ty.
It is interesting to ask how a hypothet ical proces s of evolution
would work if Lamarckian inheritance were the rule, i.e., if
characteristics achieved by somatic homeostas is were inherited. The
answer i s simple: it would not work, for the following r easons :
(1)The question turns upon the concept of economy in the use
of homeostatic circuits, and it would be the reverse of economical
to fix by genotypic change all the variables which accompany a
given desirable and homeostatically achieved characteristic. Every
such characteristic is achieved by ancillary homeostatic changes all
around the circuits, and it is most undesirable that these ancillary
changes should be fixed by inheritance, as would logically happen
according to any theory involving an indiscriminate Lamarckian
inheritance. Those who would defend a Lamarckian theory must be
prepared to suggest how in the genotype an appropriate selection
can be achieved. Without such a selection, the inheritance of
acquired characteristics would merely in-crease the proportion of
nonviable genotypic changes.
(2)Lamarckian inheritance would disturb the relative timing of
the processe s upon which evolution must—according to the presen t
hypothesis—d epend. It is essential that there be a time lag between
the unecono mical but reversible somatic achievement of a given
characteristic and the economical but more enduring alterations of
the genotype. If we look upon every soma as a working model
which can be modified in various ways in the workshop, it is clear
that sufficient but not infinite time must be given for these
workshop trials before the results of these trials are incorporated
into the final blueprint for mass production. This delay is provided
by the indirection of stochastic process. It would be unduly
shortened by Lamarckian inheritance.
The principle involved here is gene ral and by no means trivial. It
obtains in all homeosta tic systems in which a given effect can be
brought abou t by means of a homeosta tic circuit, which circuit can,
in turn, be modified in its characteristics by some- highe r system of
control. In all such systems (ranging from the house thermostat to
systems of government and administration) it is important that the
359

highe r system of control lag behind the event sequence s in the
periphera l homeostatic circuit.
In evolu tion two control systems are present: the homeostases of
the body which deal with tolerab le internal stress, and the action of
natural selection upon the (gene tically) nonviab le members of the
popula tion. From an enginee ring point of view, the probl em is to
limit communica tion from the lower , revers ible somatic system to
the highe r irreversible genot ypic s ystem.
Another aspect of the propo sed hypothes is abou t which we can
only specul ate is the probab le relative frequenc y of the two classes
of geno typic change : those which initiate something new and those
which affirm some homeostat ically achieved characteristic. In the
Metazoa and multice llular plants, we face complex networks of
multiple interlock ing homeostat ic circuits, and any given mutation
or gene recombination which initiates change will probabl y require
very various and multipl e somatic charac teristics to be achieved by
homeostas is. The hypothet ical preg iraffe with the mutant gene "long
neck" will need to modify not only its heart and circula tory system
but also perhaps its semicircul ar canals, its interverteb ral discs, its
postu ral reflexes, the ratio of length and thickness of many muscles,
its evasive tactics vis-a-vis preda tors, etc. This sugges ts that in such
complex organisms, the merely affirmative geno typic changes must
far outnumber those which initiate change, if the species is to avoid
that cul-de-sac in which t he flexibility of the soma approa ches z ero.
Converse ly, this picture sugge sts that most organisms, at any
given time, are probab ly in such a state that there are multiple
possib ilities for affirmative geno typic change . If, as seems probable,
both mutation and gene redistribution are in some sense rando m
pheno mena, a t least the chance s are con sider able t hat one o r othe r of
these multiple possibilities will be met.
Finally, it is appropr iate to discuss what evidence is avail-able or
might be sough t to suppo rt or disprov e such a hypothesis . It is clear
at the outset that such a testing will be difficult. The affirmative
mutations upon which the hypothes is depends will usual ly be
invisible. From among the many members of a popula tion which
are achiev ing a given adjustment to environmental circumstance s by
somatic chang e, it wil l not be poss ible i mmediate ly to pick out those
few in which the same adjust ment is provided by the geno typic
360

method. In such a case, the genotypicall y changed individuals will
have to be identified by breeding and raising the offspring under
more normal condi tions .
A still greater difficulty arises in cases where we would
invest igate those homeostatically acqui red characteristics which are
achieved in respon se to some innov ative genot ypic change . It will
often be impossible , by mere inspec tion of the organism, to tell
which of its characteristics are the primary results of genot ypic
change and which a re seconda ry somatic a djust ments to t hese. In the
imaginar y case of the pregi raffe with a somewhat elongat ed neck
and an enlarged heart, it may be easy to guess that the modifica tion
of the neck is geno typic while that of the heart is somatic. But all
such guesses will depend upon the very imperfec t present knowl-
edge of what a n organism can ach ieve in way of somatic adjus tment.
It is a major traged y that the Lamarckian controve rsy has
deflected the attention of gene ticists away from the pheno menon of
somatic adap tability. After all, the mechanisms, thresholds , and
maxima of individu al phenot ypic chang e under stress must sure ly be
geno typicall y determined.
Another difficulty, of rather similar nature, arises at the
population level, where we encounter another "economics" of
potential change, theoretically distinguishable from that which
operates within the individua l. The popu lation of a wild species is
today convent ional ly regarded as geno typically heterogeneou s in
spite of the high degree of super ficial resemblance between the
individu al phenot ypes. Such a popu lation expectab ly function s as a
storehous e of genot ypic possibilities. The econo mic aspect of this
storehous e of possibilities has, for example, been stressed by
Simmonds.6 He points out that farmers and breeders who demand
100 pe r cent ph enotypic uniformity in a h ighly select c rop ar e in fa ct
throwing away most of the multiple gene tic possibilities
accumulated through hundr eds of gene rations in the wild
popu lation. From this Simmonds argues that there is urgent need for
institutions which shall "conse rve" this storehouse of variability by
maintain ing uns elected popu lations.
6 N. W. Simmonds, "Variability in Crop Plants, Its Use and Conser vation," Biol.
Review, 1962, 37: 422-62.
361

Lerne r7 has argued that self-corrective or buffering mechanis ms
opera te to hold constant the composit ion of these mixtures of wild
genot ypes and to resist the effects of artificial selection. There is
therefor e at least a presu mption that this econo mics of variability
within the popu lation wil l turn out to be of the multiplicative k ind.
Now, the difficulty of discriminating between a chara cteristic
achieved by somatic homeostasi s and the same characteristic
achieved (more econo mical ly) by a geno typic short cut is clearly
going to be compounded when we come to consider popul ation s
instead of physiolog ic individua ls. All actual expe rimentation in the
field will inevitably work with popu lations, and, in this work, it will
be necessa ry to discriminate the effects of that econo mics of
flexibility which operates inside the individua ls from the effects of
the econo mics of variab ility which operates at the popul ation level.
These two orders of econo mics may be easy to separate in theory,
but to separ ate them in expe rimentation will surely be difficult.
Be all tha t as it m ay, let us cons ider what eviden tial sup-port m ay
be available for some of the propo sitions which are crucial to the
hypothes is:
(1) That the pheno mena of somatic adjustment are appropriately
described in terms of an econo mics of flexibility. In general, we
believe that the presence of stress A may reduce an organism's
ability to respond to stress B and, guided by this opinion, we
commonly protect the sick from the weather. Those who have
adjusted to the office life may have difficulty in climbing
mountains, and trained mountain climbers may have difficulty
when confined to offices; the stresses of retirement from business
may be lethal; and so on. But scientific knowledge of these
matters, in man or other organisms, is very slight.
(2) That this econo mics of flexibility has the logica l structure
described above— each successiv e demand upon flexibility
fraction ating the set of available possibilities. The proposition is
expectable, but so far as I know there is no evidence for it. It is,
however, worthwhile to examine the criteria which determine
whether a given "economic" system is more appropriately
7 I. M. Lerner, Genetic Homeostasis, Edinburgh, Oliver and Bo yd, 1954 .
362

described in additive or multiplicative terms. There would seem to
be two such criteria:
(a)A system will be additive insofar as the units of its currency
are mutually interchangeable and, therefore, can-not meaningfully
be classified into sets such as were used earlier in this paper to
show that the economics of flexibility must surely be
multiplicative. Calories in the economics of energy are completely
interchangeable and unclassifiable, as are dollars in the individual
budget. Both these systems are therefore additive. The
permutations and combinations of variables which define the states
of an organism are classifiable and—to this extent—
noninterchangeable. The system is therefore multiplicative. Its
mathematics will resemble that of information theory or negative
entropy rather than that of money or energy conservation.
(b)A system will be additive insofar as the units of its currenc y
are mutuall y independ ent. Here there would seem to be a difference
between the econo mic system of the individua l, whose budge tary
proble ms are additive (or sub-t ractive) and those of society at large,
where the overall distribution or flow of wealth is governed by
complex (and perhaps imperfec t) homeostat ic systems. Is there,
perhaps , an econo mics of econo mic flexibility (a metaecono mics)
which is multiplicative and so resembles the econo mics of
physiologi cal flexibi lity discussed above? Notic e, however , that the
units of this wider econo mics will be not dollars but patterns of
distribut ion of weal th. Similarly, Lerner's "genet ic homeostas is,"
insofa r as it is truly homeosta tic, will have m ultiplicative character.
The matter is, howeve r, not simple and we cannot expect that
every system will be either totally multipl icative or totally additive.
There will be intermediate cases which combine the two
characteristics. Speci fically, wher e several independ ent alterna tive
homeostatic circuits control a single variab le, it is clear that the
system may show additive characteristics—and even that it may pay
to incorpor ate such alternative pathwa ys in the system provided the y
can be effectively insulated from each other. Such systems of
multipl e alternative controls may give surviva l advantage insofar as
the mathematics of addition and subtraction will pay better than the
mathematics of logical fraction ation .
363

(3) That innova tive geno typic change commonl y makes demands
upon the adjustive ability of the soma. This proposition is
orthodox ly believed by biologis ts but cannot in the nature of the
case be verified by direct evidence .
(4) That success ive genotyp ic innova tions make multiplicative
demands upon the soma. This proposition (which involves both the
notion of multiplicative economics of flexibility and the notion
that each innovative genotypic change has its somatic price) has
several interesting and perhaps verifiable implications.
(a)We may expect that organisms in which numerous recent
genotypic changes have accumulated (e.g., as a result of selection,
or planned breeding) will be delicate, i.e., will need to be protected
from environmental stress. This sensitivity to stress is to be
expected in new breeds of domesticated animals and plants and
experimentally produced organisms carrying either several mutant
genes or unusual (i.e., recently achieved) genotypic combinations.
(b)We may expect that for such organisms further genotypic
innovation (of any kind other than the affirmative changes
discussed above) will be progressively deleterious.
(c)Such new and special breeds should become more resistant
both to environmental stress and to genotypic change, as selection
works upon successive generations to favor those individuals in
which "genetic assimilation of acquired characteristics" is achieved
(Proposition 5).
(5) That environmental ly induced acquir ed chara cteristics may,
under approp riate cond itions of selection, be replaced by similar
characteristics which are genet ically determined. This phenomenon
has been demonstra ted by Wadding ton8 for the bithorax pheno types
of Dro sophil a.
He calls it the "genetic assimilation of acqu ired charac teristics."
Similar pheno mena have also probab ly occur red in various
exper iments when the expe rimenters set out to prove the inheritanc e
of acqu ired characteristics but did not achieve this proof through
failure to control the condi tions of selection. We have, however , no
eviden ce at all as to the frequ ency of this phenomenon of genetic
8 C. H. Waddington, "Genetic Assimilation of an Acquired Chara cter,"
Evolution, 1953, 7: 118; also Waddington, The Strategy of Genes, London,
Allen and Unwin, 1957.
364

assimilation . It is worth noting , howeve r, that, according to the
arguments of this essay, it may be impossibl e, in princ iple, to
exclude the factor of selection from experiments which would test
"the inhe ritance of acquir ed chara cteristics." It is precisely my thesis
that the simulat ion of Lamarckian inheritance will have surviva l
value und er circumstance of unde fined or multiple stress.
(6). That it is, in gene ral, more econo mical of flexib ility to
achieve a given chara cteristic by genot ypic than b y soma-tic change .
Here the Waddington exper iments do not throw any light, because it
was the exper imenter who did the selecting. To test this propo sition,
we need experiments in which the popula tion of organisms is placed
unde r doub le stress: (a) that stress which will induce the char acteris-
tic in which we are interes ted, and (b) a second stress which will
selectively decimate the popu lation, favoring, we hope, the survival
of those individu als whose flexib ility is more able to meet this
second s tress after adjusting to the first. According to the hypothesis ,
such a system shou ld favor those individua ls which achieve their
adjustment to the first stress by geno typic proce ss.
(7) Finally, it is interesting to consider a corollary which is the
conv erse of the thesis of this essay. It has been argued here that
simulated Lamarckian inheritance will have surviva l value when the
popu lation must adjus t to a stress which remains const ant over
success ive gener ation s. This case is in fact the one which has been
examined by those who would demonstrate an inheritance of
acquired characteristics. A converse problem is presented by those
cases in which a population faces a stress which changes its
intensity unpredictably and rather often—perhaps every two or
three generations. Such situations are perhaps very rare in nature,
but could be produced in the laboratory.
Under such variable circumstances, it might pay the organisms
in survival terms to achieve the converse of the genetic
assimilation of acquired characteristics. That is, they might
profitably hand over to somatic homeostatic mechanisms the
control of some characteristic which had previously been more
rigidly controlled by the genotype.
It is evident, however, that such experimentation would be very
difficult. Merely to establish the genetic assimilation of such
characteristics as bithorax requires selection on an astronomical
365

scale, the final population in which the genetically determined
bithorax individuals can be found being a selected sample from a
potential population of something like 1050. or 1060 individuals. It is
very doubtful whether, after this selective process, there would still
. exist in the sample enough genetic heterogeneity to undergo a
further converse selection favoring those individuals which still
achieve their bithorax phenotype by somatic means.
Nevertheless, though this converse corollary is possibly not
demonstrable in the laboratory, something of the sort seems to
operate in the broad picture of evolution. The matter may be
presented in dramatic form by considering the dichotomy between
"regulators" and "adjusters."9 Prosser proposes that where internal
physiology contains some variable of the same dimensions as some
external environmental variable, it is convenient to classify
organisms according to the degree to which they hold the internal
variable constant in spite of changes in the external variable. Thus,
the homoiothermic animals are classified as "regulators" in regard
to temperature while the poikilothermic are "adjusters." The same
dichotomy can be applied to aquatic animals according to how
they handle internal and external osmotic pressure.
We usually think of regulators as being in some broad
evolu tiona ry sense "higher" than adjus ters. Let us now consider
what this might mean. If there is a broad evolution ary trend in favor
of regulators , is this trend consis tent with what has been said above
about the survival benef its which accru e when contro l is transferred
to genot ypic m echanisms?
Clear ly, not only the regulators but also the adjust ers must rely
upon homeosta tic mechanisms. If life is to go on, a large number of
essen tial physiolog ical variable s must be held within narrow limits.
It the internal osmotic pressu re, for example, is allowed to change ,
there must be mechanis ms which will defend these essen tial
variable s. It follows that the differenc e between adjuste rs and
regulators is a matter of wher e, in the complex network of
physiologic causes and effects, homeostatic pro cess opera tes.
9 C. L. Prosser, "Physiological Variation in Animals," Biol. Review, 1955, 30:
22-262.
366

In the regula tors, the homeostatic proces ses operate at or close to
the input and output point s of that network which is the individu al
organis m. In the adjus ters, the environmental variab les are per mitted
to enter the bod y and the or ganism must then cope with their effects,
using mechanisms which will involve deeper loops of the total
network.
In terms of this analysis, the polarity between adjusters and
regula tors can be extrapola ted another: step to include what we may
call "extraregula tors" which achieve homeostatic controls outside
the body by changing and contro lling the environment—man being
the m ost consp icuous e xample o f this class.
In the earlier part of this essay, it was argued that in adjusting to
high altitude there is a bene fit to be obtained, in terms of an
econo mics of flexib ility, by shifting from, e.g., panting to the more
profound and less reversib le chang es of acclimation; that habit is
more econo mical than trial and error; and that genotypic contro l
may be more econo mical than acclimation. These are all centripetal
changes i n the location o f con trol.
In the broad picture of evolu tion, howeve r, it seems that the trend
is in the opposi te direction: that natural selection, in the long run,
favors regula tors more than adjus ters, and extraregula tors more than
regula tors. This seems to indicate that there is a long time
evolutionar y advan tage to be gained by centrifuga l shifts in the
locus o f con trol.
To specula te about proble ms so vast is perhaps romantic, but it is
worth noting that this contra st between the overa ll evolu tiona ry
trend and the trend in a popu lation faced with constant stress is
what we might expect from the converse corollary here being
considered. If constant stress favors centripetal shift in the locus of
control, and variable stress favors centrifugal shift, then it should
follow that in the vast spans of time and change which determine
the broad evolutionary picture, centrifugal shift of control will be
favored.
367

Summa ry
In this essay the author uses a dedu ctive approach. Starting from
premises of conven tiona l physiolog y and evolut ionary theor y and
applying to these the arguments of cybernetics, he shows that there
must be an economics of somat ic flexibility and that this econo mics
must, in the long run, be coercive upon the evolutionar y process .
Extern al adap tation by mutation or genot ypic reshuffling, as
ordin arily thought of, will inevitably use up the available somatic
flexibi lity. It follows—if evolut ion is to be continuous— that there
must also be a class of geno typic changes which will confe r a bonus
of somatic flexibi lity.
In gene ral, the somatic achieve ment of change is uneconomical
becaus e the process depend s upon homeostas is, i.e., upon whol e
circuits of interdependent variable s. It follows that inheritance of
acqui red charac teristics would be lethal to the evolu tiona ry system
becaus e it would fix the values of these variables all around the
circuits. The organis m or species would , however , bene fit (in
survival terms) by genotypic change which would simula te
Lamarckian inheritance , i.e., would bring about the adapt ive
component of somatic homeostas is without involv ing the whol e
homeosta tic circuit. Such a genot ypic change ( erroneousl y called t he
"Bald-win effect") would confer a bonus of somatic flexibility and
would t herefore have m arked su rviva l value.
Finally, it is sugges ted that a contrary argument can be applied in
those cases wher e a popula tion must acclimate to variabl e stress.
Here n atural selection should favor an anti-Baldwin effect.
368

Problems in Cetac ean and Other
Mammalian Comm unication*
The Communi cation o f Preve rbal Mamma ls*
Of the Cetace a I have had little expe rience. I once dissected in
the Cambridge Zoolog ical Labora tories a specimen of Phocoena
bough t from the local fishmonger, and did not really encounte r
cetace ans again until this year, when I had an oppor tunity to meet
Dr. Lilly's dolph ins. I hope that my discussion of some of the
question s that are in my mind as I approach these peculiar mammals
will assist you in examining either these or related quest ions.
My previous work in the fields of anth ropolog y, animal etholog y,
and psychiatric theory provides a theoretical framework for the
transa ction al analysis of behav ior. The premises of this theoretical
position may be briefly summarized: (1) that a relationsh ip between
two (or more) organis ms is, in-fact, a sequence of S-R sequences
(i.e.,. of contex ts in which proto-l earning occu rs) ; (2) that deutero-
learning (i.e., learning to learn) is, in fact, the acqu iring of
information about the contingenc y patterns of the contex ts in which
proto-l earning occurs; and (3) that the "chara cter" of the organism is
the aggreg ate of its deutero-le arning and therefore reflects the
contextu al patterns of past proto learn ing.10
* This article appeared as Chapter 25, pp. 569-799, in Whales, Dolphins
and Porpoises, edited b y Kenneth S. Norris , Uni versit y of Cali fornia Pr ess, 1966.
Reprinted by permission of The Regents of the Universit y of California.10 J. Ruesch and G. Bateson, Communication: The Social Matrix of Psychiatry,
New York, Norton , 1951 .
369

These premises are essentially a hierarch ic structuring of learning
theory along lines related to Russel l's Theory of Logic al Types.11
The premises, following the Theory of Types, are primarily
approp riate for the analysis of digital communica tion. To what
exten t they may be applicab le to analogi c communicat ion or to
systems that combine the digital with the analog ic is probl ematic. I
hope that the study of dolphin communicat ion will throw light on
these fundamental probl ems. The point is not either to discover that
dolph ins have complex language or to teach them Engl ish, but to
close gaps in our theoretical knowl edge of communication by
studying a system that, whether rudimentar y or complex, is almost
certainly of a totally unfamiliar kind.
Let me start from the fact that the dolphin is a mammal. This fact
has, of cours e, all sorts of implic ation s for anatomy and physiolog y,
but it is not with these that I am conce rned. I am interes ted in his
communication , in what is called his "behav ior," looked at as an
aggreg ate of data percep tible and meaningful to other members of
the same species. It is meaningfu l, first, in the sense that it affects a
recipien t animal's behavio r, and, second, in the sense that
perceptib le failure to achieve appropri ate meaning in the first sense
will affect the behavio r of both animals. What I say to you may be
totally ineffective, but my ineffectiveness, if perceptible, will affect
both you and me. I stress this point becau se it must be remembered
that in all relationships between man and some other animal,
espec ially when that animal is a dolph in, a very large propor tion of
the behavior of both organisms is determined by this kind of
ineffectivenes s.
When I view the behav ior of dolphins as communication, the
mammalian label implies, for me, something very definite. Let me
illustrate what I have in mind by an example from Benson
Ginsbur g's wolf pack in the Brookfie ld Zoo .
Among the Canidae, weaning is performed by the mother. When
the pupp y asks for milk, she presse s down with her open mouth on
the back of his neck, crush ing him down to the ground . She does
this repeatedly until he stops asking . This method is used by
11 A. N. Whitehead and B. Russell , Principia Mathematica, London, Cambridg e
Universit y Press, 1910 .
370

coyotes, dingo es, and the domestic dog. Among wolv es the system
is different. The puppie s graduate smoothl y from the nipple to
regurgitated food. The pack comes back to the den with their bellies
full. All regurgitate what they have got and all eat togeth er. At some
point the adults start to wean the pupp ies from these meals, using
the method employed by the other Canidae; the adult crushes the
pupp y down by pressing its open mouth on the back of the pupp y's
neck. In the wolf this function is not confined to the mother, but is
performed by adults of both sex es.
The pack leader of the Chicago pack is a magnificent male
animal who endlessly patrols the acre of land to which the pack is
confined. He moves with a beautiful trot that appears tireless,
while the other eight or nine members of the pack spend most of
their time dozing. When the females come in heat they usually
proposition the leader, bumping against him with their rear ends.
Usually, however, he does not respond, though he does act to
prevent other males from getting the females. Last year one of
these males succeeded in establishing coitus with a female. As in
the other Canidae, the male wolf is locked in the female, unable to
withdraw his penis, and this animal was helpless. Up rushed the
pack leader. What did he do to the helpless male who dared to
infringe the leader's prerogatives? Anthropomorphism would
suggest that he would tear the helpless male to pieces. But no. The
film shows that he pressed down the head of the offending male
four times with his open jaws and then simply walked away.
What are the implications for research from this illustration?
What the pack leader does is not describab le, or only insuf ficiently
describ ed, in S-R terms. He does not "negat ively reinforce" the othe r
male's sexual activity. He asserts or affirms the nature of the
relationsh ip between himself and the other. If we were to translate
the pack leader 's action into words , the words would not be "Don't
do that." Rathe r, they would translate the metaphoric action: "I am
your senior adult male, you pupp y!" What I am trying to say about
wolve s in particular, and about preverba l mammals in genera l, is
that their discou rse is primaril y about the rules and the contingenc ies
of relationship .
Let me offer a more familiar example to help bring home to you
the generality of this view, which is by no means orthodox among
371

ethologists. When your cat is trying to tell you to give her food,
how does she do it? She has no word for food or for milk. What
she does is to make movements and sounds that are
characteristically those that a kitten makes to a mother cat. If we
were to translate the cat's message into words, it would not be
correct to say that she is crying "Milk!" Rather, she is saying
something like "Ma-ma!" Or, perhaps still more correctly, we
should say that she is asserting "Dependency! Dependency!" The
cat talks in terms of patterns and contingencies of relationship, and
from this talk it is up to you to take a deductive step, guessing that
it is milk that the cat wants. It is the necessity for this deductive
step which marks the difference between preverbal mammalian
communication and both the communication of bees and the
languages of men.
What was extraordinary—the great new thing—in the evolution
of human language was not the discovery of abstraction or
generalization, but the discovery of how to be specific about
something other than relationship. Indeed, this discovery, though it
has been achieved, has scarcely affected the behavior even of
human beings. If A says to B, "The plane is scheduled to leave at
6.30," B rarely accepts this remark as simply and solely a
statement of fact about the plane. More often he devotes a few
neurons to the question, "What does A's telling me this indicate for
my relationship to A?" Our mammalian ancestry is very near the
surface, despite recently acquired linguistic tricks.
Be that as it may, my first expectation in studying dolph in
communication is that it will prove to have the gene ral mammalian
characteristic of being primaril y about relation ship. This premise is
in itself perhaps sufficient to accoun t for the sporad ic development
of large brains among mammals. We need not complain that, as
eleph ants do not talk and whales invent no mousetraps , these
creature s are not overtly intelligen t. All that is needed is to suppose
that large-brained creatu res were, at some evolutionar y stage ,
unwise enough to get into the game of relationship and that, once
the species was caugh t in this game of interpr eting its members'
behavior toward one another as relevant to this complex and vital
subject, there was survival value for those individuals who could
play the game with greater ingenuity or greater wisdom. We may,
372

then, reasonably expect to find a high complexity of
communication about relationship among the Cetacea. Because
they are mammals, we may expect that their communication will
be about, and primarily in terms of, patterns and contingencies of
relationship. Be-cause they are social and large-brained, we may
expect a high degree of complexity in their communication.
Methodo logic al Cons idera tions
The above hypothes is introdu ces very special difficulties into the
proble m of how to test what is called the "psychology" (e.g.,
intelligence, ingenui ty, discrimination, etc.) of individu al animals. A
simple discrimination experiment, such as has been run in the Lilly
labora tories, and no doubt elsewhere , involves a series of steps: (1)
The dolph in may or may not perceive a difference between the
stimulus objects, X and Y. (2) The dolph in may or may not perceive
that this difference is a cue to behavio r. (3) The dolph in may or may
not perceiv e that the behavio r in question has a good or bad effect
upon reinforc ement, that is, that doing "righ t" is condi tiona lly
followed by fish. (4) The dolph in may or may not choose to do
"right ," even after he knows which is right. Succe ss in the first three
steps merely provides the dolphin with a further choice point. This
extra degre e of freedo m must be the first focus o f our investigations.
It must be our first focus for methodolog ical reason s. Consider
the arguments that are conv ention ally based upon experiments of
this kind. We argue alwa ys from the later steps in the series to the
earlier step s. We say, "If the animal was able to ach ieve step 2 in our
experiment, then he must have been able to achiev e step 1." If he
could learn to behave in the way that would bring him the reward ,
then he must have had the necessar y sensor y acuity to discriminate
between X and Y, and s o on.
Precise ly becaus e we want to argue from observa tion of the
animal's succe ss in the later steps to conclus ions about the more
elementary steps, it becomes of prime importance to know whether
the organism with which we are dealing is capable of step 4. If it is
capable, then all arguments about steps 1 through 3 will be
373

invalidated unless appropriate methods of controlling step 4 are
built into the experimental design. Curiously enough, though
human beings are fully capable of step 4, psychologists working
with human subjects have been able to study steps 1 through 3
without taking special care to exclude the confusions introduced
by this fact. If the human subject is "cooperative and sane," he
usually responds to the testing situation by repressing most of his
impulses to modify his behavior according to his personal view of
his relationship to the experimenter. The words coop erative and
sane imply a degree of consistency at the level of step 4. The
psychologist operates by a sort of petitio princ ipii: if the subject is
cooperative and sane (i.e., if the relational rules are fairly
constant), the psychologist need not worry about changes in those
rules.
The proble m of method becomes entirely different when the
subje ct is noncoopera tive, psychopath ic, schizophren ic, a naught y
child, or a dolph in. Perhaps the most fascinating chara cteristic of
this animal is derived precise ly from his ability to operate at this
relatively high l evel, an ab ility that is still to be demonstrated .
Let me now conside r for a moment the art of the animal trainer.
From conversa tions with these highly skilled peop le —trainers of
both dolphin s and guide dogs— my impression is that the first
requirement of a trainer is that he must be able to preven t the animal
from exerting choice at the level of step 4. It must continu ally be
made clear to the animal that, when he knows what is the right thing
to do in a given context, that is the only thing he can do, and no non-
sense about it. In other words, it is a primary cond ition of circus
succe ss that the animal shall abroga te the use of certain higher levels
of his intelligence . The art of the hypnotist is similar .
There is a story told of Dr. Samuel Johnson. A silly lady made
her dog perform tricks in his presence . The Doctor seemed
unimpressed. The lady said, "But Dr. Johnson, you don't know how
difficult it is for the dog." Dr. John son re-plied, "Difficult, madam?
Would it were impossib le!"
What is amazing about circus tricks is that the animal can
abrogate the use of so much of his intelligence and still have
enough left to perform the trick. I regard the conscious intelligence
374

as the greatest ornament of the human mind. But many authorities,
from the Zen masters to Sigmund Freud, have stressed the
ingenuity of the less conscious and perhaps more archaic level.
Com munica tion About Rela tionsh ip
As I said earlier, I expect dolphin communicat ion to be of an
almost totally unfamiliar kind. Let me expand on this point. As
mammals, we are familiar with, though largely unconsc ious of, the
habit of communicating about our relationships . Like other
terrestrial mammals, we do most of our communicating on this
subject by means of kines ic and paralinguis tic signals , such as
bodily movements, involun tary tensions of volun tary muscles ,
changes of facial expression, hesitations, shifts in tempo of speech
or movement, overtones of the voice, and irregul arities of
respiration. If you want to know what the bark of a dog "means,"
you look at his lips, the hair on the back of his neck, his tail, and so
on. These "expressive" parts of his body tell you at what object of
the environment he is barking , and what patterns of relationship to
that objec t he is likely to follow in the next few seconds . Above all,
you look a t his sense organs: his eyes, his ears, and h is no se.
In all mammals, the organs of sense beco me also organs for the
transmission of messages about relationsh ip. A blind man makes us
unco mfortabl e, not because he cannot see tha t is his proble m and we
are only dimly aware of it—bu t because he does not transmit to us
through the movement of his eyes the messages we expect and need
so that we may know and be sure of the state of our relationship to
him. We shall not know much about dolphin communicat ion until
we know what one- dolphin can read in anothe r's use, direction,
volume, and pitch o f echoloca tion.
Perhaps it is this lack in us which makes the communication of
dolphins seem mysterious and opaque, but I suspec t a more
375

profound explanation. Adaptat ion to life in the ocean has stripped
the whales of facial expres sion. They have no external ears to flap
and few if any erectile hairs. Even the cervica l vertebrae are fused
into a sol id block i n many species , and evo lution has st reamlined the
body, sacrificing the expre ssiven ess of separate parts to the
locomotion of the whole . Moreover , condi tions of life in the sea are
such that even if a dolph in had a mobile face, the details of his ex-
pression would be visible to other dolph ins only at rather short
range , even i n the clearest water s.
It is reasonabl e, then, to suppose that in these animals
vocal ization has taken over the communica tive function s that most
animals perform by facial expres sion, wagging tails, clenched fists,
supina ted hands, flaring nostrils, and the like. We might say that the
whale is the communication al opposite of the giraffe; it has no neck,
but has a voice. This specu lation alone would make the
communication of dolphins a subjec t of great theor etical interes t. It
would be fascinating, for example, to know whether or not, in an
evolu tiona ry shift from kinesics to vocal ization, the same general
structu re of categor ies is retained .
My own impression— and it is only an impress ion unsupported
by testing—is that the hypothesis that dolphins have substituted
paralingu istics for kines ics does not quite fit in with my expe rienc e
when I listen to their sounds. We terrestrial mammals are familiar
with pa ralinguis tic co mmunication; we us e it ou rselv es in g runts and
groans , laugh ter and sobbing , modulat ions of breath while speak ing,
and so on. Therefo re we do not find the paralinguis tic sound s of
other mammals to tally opaqu e. We learn rather easily to recogni ze in
them certain kinds of greeting, patho s, rage, persuas ion, and
territoriality, though our guesse s may often be wrong. But when we
hear the sounds of dolphin s we cannot even guess at their
significance . I do not quite trust the hunch that would explain the
sounds of dolphins as merely an elabora tion of the paralinguis tics of
other mammals. (To argue thus from our inability is, however ,
weaker than to argue from wha t we c an do .)
I persona lly do not believe that the do lphins have any-thing that a
human linguis t would call a "language ." I do not think that any
animal withou t hand s would be stupid enough to arrive at so
outlandish a mode of communication.
376

To use a syntax and category system appropriate for the
discussion of things that can be handled, while really discussing
the patterns and contingencies of relationship, is fantastic. But that,
I submit, is what is happening in this room. I stand here and talk
while you listen and watch. I try to convince you, try to get you to
see things my way, try to earn your respect, try to indicate my
respect for you, challenge you, and so on. What is really taking
place is a discussion of the patterns of our relationship, all
according to the rules of a scientific conference about whales. So it
is to be human.
I simply do not believe that dolphins have language in this
sense. But I do believe that, like ourselves and other mammals,
they are preoccupied with the patterns of their relationships. Let us
call this discussion of patterns of relationship the t function of the
message. After all, it was the cat who showed us the great
importance of this function by her mewing. Preverbal mammals
communicate about things, when they must, by using what are
primarily µ-function signals. In contrast, human beings use
language, which is primarily oriented toward things, to discuss
relationships. The cat asks for milk by saying "Dependency," and I
ask for your attention and perhaps respect by talking about whales.
But we do not know that dolphins, in their communication, re-
semble either me or the cat. They may have a quite different
system.
Analogic versus Digital Communi cation
There is ano ther s ide of t he prob lem. How does i t happ en tha t the
paralinguis tics and kines ics of men from strange cultures, and even
the paralingu istics of other terrestrial mammals, are at least partly
intelligible to us, whereas the verba l langu ages of men from strange
cultures seem to be totally opaque? In this respect it would seem that
the vocalizations of the dolph in resemble human language rather
than t he kinesics or paralinguis tics of terrestrial mammals.
We know , of course , why gestures and tones of voice are partly
intelligible while foreign languag es are unintelligible. It is because
377

languag e is digital and kinesics and paralinguis tics are analogic .12
The essen ce of the matter is that in digital communication a number
of purely convent ional signs -1, 2, 3, X, Y, and so on—ar e pushed
around according to rules called algorithms. The signs themselves
have no simple conn ection (e.g., correspondence of magnitude) with
what they stand for. The numeral "5" is not bigger than the numeral
"3." It is true that if we remove the crossba r from "7" we obtain the
numeral "1" ; but t he cro ssbar doe s not, i n any sens e, stand for "6 ." A
name usual ly has only a purely conven tiona l or arbitrary conne ction
with the class named. The numeral "5" is only the name of a
magnitude . It is non-sen se to ask if my telephone number is larger
than yours, because the telephone exchange is a purely digital com-
puter. It is not fed with magnitude s, but only with names of po sitions
on a matrix.
In analog ic communication, howeve r, real magnitudes are used,
and they correspond to real magnitud es in the subject of discou rse.
The linked range finder of a camera is a familiar example of an
analogu e computer. This devic e is fed with an angle that has real
magnitude and is, in fact, the angle that the base of the range finder
subtend s at some point on the object to be photographed . This angle
controls a cam that in turn moves the lens of the camera forwa rd or
back . The secret of the device lies in the shape of the cam, which is
an analog ic repres entation (i.e., a picture, a Cartesian graph) of the
functiona l relation ship between distance of object and distance of
image.
Verbal language i s almost (but no t quite) pu rely digital. The word
"big" is not bigge r than the word "little"; and in gener al there is
nothing in the pattern (i.e., the system of interrelated magnitudes ) in
the word "table " which would correspond to the system of
interrelated magnitudes in the objec t deno ted. On the other hand, in
kines ic and paralinguistic communication, the magnitude of the
gesture, the loudness of the voice, the length of the pause, the
12 The difference between digital and analogic modes of communication may
perhaps be made clear by thinking of an English-speaking mathematician confronted
with a paper by a Japanese colleague. He gazes uncomprehendingly at the Japanese
ideographs, but he is able partly to understand the Cartesian graphs in the Japanese
publication. The ideographs, though they may originally have been analogic pictures,
are now purely digital; the Cartesian graphs are analo gic.
378

tension of the muscle, and so forth—these magnitudes commonly
correspond (directly or inversely) to magnitudes in the relationship
that is the subject of discourse. The pattern of action in the
communication of the wolf pack leader is immediately intelligible
when we have data about the weaning practices of the animal, for
the weaning practices are themselves analogic kinesic signals.
It is logical, then, to consider the hypothesis that the
vocalization of dolphins may be a digital expression of µ
functions. It is this possibility that I especially have in mind in
saying that this communication may be of an almost totally
unfamiliar kind. Man, it is true, has a few words for µ functions,
words like "love," "respect," "dependency," and so on. But these
words function poorly in the actual discussion of relationship
between participants in the relationship. If you say to a girl, "I love
you," she is likely to pay more attention to the accompanying
kinesics and paralinguistics than to the words themselves.
We humans become very uncomfortable when somebody starts
to interpret our postures and gestures by translating them into
words about relationship. We much prefer that our messages on
this subject remain analogic, unconscious, and involuntary. We
tend to distrust the man who can simulate messages about
relationship. We therefore have no idea what it is like to be a
species with even a very simple and rudimentary digital system
whose primary subject matter would be µ functions. This system is
something we terrestrial mammals cannot imagine and for which
we have no empathy.
Research P lans
The most specula tive part of my pape r is the discuss ion of plans
for the testing and amplific ation of such a body of hypotheses . I
shall be guided by the following heuristic assumptions :
(1) The epistemology in whose terms the hypotheses are
constructed is itself not subject to testing. Derived from Whiteh ead
379

and Russ ell,13 it serves to guide our work . Shou ld the work prove
reward ing, the succe ss will be only a weak verification of the
epistemology.
(2)We do not even know what a primitive digital system for the
discu ssion of patterns of relationsh ip might look like, but we can
guess that it would not look like a "thing" language . (It might, more
probab ly, resemble music.) I shall therefore not expec t the
techn iques for cracking human linguis tic code s to be immediately
applicabl e to the voca lization of dolphin s.
(3)The first requirement, then, is to identify and to classify the
varieties and the component s of relationship existing among the
animals through detailed ethologi cal study of their action s,
interac tions , and social organization. The elements of which these
patterns are built are doubt less still present in the kinesics and
actions of the specie s. We there-fore begin with a listing of the
kines ic signa ls of individu al dolphins , and then try to relate them to
the contexts in which t hey are us ed.
(4)No doub t, just as the pack leader's behav ior tells us that "dominance" among wolves is metaphor ically related to wean ing,
so also the dolph ins will tell us their kinesic metaphors for "dominance ," "dependenc y," and other µ functions . Gradual ly this
system of signal s will fit togeth er piece by piece to form a picture
of the varieties of relationsh ip existing even among animals
arbitrarily conf ined together in a tank.
(5)As we begin to unders tand the metaphor system of the
dolph in, it will become possib le to recogn ize and classify the
contex ts of his voca lization. At this point the statistical techn iques
for cracking codes may conceiv ably become useful.
The assumptions regarding the hierarch ic structu re of the
learning proce ss—upon which this whol e pape r is based —provide
the basis for various kinds of exper imentat ion. The contex ts of
proto -learning may be variousl y constructed with a view to
observ ing in what types of contexts certain types of learning most
readily occu r. We shall pay special attention to those contexts that
involve either relationsh ips between two or more animals and one
person , or relationships between two or more people and one
13 'Whitehead a nd Russell , op. cit.
380

animal. Such contexts are miniature models of social organization
within which the animal may be expected to show characteristic
behaviors and to make characteristic attempts to modify the
context (i.e., to manipulate the humans).
Comm ents
Mr. Wood: In the course of twelve years in Marine Studios in
Florida, I spent a great deal of time watching what was perhaps the
most natural assemblage of Tursiop s in captivity, includ ing animals
of variou s ages, usual ly two or more of them in the process of
growing up, and I saw remarkabl y little of what you are going to
look for in a much more restricted group of animals in the Virgin
Islands .
One time I saw something very interes ting. Early one morning
abou t six o r six-thirty, over a p eriod o f at least h alf an hou r, the adu lt
male assumed a position next to one of the females in the tank who
was hang ing motionless in the cur-rent. He would go up
occasiona lly and move away and then come back and assume a
position beside her, and he would stroke her side with his right
flippe r repeatedly. There was no indication that this had sexual
signif icanc e. There was no erection on the part of the male, and no
observabl e respon se on the pa rt of t he fe male. But i t was as c lear-cut
a nonvoca l signal as I ever observed in the tank.
Mr. Bateson: I would like to say that the amount of signal ing that
goes on is much greater than is evident at first sight. There are, of
course, the rather specific kinds of signals which are very importan t.
I am not denying that. I mean the touching , and so on. But the shy
individu al, the traumatized female, staying almost stationa ry three
feet be-low the surface while two other individua ls fool around , is
getting a great deal of attention just by sitting there and staying. She
may not be actively transmitting, but in this busines s of bodily
communica tion, you don't have to be actively transmitting in order
to have your signa ls picked up by other peop le. You can just be, and
just by being she attracts an enormous amount of attention from
these other two indiv idual s who come over, pass by, pause a little as
381

they pass, and so on. She is, we would say, "withdrawn," but she is
actually about as withdrawn as a schizophrenic who by being
withdrawn becomes the center of gravity of the family. All other
members of the group move around the fact of her withdrawal,
which she never lets them for-get.
Dr. Ray: I tend to agree with Mr. Bateson . We are working at the
New York Aquar ium with the beluga whale , and I believe these
animals are much more expressive than we like to suspec t. I think
one of the reasons they don't do very much in captivity is that they
are bo red to t ears most of the t ime. There is nothing m uch of interest
in their tank environment, and I would like to suggest that we have
to manipula te their captiv ity much more clever ly than we do. I don't
mean handl ing the whale s. They don't like that. But the introduct ion
of different types of animals, or clever little things that we might do
would get them to respond more. Capt ive cetaceans are like
monkeys in a cage. They are highly intelligent and highl y
develop ed, and t hey are bo red.
Another factor is our skill in observation, and in the beluga
whale , at least, we have been able to notice visua lly the sounds they
are m aking by watching th e change i n the shap e of the m elon, which
is extremely marked in this animal. It can swell on one side or the
other, or take several different shapes correlated with sound
produc tion. So, by very careful observa tion and/or skilled
manipula tion, I think a great deal can be done with these animals
rather simply.
Mr. Bateson: I had meant to point out that all sense organs
among mammals, and even among ants, beco me major organs for
the transmission of messages, such as, "Where are the other fellow's
eyes focused?" and, "Are his pinnae focused in one direction or
anothe r?" In this way sense organs beco me transmitting organs for
signa ls.
One of the things we must absolutely acquire if we are going to
unders tand dolphins is a knowledge of what one animal knows and
can read from anoth er animals' use of sonar. I suspect the presence
of all sorts of courtesy rules in this busin ess; it probabl y isn't polite
to sonar scan your friends too much, just as among human beings it
is not polite, really, to look at anothe r's feet in detail. We have many
382

taboos on observing one anothers' kinesics, because too much
information can be got in that way.
Dr. Purves: It seems to me that the dolphin or the cetacean must
suffer from an even greater disadvantage than man has in the past,
because—I have forgotten the authority—it has been said that the
origin of human speech is an analogue language. In other words, if
you use the word "down," you lower the hand and lower the lower
jaw at the same time. If you say "up," you raise the hand and raise
the lower jaw. And if you use the word "table," and, better still,
pronounce it in French, your mouth widens out and you make a
horizontal gesture. However complicated the human language is, it
has its origin in an analogue language. The poor porpoise has
nothing like this to start from. So he must have been highly
intelligent to have developed a communication system completely
de novo .
Mr. Bateson : What has happened to this creature is that the
information we get visually and the other terrestrial animals get
visually must have been pushed into voice. I still maintain that it is
appropriate for us to start by investigating what is left of the visual
material.
383

A Re-exami nation of "Bateson's Rule "*
Introduct ion
Nearl y eighty years ago, my father, William Bateson , be-came
fascinated by the pheno mena of symmetry and metameric regularity
as exhibited in the morpholog y of animals and plants. It is difficult
today to defin e precisely what he was after, but, broad ly, it is clear
that he believed that an entirely new concep t of the nature of living
things would develop from the study of such pheno mena. He held,
no doubt correctly, that natural selection could not be the only
determinant of the direction of evolu tiona ry chang e and that the
genes is of variation could not be a rando m matter. He therefore set
out to demonstra te regularity and "lawfulnes s" among the
pheno mena of variability.
In his attempt to demonstrate a sort of order which the biologis ts
of his day had largely ignored, he was guided by the notion , neve r
clearly formulated, that the place to look for regularity in variation
would be precisely where variation had its impact upon what was
already regula r and repetitive. The pheno mena of symmetry and
metamerism, themselves striking ly regular, must surely have been
brought abou t by regula rities or "laws" within the evolutionar y
proce ss and, therefore, the variations of symmetry and metamerism
should precisely exemplif y these laws a t work.
In the languag e of today, we might say that he was groping for
those orderly chara cteristics of living things which illustrate the fact
that organis ms evolve and develop with-i n cybernet ic,
organizational , and oth er communicationa l limitations.
It was f or this study that he coined t he word "gene tics."14
He set out to examine the materia l in the world's museums,
private collections, and journ als bearing upon the teratology of
* This essay has been accepted for publication in the Journal of Genetics, and
is here reproduced with the permission of that journal .
384

animal symmetry and metamerism. The de-tails of this survey were
published in a large book15 which is still of cons iderab le interest.
To demonstrate regular ity within the field of teratolog ical
variation, he attempted a classification of the various sorts of
modifica tion that he encoun tered. With this classification I am not
here conce rned, except that in the survey he happened upon a
generalization which can be called a "discovery." This discover y
came to be called "Bateson's Rule" and remains one of the
unexp lained m ysteries of biolog y.
The purpose of the present note is to place Bateson's Rule in a
new theoretical perspe ctive determined by cybernetics, information
theory, and the like.
Briefly, Bateson's Rule asserts in its simplest form that when an
asymmetrical lateral appendage (e.g., a right hand) is redup licated,
the resulting redup licated limb will be bilaterally symmetrica l,
consisting of two parts each a mirror image of the other and so
placed that a plane of symmetry could be imagined between them.
He himself was, however , very doub tful whethe r such simple
redupl ication ever occurs . He believed and accumulated evidence to
show that, in a very large propor tion of such cases, one component
of the redupli cated system was it-self double . He asserted that in
such systems the three components are normally in one plane; that
the two componen ts of the doub let are mirror images of each other;
and that that component of the doublet which is the nearer to the
primary appendage is a mirror image of the primary.
This gene ralization was shown by my father to hold for a very
large number of examples of redupl ication in the vertebrates and in
arthropod s, and for a few cases in other phyla where the museum
materia l was , of course, more scarce .
Ross Harrison16 believed that Bateson underes timated the
importance of simple redup lication.
14 1W. Bateson, "The Progress of Genetic Research," In-augural Address, Royal
Horticultural Society Report, 1906. 15 'W. Bateson, Materials for the Study of Variation, London, Macmillan and Co.,
1894.16 R. G. Harrison, "On Relations of Symmetry in Transplanted Limbs,"
Journal of Experimental Zoology, 1921, 32: 1-118.
385

Whether or not simple redupl ication is a real and common
pheno menon, I shall begin this essay with a discuss ion of the logical
probl ems which i t would p resen t.
The Probl em Rede fined
In 1894 , it appeared that the problem centered around the
quest ion: What causes the develop ment of bilateral symmetry in a
contex t wher e it does no t belong?
But modern theory has turned all such quest ions upside down.
Information , in the technic al sense , is that which excludes certain
alterna tives. The machine with a governor does not elect the stead y
state; it prevents itself from staying in any alterna tive state; and in
all such cybernet ic systems, corrective action is brought abou t by
difference. In the jargon of the engin eers, the system is "error
activated." The difference between some presen t state and some "preferred " state activates the corrective response .
The techni cal term. "information" may be succin ctly de-fined as
any difference which makes a difference in some later event . This
definition is funda mental for all analysis of cybernet ic systems and
organization. The definition links such analysis to the rest of
science, where the cause s of even ts are commonly not differences
but forces, impacts, and the like. The link is classically exemplified
by the heat engin e, where available energy (i.e., negative entrop y) is
a func tion of a difference between two te mperatures. In t his classical
instance , "infor mation" a nd "nega tive entropy" overlap.
Moreov er, the energy relations of such cybernetic systems are
commonly inve rted. Becaus e organisms are able to sto re energy, it is
usual that the energy expendi ture is, for limited periods of time, an
inverse function of energy in-put. The amoeba is more active when
it lacks food, and the stem of a green plant grows faster on that side
which i s turned away from the light.
Let us therefore invert the question abou t the symmetry of the
total redup licated appendage: Why is this doubl e append age not
asymmetrical like the corresponding appendages of normal
organisms?
386

To this quest ion a formal and gene ral (but not particula r) answer
can be const ructed on the following l ines:
(1) An unfer tilized f rog's egg i s rad ially symmetrical, with a nimal
and vegeta l poles but no differen tiation of its equato rial radii. Such
an egg develops into a bilaterally symmetrica l embryo, but how
does it select one meridian to be the plane of bilateral symmetry of
that embryo? The answer is known—that , in fact, the frog's egg
receiv es information from the outside. The point of entry of the
spermatozoon (or the prick of a fine fiber) marks one meridian as
different from all others , and that meridian is the future plane of
bilateral symmetry.
Conver se cases can also be cited. Plants of many families bear
bilaterally symmetrica l flowers. Such flowers are all clearly deriv ed
from triadic radial symmetry (as in orchid s) or from pentad ic
symmetry (as in Labia tae, Legu minosae , etc.) ; and the bilateral
symmetry is achiev ed by the differentiation of one axis (e.g., the
"standa rd" of the familiar swee t pea) of this radial symmetry. We
again ask how it is possible to select one of the similar three (or
five) axes. And again we find that each flower receives information
from the outside. Such bilaterally symmetrical flowers can only be
produced on branch stems, and the differentiation of the flower is
always oriented to the manner in which the flower -bear ing branch
stem comes off from the main stem. Very occas ional ly a plan t which
normally bears bilaterally symmetrica l flowers will form a flower at
the terminus of a main stem. Such a flower is necess arily only radial
in its symmetry—a cup-shaped monstros ity. (The problem of bi-
laterally asymmetrica l flowe rs, e.g., in the Catase tum group of
orchids , is interesting. Presu mabl y these must be borne, like the
lateral appendages of animals, upon branch es from main stems
which are themselves alread y bilaterally symmetrical, e.g., dorso-
ventrally flattened .)
(2) We note then that, in biologi cal systems, the step from radial
symmetry to bilateral symmetry commonly requi res a piece of
information from the outside. It is, however , conc eivab le that some
divergent process might be touched off by minute and rando mly
distribut ed differences , e.g., among the radi i of the frog's egg. In thi s
case, of course , the selection of a particula r meridian for special
development would itself be rando m and could not be oriented to
387

other parts of the organism as is the plane of bilateral symmetry in
sweet peas and l abiate flower s.
(3) Similar conside rations apply to the step from bilateral
symmetry to asymmetry. Again either the asymmetry (the
differen tiation of one half from the other) must be achieved by a
rando m process or it m ust be achieved b y information received fr om
the outside, i.e., from neighboring tissues and organs. Ever y lateral
appendage of a vertebrate or arthropod is more or less
asymmetrical17 and the asymmetry is neve r set rando mly in relation
to the rest of the animal. Right limbs are not borne upon the left side
of the body, except under expe rimental circumstances . Therefore the
asymmetry must depend upon the outside information , presumabl y
derived f rom the neighboring tissues .
(3) But i f the step f rom bilateral symmetry to asymmetry requir es
additiona l information , then it follows that in absence of this
additiona l information , the appendage which shou ld have been
asymmetrical can on ly be bilaterally symmetrica l.
The probl em of the bi lateral s ymmetry of redupl icated limbs thus
becomes simply a problem of the loss of a piece of information.
This follows from the general logical rule that every reduction in
symmetry (from radial to bilateral or from bilateral to
asymmetrical) requires additional in-formation.
It is not claimed that the above argument is an explana tion of all
the pheno mena which illustrate Bateson's Rule. Indeed, the
argument is offered only to show that there are simple ways of
thinking about these pheno mena which have sca rcely been explored .
What is propo sed is a family of hypotheses rather than a single one.
A critical examination of what has been said above as if it were a
single hypothesi s will, how-eve r, provide a further illustration of the
method.
17 In this connection, scales and feathers and hairs are of special interest. A
feather would seem to have a very clear bilateral symme try in which the plane of
symme try is related to the antero- posterior di fferentiation of the bird. Superposed on
this is an asymme try like that of the individual bilateral limbs. As in the case of lateral
limbs, corresponding feathers on opposite sides of the body are mirror images of
each other. Every feather is, as it were, a flag whose shape and colorin g denote the
values of deter mining variables at the point and time of its growth.
388

In any given case of redupl ication, it will be necessa ry to decide
what particular piece of information has been lost, and the argument
so far given shou ld make this decision easy. A natural first guess
would be that the deve loping appendage needs three sorts of
orienting information to en-abl e it to achieve asymmetry: proximo-
distal information ; dorso-ven tral information ; and antero-pos terior
information. The simplest hypothesi s sugges ts that these might be
separate ly received and therefore that one of these sorts of
information will be lost or absent in any given case of redupl ication.
It should then be easy to classify cases of redupl ication ac-cord ing
to which piece of orienting information is missing . There should be
at most three such t ypes of redupli cation, and t hese should be clearly
distinct.
Supernume rary D ouble Legs in Co leopt era
But in the only set of cases wher e this deduct ion can be tested,
facts clearly do not fit the hypothesis. The cases are those of
supernumerary pairs of appendages i n bee tles. About a hundr ed such
cases were known in 1894 , and of these Bateson18 describ es abou t
half and f igure s thirteen.
The formal relations are remarkabl y unifo rm and leave no doub t
that a single type of explan ation shou ld apply to the symmetry in all
cases.
18 W. Bateson, Materials . . . , op. cit., pp. 477-503.
389

390Fig. 2 Pterostichus muhlfeldii , No. 742. Semidiag rammatic representation of the
left middle tibia bearing the extra tarsi upon the antero- ventral border of the apex.
L, the normal tarsus; R, the extra right; L' the extra left tarsus. ( The property of
Dr. Kraatz. ) From Bateson , W., Materials for the Study of Variation , London:
Macmillan, 1894 , p 485.
Fig. 1 Carabus scheidleri, No. 736. The normal right fore leg, R, bearing an
extra pair of legs, SL and SR', arising from the ventral surface of the coxa, C. Seen
from in front. (The propert y of Dr. Kraatz.) From Bateson , W., Materials for the
Study of Variation, London: Macmillan, 1894, p 483.

391

Typical ly19 one leg (rarely more than one) of a beet le is abnormal
in bearing a branch at some point in its length. This branch is
19 See Figures 1 and 2, pages 385 and 386.
392Fig. 5 A mechanical device for showing the relations that extra legs in
Secondar y Symmetry bear to each other and to the normal leg from which
they arise. The model R represents a normal right leg. SL and SR represent
respectivel y the extra right and extra left legs of the supernumerar y pair. A
and P, the anterior and posterior spurs of the tibia. In each leg the
morphologically anterior surface is shaded, the posterior being white. R is
seen from the ventral aspect and SL and SR are in Position VP. From
Bateson , W., Materials for the Study of Variation, London: Macmillan,
1894 , p. 480 .

regula rly a doubl et, consis ting of two parts which may be fused at
the point of branching off from the primary leg but which are
commonly separate at their distal end s.
Distally from the point of branching there are thus three
component s—a primary leg and two supernu merar y legs. These
three lie in one plane and have the following symmetry: the two
component s of the supernumerary doub let are a complementary pair
—one being a left and the other a right— as Bateson's Rule would
sugg est. Of these two, the leg nearest to the primary leg is
comple mentar y to it.
These relation s are repres ented in Figure 3. (See page 387.) Each
component is shown in diagra mmatic cross section, and their dorsa l,
ventral, anterior, and pos terior fa ces ar e indi cated b y the le tters D, V,
A, and P , respec tively.
What is surpr ising about these abnormalities—in that it conflicts
with the hypothesis offered above—is that there is no clear
discont inuity by which the cases can be classified according to
which sort of orien ting information has been lost. The
supernumerary double t may be borne on any part of the
circumference of the primary leg.
Figure 3 illustrates the symmetry of a double t occur ring in the
dorsal region . Figure 4 (page 387) illustrates the symmetry of a
doub let in the dorso -ante rior region.
It appea rs, then, that the planes of symmetry are parallel to a
tangent of the circumferenc e of the primary leg at the point of
branching but, since the points of branching may be anywhere on
the circumference , a continuous series of possible bilateral
symmetries is genera ted.
Figure 5 (page 388) is a machine invented by W. Bateson to
demonstrate this continuous s eries of possible bilateral symmetries .
If the bilateral symmetry of the doub let is due to a loss of
orienting information , we should expect the plane of that bilateral
symmetry to be at right angles to the direction of the lost
information; i .e., if dorso -vent ral information were lo st, the resulting
limbs or doubl et should conta in a plane of symmetry which would
be at right angle s to the dor so-vent ral line.
(The argument fo r this expec tation m ay be spelled out as follows :
a gradien t in a lineal sequence creates a difference between the two
393

ends of the sequen ce. If this gradient is not presen t, then the ends of
the sequence will be similar, i.e., the sequence will be symmetrical
about a plane of symmetry transve rse to itself. Or, consid er the case
of the frog's egg. The two poles and the point of entry of the
spermatozoon determine a plane of bilateral symmetry. To achieve
asymmetry, the egg requi res information at right angles to this plane,
i.e., something which will make the right half differen t from the left.
If this something is lost, then the egg will revert to the original
bilateral symmetry, with the original plane of symmetry transve rse
to the direction of the lost information. )
As noted above , the supernu merar y doub lets may originate from
any face of the primary leg, and therefo re all intermediates occur
between the expec tedly discon tinuous types of loss of information .
It follows that if bilateral symmetry in these doub lets is due to loss
of information , then the information lost cannot be classified as
antero-pos terior, dorso-ven tral, or proxi mo-distal.
The hypothesis must therefore be corrected.
Let us retain the genera l notion of lost information , and the
corollary of this that the plane of bilateral symmetry must be at right
angle s to the direction of the information t hat was l ost.
The next simplest hypothesis sugges ts that the lost information
must have been centro-per iphera l. (I here retain this bipola r term
rather than use the simpler " radia l.")
Let us imagine, then, some centro -periphera l differen ce —
possib ly a chemical or electrical gradient within the cross section
of the primary leg; and suppo se that the loss or blurring of this
differenc e at some point along the length of the primary leg
determines that any branch limb produced at this point shall fail to
achieve a symmetry.
It will follow , naturally, that such a branch limb (if produced)
will be bilaterally symmetrica l and that its plane of bilateral
symmetry will be at right angle s to the direction of the lost gradient
or difference .
But, clearly, a centro-periph eral difference or gradien t is not a
primary componen t of that information system which determined
the asymmetry of the primary leg. Such a gradient might, however,
inhibit branching, so that its loss or blurring would result in
production of a supernumerary branch at the point of loss.
394

The matter becomes superf icially paradoxic al: the loss of a
gradien t which might inhibit branching results in branch formation,
such tha t the br anch canno t achiev e asymmetry. It appea rs, then, that
the hypothet ical Centro -periphera l gradien t or difference may have
two sorts of command functions : (a) to inhib it branching ; and (b)
to determine an asymmetry in that branch which can only come into
existence at all if the Cent ro-pe riphe ral gradient is absen t. If these
two sorts of message functions can be shown to overlap or be in
some sense synonymous, we shall have gene rated an econo mical
hypotheti cal description o f the pheno mena.
We therefore address ourse lves to the quest ion: Is there an a
priori case for expec ting that the absence of a gradien t which
would prohib it branch ing in the primary leg will permit the
formation of a branch which will lack the information necessar y to
determine asymmetry across a plane at right angles to the missing
gradien t?
The question must be inverted to fit the upside-downness of all
cybernet ic explanation. The conc ept "infor mation necess ary to
determine asymmetry" then becomes "infor mation necessar y to
prohibit bilateral symmetry."
But anything which "prohib its bilateral symmetry" will also
"prohibi t branching, " since the two components of a branching
structure constitute a symmetrica l pair (even though the component s
may be radially symmetrical).
It therefo re beco mes reasonabl e to expec t that loss or blurring of
a Centro -periphera l gradien t which prohibits branch formation will
permit the formation of a branch which will, howeve r, itself be
bilaterally symmetrica l abou t a plane parallel to the circumference
of the primary limb.
Meanwhile, within the primary limb, it is possible that a
Centro-peripheral gradient, by preventing branch formation, could
have a function in preserving a previously deter-mined asymmetry.
The above hypotheses provide a possible framework of
explanation of the formation of the supernu merar y doub let and the
bilateral symmetry within it . It re mains to cons ider the or ienta tion of
the componen ts of that doub let. According to Bateson' s Rule , the
component neares t to the primary leg is in bilateral symmetry with
it. In other words , that face of the supernumerary which is toward
395

the primary is the morpholog ical counte rpart of that face of the
peripher y of the primary from which the branch sprang.
The simplest, and perhaps obvious, explana tion of this regul arity
is that in the process of branch ing there was a sharing of
morphologi cally differen tiated structures between branch and
primary and that these shared structu res are, in fact, the carriers of
the necessa ry information. Howeve r, since information carried this
way will clearly have prope r-ties very differen t from those of
information carried by gradients, it is approp riate to spell the matter
out in some detail.
Consider a radially symmetrical cone with circular base. Such a
figure is different iated in the axial dimension , as between apex and
base. All that is neces sary to make the cone fully asymmetrical is to
differen tiate on the circumference of the base two poin ts which shall
be differen t from each other and shall not be in diametrica lly
opposi te positions, i.e., the base must contain such different iation
that to name its parts in clockwis e order gives a result differen t from
the result of naming the pa rts in anticlockwi se order.
Assu me now that the supernu merar y bran ch, by its very origin as
a unit growing out from a matrix, has proximo-distal differentiation ,
and that this differentiation is analogous to the differen tiation in the
axial dimension of the cone. To achieve complete asymmetry, it is
then only necess ary that the developing limb receive directiona l
information in some arc of its circumference . Such information is
clearly immediately available from the circumstance that, at the
point of branch ing, the second ary limb must share some cir-
cumferenc e with the primary. But the shared points which are in
clockwi se order on the periphe ry of the primary will be in
anticlockwi se order on the periphe ry of the branch. The information
from the shared arc will therefor e be such as to determine both that
the resulting limb will be a mirror image of the primary and that the
branch will face a ppropriately toward the primary.
It is now possib le to constru ct a hypotheti cal sequence of events
for the redup lication s in the legs o f bee tles:
(1)A primary leg develops asymmetry, deriving the
necessary information from surrounding tissues.
(2)This information, after it has had its effect, continues to
exist, transformed into morphological differentiation.
396

(3)The asymmetry of the normal primary leg is hence-forth
maintained by a centro-peripheral gradient which normally
prevents branching.
(4)In the abnormal specimens, this centro-peripheral
gradient is lost or blurred—possibly at some point of lesion or
trauma.
(5)Following the loss of the centro-peripheral gradient,
branching occurs.
(6)The resulting branch is a doublet; lacking the gradient
information which would have determined asymmetry, it must
therefore be bilaterally symmetrical.
(7)That component of the doublet which is next to the
primary is oriented to be a mirror image of the primary by the
sharing of differentiated peripheral structures.
(8)Similarly each component of the doublet is itself
asymmetrical, deriving the necessary information from the
morphology of shared peripheries in the plane of the doublet.
The above specul ation s are intended to illustrate how the
explanatory principle of loss of information might be applied to
some of the regularities subsu med under Bateson's Rule. But it will
be noted that the data on symmetry in the legs of beetles have, in
fact, been ove rexpla ined.
Two distinct but not mutually exclusive—t ypes of ex-
planat ion have been invoked: (a) the loss of information which
shou ld have been derived from a centro-pe riphe ral gradien t, and (b)
information de rived from shared peripheral morpholog y.
Neither of these types of explanation is sufficien t by itself to
explain the pheno mena, but when combined the two principles
overlap so that some details of the total picture can be referred
simultaneou sly to both principles.
Such redundan cy is, no doubt , the rule rather than the excep tion
in biologi cal systems, as it is in all other systems of organization,
differentiation, and communica tion. In all such systems, redundanc y
is a major and necessa ry source of stability, predictab ility, and
integration.
Redundanc y within the system will inevitably appear as
overlapping between our explanations of the system. Indeed,
without overlapping, our explanations will commonly be
insufficient, failing to explain the facts of biological integration.
397

We know little abou t how the pathwa ys of evolution ary change
are influenced by such morphogene tic and physiological redund an-
cies. But certainly such internal redund ancies must impose
nonrando m characteristics upon the pheno mena of variation.20
Reduplic ated Limbs in Amphib ia
At this point it is interesting to turn from analysis of
redup lication in beetles' legs to another body of data in which
redup lication commonly occurs and has been referred to Bateson's
Rule.21 These are the data on redupli cation in the exper imental ly
transplan ted limbs of larval newt s.
(1) There are some cases, mostly of heterotopic trans-plan ts in
which the grafted limb bud deve lops into a simple and appa rently
equal binary system, in which the two component s are in mirror
image symmetry. I was shown about three years ago a very striking
prepa ration by Dr. Emerson Hibba rd of the California Institute of
Technolog y. In this specimen the limb bud had been rotated through
180°, so that the anterior edge of the bud faced toward the posterior
end of the host, and had been implanted in a median dorsal position
on the posterior region of the head of the host. This transplan t had
develop ed into two remarkab ly complete legs in mirror image
relationship . This binary system was conne cted to the head of the
host only by a slende r bridge of tissue.
Such prepa rations, wher e the product is binary and the parts
equal , certainly look l ike what would b e expec ted fro m a simple loss
of one dimension of orienting information . (It was Dr. Hibba rd's
specimen that sugg ested to me that the hypothesis of lost
information might be applicable to the amphibian material.)
(2) However , apart from these instances of equal binary
redup lication, the amphibian materia l does not at all fit with any
hypothes is that would explain the redupli cation as due to a simple
20 G. Bateson , "The Role of Somatic Change in Evolution," Evolution , 1962,
17: 529-39.21 Harrison, op. cit.; also F. H. Swett, "On The Production of Double Limbs in
Amphibians," Journal of Experimental Zoology , 1926, 44: 419-72.
398

loss of information. Indeed, if Bateson's Rule were restricted to
cases where the explana tion is formally analogous to that which fits
the redupl ication in the beetles' legs, then the amphibian cases
would probab ly not fall under this rub ric.
The limitat ions of a hypothesis are, howeve r, as important as its
application s, and I shall therefor e summarize here the very complex
data on orthotop ic transp lants .
One schematic paradig m will suffice: if the right anterior limb
bud is excised, turned through 180° and replac ed in the wound, it
will grow to be a left limb. But this primary limb may subsequen tly
form seconda ry limb buds at its base, usual ly either immediately
anterior or posterior to the point of insertion. The seconda ry will be
a mirror image of the primary, and may even later deve lop a tertiary
which will typically be formed outside the secondar y, i.e., on that
side o f the secondar y which i s farthest from the primary.
The formation of the left primary on the right side of the body is
explained22 by assuming that antero-post erior orientation is received
by the limb bud earlier than dorso-v entra l information, and that,
once received , this antero-po sterior information is irreversible. It is
suppo sed that the graft is already antero-poste riorly determined at
the time of grafting but later receive s dorso-ven tral information
from the tissues with which it is now in contac t. The result is a limb
whose dorso-ven tral orientation is correct for its new setting but
whose antero-poste rior orientation is revers ed. It is tacitly assumed
that the proximo-distal orientation of the bud is undis turbed . The
result is a limb which is reversed in regard to one of its three sorts
of asymmetry. Such a limb must logically be a left.
This explanat ion I accept and proceed to consid er the
redupl ications.
These differ in four important respects from the redupl ications in
beetles' legs discu ssed a bove:
(a)In the beetles, the reduplication is usually equal. The two
halves of the supernumerary doublet are equal in size, and are
usually approximately equal in size to the corresponding parts of
the primary leg. Such differences as do appear among the three
22 Swett, op. cit.; also Harrison , op. cit.
399

components are such as might expectably result from trophic
differences. But in the larval newts, great differences in size occur
between the components of the reduplicated system, and it appears
that these differences are determined by time. The secondaries are
smaller than the primaries because they are produced later and,
similarly, the rare tertiaries are later and smaller than the
secondaries. This spacing of events in time indicates clearly that
the primary limb received all the information necessary to
determine its own asymmetry. It received, in-deed, "wrong"
information and grew to be a left leg on the right side of the body
but it did not suffer from such a deficiency of information as
would make it immediately fail to achieve asymmetry. The
reduplication cannot simply be ascribed to loss of orienting
information in the primary.
(b)The reduplications in beetles' legs may occur at any point
along the length of the leg. But those of amphibian larvae usually
arise from the region of attachment of the limb to the body. It is
not even sure that the secondary always shares tissue with the
primary.
(c)In the case of the beetles, the supernumerary doublets form a
continuous series, being given off from any portion of the
periphery of the primary. In contrast, the reduplication of limbs in
amphibian larvae is localized either anterior or posterior to the
primary.
In the beetles it is clear that the two supernumerary components
form together a single unit. In many cases there is actual
compounding of the two components (as in Figure 1). In no case23 is
that component of the double t which is nearer to the primary
compounded with it rather than with the other supernu merar y. In the
amphibian preparations , on the other hand, it is not clear that
seconda ry and tertiary form a subun it. The relation between tertiary
and secondar y seems no clo ser than between seconda ry and pri mary.
Above all, the relation i s asymmetric in the time dimension .
23 Bateson (Materials . . . , op. cit., p. 507) describes and figures one
doubt ful exception to this statement. This is a redupli cation in the left hind tarsus o f
Platycerus caraboides.
400

These profound formal differences between the two bodies of
data indicate that the explanations for the amphibi an data must be of
a different order. It would seem that the processes are located not in
the shaft of the limb but in its base and the tissues surround ing the
base. Tentatively we may guess that the primary in some way
proposes the later formation of a seconda ry by a reversal of gradient
information, and that the seconda ry similarl y proposes a revers ed
tertiary. Models for such systems are available in cybernet ic theory
in those circuit structures which propose Russe llian paradoxes.24 To
attempt to const ruct any such model at the present time would be
premature .
Sum mary
This essay on the symmetry of redup licated lateral append ages
starts from an explan atory principle, viz., that any step of
ontogenet ic different iation which reduces the symmetry of an organ
(e.g., from radial to bilateral symmetry, or from bilateral symmetry
to asymmetry) requires additional orienting information . From this
principle it is argued that a normally asymmetrical lateral
appendage, lacking some necessary piece of orienting information ,
will only be able to achieve bilateral symmetry, i.e., instead of a
normal asymmetrica l appendage , the result will be a bilaterally
symmetrical doubl et.
To examine this explan atory principle, the writer has at-tempted
to const ruct a hypothesi s to explain Bateson's Rule as this regularity
is exemplified in the rare supernu merar y doub le legs of Coleopte ra.
In the construction of this hypothesis, it was assumed that
morphogenetic orienting information may unde rgo transfor mation
from one type of coding to another, and that each transfor m or code
is sub ject to characteristic limitations :
(a) The information may be embodied in gradien ts (perhaps
biochemical). In this coding, the information can be diffused from
neighboring tissues and provide the first determinants of
24 G. Bateson, "Mini mal Require ments for a Theor y of Schizophrenia," A.M.A.
Archives of General Psychiatry, 1960, 2: 477-91 .
401

asymmetry in the developing appendage. It is suggested that
information coded in this way is only briefly available, and that
once the asymmetry of the limb is established, the information
continues to exist, but trans-formed into morphology.
(b) It is suggest ed that information coded as morphological
differenc e is essent ially static. It canno t be diffused to neighbo ring
tissues and it cannot inhibit branching . It can, however , be used by a
branch which at its incept ion shares tissue with the primary limb
from which it branch es off. In this case, the information passed on
by the method of shared peripher y will be necess arily inverted: if
the primary be a right, the branch will be a left.
(c) The information in morphologic al form being (by hypothes is)
unable to inhibit branching , the asymmetry of a growing primary
must be preserved by a centro -periphera l gradient—no t itself a
determinant of that asymmetry.
(d) It is sugge sted that the loss of such a centro-per ipheral
gradient might have two effects : that of permitting branch ing and
that of depriving the resulting branch of one dimension of neces sary
orienting information; so that the branch can only be a bilaterally
symmetrica l unit with a plane of symmetry at right angles to the lost
centroperiphera l gradien t.
The data on redupl ication in the expe rimentally trans-p lanted
limb buds of amphibi a are also exa mined. It i s argued that th ese data
are not to be explained by simple loss of orienting information .
Simple loss, it is sugges ted, will expe ctably result in equal and
synchronous bilateral symmetry. The amphibian redupl icates are, in
gener al, unequa l and successive . In a few cases, synchronous and
equal re -dupl ication occurs in the a mphibian exper iments, especially
in heterotopic implants . Such cases could perhaps be regarded as
due t o simple loss of orienting information .
402

Postscript, 1971
Compare the bilateral symmetry in the supernumerary double t of
the beetle's leg with the bilateral symmetry in the sweet pea or
orchid flower . Both in the plant and in the animal, the bilaterally
symmetrical unit comes off from a point of branching .
In the plant, the morpholog y of the fork provides information
enabling the flower to be not radially but bilaterally symmetrica l,
i.e., information which will differen tiate the "dorsa l" standard from
the ven tral lip of the flower .
In the doub let on the beetle's leg, the plane of bilateral symmetry
is orthogona l to that in the flower .
We might say that the information which the beetle's leg has lost
is precisely that information which the plant creates by the act of
branching .
403

Comm ent on Part IV
The paper s placed togeth er in this part are diverse in that while
each paper is a branch from the main stem of the argument of the
book, these branches come off from very different location s. "The
Role o f Somatic Change in Evolut ion" is an expansion of the though t
behind "Min imal Requ irements for a Theory of Schizophren ia,"
while "Probl ems in Cetace an and Other Mammalian
Communica tion" is an application of "The Logic al Catego ries of
Learn ing and Communica tion" to a particula r type of animal."A Re-examination of Bateson's Rule" may seem to break new
ground , but is related to the remainder of the book in that it ex-tends
the notion of informational contro l to include the field of
morphogenesi s and, by discuss ing what happens in absence of
needed information , brings out the importan ce of the context into
which i nformation i s received.
Samuel Butler, with uncann y insigh t, once commented upon the
analog y between dreams and parthenog enesis . We may say that the
monstrous doub le legs of the beetles share in this analogy: they are
the projection of the receptive context deprived of information
which s hould h ave co me from an external sou rce.
Mess age material, or information, comes out of a context into a
contex t, and in other parts of the book the focus has been on the
contex t out of which information came. Here the focus is rather
upon the internal state of the organis m as a context into which the
information must be received .
Of course, neither focus is sufficient by itself for our under-
stand ing of either animals or men. But it is perhaps not an accident
that in these papers dealing with non-human organisms the "contex t"
which is discus sed is the obver se or comple ment of the "context"
upon which I have focussed attention in other parts of the book.
Consider the case of the unfertilized frog's egg for which the entry
point of the spermatozoon defines the plane of bilateral symmetry of the
future embryo.
The prick of a hair from a camel's hair brush can be substituted and
still carry the same message. From this it seems that the external context
404

out of which the message comes is relatively undefined. From the entry
point alone, the egg learns but little about the external world. But the
internal context into which the message comes must be exceedingly
complex.
The unfertilized egg, then, embodies an immanent question to
which the entry point of spermatozoon provides an answer; and this way
of stating the matter is the contrary or obverse of the conventional view,
which would see the external context of learning as a "question" to which
the "right" behavior of the organism is an answer.
We can even begin to list some of the components of the immanent
question. First there are the already existing poles of the egg and,
necessarily, some polarization of the intervening protoplasm towards
these poles. Without some such structural conditions for the receipt of the
prick of the spermatozoon, this message could have no meaning. The
message must come into an appropriate structure.
But structure alone is not enough. It seems probable that any meridian
of the frog's egg can potentially become the plane of bilateral symmetry
and that, in this, all meridians are alike. It follows that there is, to this
extent, no structural difference between them. But every meridian must
be ready for the activating message, its "readiness" being given direction
but otherwise unrestricted by structure. Readiness, in fact, is precisely
not-structure. If and when the spermatozoon delivers its message, new
structure is generated.
In terms of the economics of flexibility, discussed in "The Role of
Somatic Change in Evolution" and later in "Ecology and Flexibility in
Urban Civilization" (Part VI), this "readiness" is uncommitted
potentiality for change, and we note here that this uncommitted
potentiality is not only always finite in quantity but must be appropriately
located in a structural matrix, which also must be quantitatively finite at
any given time.
These considerations lead naturally into Part V, which I have titled
"Epistemology and Ecology." Perhaps "epistemology" is only another
word for the study of the ecology of mind.
405

Part V: Epistemology and
Ecology

Cyberneti c Explanation*
It may be useful to descr ibe some of the pecul iarities of
cybernet ic explana tion.
Causa l explanation is usually positive. We say that billiard ball B
moved in such and such a direction becaus e billiard ball A hit it at
such and such an angle. In contrast to this, cybernet ic explanat ion is
always negat ive. We conside r what alternative possib ilities could
conceivab ly have occu rred and then ask why many of the
alternatives were not followed , so that the particula r event was one
of those few which could, in fact, occur. The classica l example of
this type of explanation is the theory of evolution unde r natural
selection. Ac-co rding to this theory, those organis ms which were not
both physiologi cally and environ mental ly viable could not possib ly
have lived to reproduce . Therefo re, evolut ion alwa ys followed the
pathwa ys of viability. As Lewi s Carrol l has pointed out, the theory
explains quite satisfac torily why there are no bread-and -butt er-flies
today.
In cybernet ic language , the course of events is said to be subjec t
to restraints, and it is assumed that, apart from such restraints, the
pathwa ys of change would be govern ed only by equality of
probabi lity. In fact, the "restraints" upon which cybernetic
explanation depends can in all cases be regarded as factors which
determine inequa lity of probabi lity. If we find a monkey striking a
typewrit er appa rently at rando m but in fact writing meaningfu l
prose, we shall look for restraints, either inside the monkey or inside
the typewriter. Perhaps the monkey could not strike inappropriate
letters; perhaps the type bars could not move if imprope rly struck ;
perhaps incor rect letters could not survive on the pape r. Somewhere
there must have been a circuit which could identify error and
eliminate i t.
* This article is reprinted from the American Behavioral Scientist, Vol. 10, No. 8,
April 1967, pp. 29-32, by per-mission of the publish er, Sage Publications, Inc.
407

Ideally—and commonly—the actual even t in any sequence or
aggreg ate is unique ly determined within the terms of the cybernetic
explan ation . Restraints of many different kinds may combine to
gener ate this unique determination. For example, the selection of a
piece for a given position in a jigsaw puzzl e is "restrained" by many
factors. Its shape must conform to that of its several neighbor s and
possib ly that of the boundar y of the puzzl e; its color must confo rm
to the color pattern of its region ; the orientation of its edges must
obey the topolog ical regu larities set by the cutting machine in which
the puzz le was made; and so on. From the point of view of the man
who is trying to solve the puzzle, these are all clues, i.e., source s of
information which will guide him in his selection. From the point of
view o f the cybernet ic ob serve r, they are restra ints.
Similarly, from the cybernet ic point of view, a word in a
sentence, or a letter within the word , or the anatomy of some part
within an organism, or the role of a species in an ecosystem, or the
behavio r of a member within a family—these are all to be
(nega tively) explained b y an an alysis of restraints.
The negat ive form of these explanations is precisely comparable
to the form of logical proof by reduct io ad absur dum. In this species
of proof, a sufficient set of mutuall y exclu sive alterna tive
propos itions is enumerated , e.g., "P" and "not P," and the process of
proof proced es by demonstra ting that all but one of this set are
untenab le or "absurd ." It follows that the surviving member of the
set must be tenabl e within the terms of the logical system. This is a
form of proof which the nonmathe matica l sometimes find
unconvinc ing and, no doubt , the theor y of natur al selection
sometimes seems unconvincing to nonmathe matica l persons for
similar reason s—whatever those reasons may be.
Another tactic of mathematica l proof which has its count erpar t in
the construction of cybernet ic explanations is the use of "mapping"
or rigorous metaphor . An algebra ic proposition may, for example, be
mapped onto a system of geometric coordina tes and there proven by
geometric methods . In cyberneti cs, mapping appears as a technique
of explanation whenever a conceptua l "model" is invoked or, more
concr etely, when a computer is used to simulate a complex
communication al proces s. But this is not the only appearan ce of
mapping in this science. Formal proce sses of mapping, translation,
408

or transfo rmation are, in principle, imputed to every step of any
sequence of pheno mena which the cyberneticist is attempting to
explain. These mappings or trans-f ormations may be very complex,
e.g., wher e the output of some machine is rega rded as a transfor m of
the input; or they may be very simple, e.g., where the rotation of a
shaft at a given point along its length is regarded as a transform
(albeit identical) of its rotation a t some previou s point.
The relations which remain constant under such transformation
may be of any conc eivab le kind.
This parallel, between cybernet ic explana tion and the tactics of
logica l or mathematica l proof , is of more than trivial interest.
Outs ide of c ybernet ics, we look for exp lanation, but not f or an ything
which would simulate logical proof. This simulation of proof is
something new. We can say, however , with hindsigh t wisdo m, that
explanation by simulation of logical or mathematical proof was
expectab le. After all, the subject matter of cybernet ics is not event s
and objects but the informat ion "carried" by events and objects. We
consider the objects or events only as propos ing facts, propositions,
messages, percept s, and the like. The subjec t matter being
proposi tiona l, it is expec table that explan ation would simulate the
logica l.
Cybernetic ians have specialized in those explanations which
simulate reduc tio ad absu rdum and "mapping." There are perhap s
whole realms of explana tion awai ting discove ry by some
mathematic ian who will recogn ize, in the informational aspec ts of
nature, sequences which s imulate other types of proof.
Becau se the subject matter of cybernetics is the propositional or
informational aspect of the events and objects in the natur al world ,
this science is forced to procedur es rather different from those of the
other science s. The differentiation, for example, between map and
territory, which the semanticists insist that scientists shall respect
in their writings must, in cybernetics, be watched for in the very
phenomena about which the scientist writes. Expectably,
communicating organisms and badly programmed computers will
mistake map for territory; and the language of the scientist must be
able to cope with such anomalies. In human behavioral systems,
especially in religion and ritual and wherever primary process
409

dominates the scene, the name often is the thing named. The bread
is the Body, and the wine is the Blood.
Similarly, the whole matter of induc tion and dedu ction —and our
doctrinaire preferences for one or the other—will take on a new
significance when we recognize inductive and deduct ive steps not
only in our own a rgument bu t in the relationsh ips a mong data.
Of especial interest in this conn ection is the relationsh ip between
conte xt and its content. A phone me exists as such only in
combinat ion with other phone mes which make up a word . The word
is the conte xt of the phone me. But the word only exists as such—
only has "meaning"—in the larger context of the utterance, which
again has meaning only in a relationship .
This hierarchy of contexts within contexts is unive rsal for the
communication al (or "emic") aspect of pheno mena and drives the
scientist alwa ys to seek for explana tion in the ever larger units. It
may (perhaps) be true in physics that the explana tion of the
macroscop ic is to be sough t in the microscopi c. The oppos ite is
usual ly true in cybernet ics: without context, there is no
communication .
In accord with the nega tive charac ter of cybernetic ex-plana tion,
"info rmation" is quan tified in negative terms. An even t or-obje ct
such as the letter K in a given position in the text of a message
might have been any other of the limited set of twen ty-six letters in
the English language . The actua l letter excludes (i.e., eliminates by
restraint) twenty-five alternatives . In comparison with an English
letter, a Chinese ideograph would have exclud ed sever al thousand
alternatives. We say, therefore, that the Chinese ideograph carries
more information than the letter. The quant ity of information is
convent ional ly expressed as the log to base 2 of the improbab ility of
the actual event or objec t.
Probabi lity, being a ratio between quant ities which have similar
dimensions, is itself of zero dimensions. That is, the central
explanatory quantity, information, is of zero dimensions.
Quantities of real dimensions (mass, length, time) and their
derivatives (force, energy, etc.) have no place in cybernetic
explanation.
The status of energy is of special interest. In genera l in
communication al systems, we deal with sequen ces which resemble
410

stimulus-and-r esponse rather than cause-and -effect. When one
billiard ball strikes another, there is an energy transfer such that the
motion of the second ball is energized by the impact of the first. In
communica tiona l systems, on the other hand , the energy of the
response is usually provided by the respondent . If I kick a dog, his
immediately sequent ial behav ior is ener gized by his metabol ism, not
by my kick. Similarl y, when one neuron fires anothe r, or an impulse
from a microphone activates a circuit, the sequent event has its own
energy sources.
Of course, everything that happens is still within the limits
defined by the law of energy conserv ation . The dog's metabol ism
might in the end limit his response , but, in general, in the systems
with which we deal , the energy suppl ies are large compared with the
demands upon them; and, long before the suppl ies are exhaus ted,
"econo mic" limitat ions a re imposed b y the f inite nu mber of av ailable
alternatives, i.e., there is an econo mics of probabi lity. This
econo mics differs from an econo mics of energy or money in that
probability—being a ratio— is not subjec t to addition or subtraction
but only to multipl icative processe s, such as fractionation. A
telephon e exch ange at a time of emergenc y may be "jammed" when
a large fraction of its alternative pathwa ys are busy. There is, then, a
low probab ility of any given message ge tting through.
In addition to the restraints due to the limited econo mics of
alternatives, two other categor ies of restraint must be discussed :
restraints related to "feedback" and restraints related to
"redundanc y."
We consider first the concep t of feedback:
When the pheno mena of the universe are seen as linked together
by cause -and-ef fect and energy transfe r, the resulting picture is of
complexl y branch ing and interconn ecting chain s of causa tion. In
certain regions of this universe (notably organisms in environments,
ecosystems, thermostats, steam engines with governors , societies,
computers, and the like), these chains of causa tion form circuits
which are closed in the sense that causal interconn ection can be
traced around the circuit and back through whateve r position was
(arbitarily) chosen as the starting point of the description . In such a
circuit, evident ly, event s at any position in the circuit may be
expected to have effect at all positions on the circuit at later times.
411

Such systems are, howeve r, always open: (a) in the sense that the
circuit is energized from some external source and loses energy
usual ly in the form of heat to the outside; and (b) in the sense that
event s within the circuit may be influenced from the outside or may
influence outside e vents .
A very large and i mportant par t of cyberneti c theory is concerned
with the formal charac teristics of such causa l circuits, and the
condi tions of their stability. Here I shall consider such systems only
as source s of restraint.
Consider a variable in the circuit at any position and sup-pose
this variable subjec t to rando m change in value (the chang e perhaps
being imposed by impact of some even t extern al to the circuit). We
now ask how this change will affect the value of this variable at that
later time when the sequence of effects has come around the circuit.
Clear ly the answer to this last question will depend upon the
characteristics of the circuit and will, therefore , be not random.
In princ iple, then, a causa l circuit will gene rate a non-rando m
respon se to a rando m event at that position in the circuit at which the
random event occurred.
This is the gener al requi site for the creation of cybernet ic
restraint in any variable at any given position. The particular
restraint created in any given instance will, of course , depend upon
the characteristics of the particular circuit—whether its overall gain
be positive or negative, its time characteristics, its thresho lds of
activity, etc. These will togethe r determine the restraints which it
will exe rt at any given pos ition.
For purpos es of cyberneti c explan ation , when a machine is
observ ed to be (improbabl y) moving at a constant rate, even under
varying load, we shall look for restraints—e .g., for a circuit which
will be activated by changes in rate and which, when activated, will
opera te upon some variable (e.g., the fuel supp ly) in such a way as
to diminish th e chang e in rate.
When the monkey is observ ed to be (improbabl y) typing prose,
we shall look fo r some circuit whi ch is ac tivated whenever h e makes
a "mistake" and which , when activated , will delete the eviden ce of
that mistake at the position where i t occur red.
The cybernetic method of negat ive explanat ion raises the
quest ion: Is there a difference between "being right" and "not being
412

wrong"? Should we say of the rat in a maze that he has "learned the
right path" or should we say only that he has learned "to avoid the
wrong paths" ?
Subjectively, I feel that I know how to spell a number of English
words , and I am certainly not aware of discard ing as unrewa rding
the letter K when I have to spell the word "many." Yet, in the first
level cyberneti c explanat ion, I should be viewed as actively
discard ing the alternative K when I spell "many."
The question is not trivial and the answer is both subtle and
fundamental : choices are not all at the same level. I may have to avoid
error in my choic e of the word " many" in a given con text, discarding
the alterna tives, "few," "several," "frequent ," etc. But if I can
achieve this higher level choice on a negative base, it follows that
the word "many" and its alternatives somehow must be conceiv able
to me—must exist as distingu ishabl e and possibly labeled or coded
patterns in my neural proce sses. If they do, in some sense, exist,
then it follows that, after making the higher level choice of what
word to use, I shall not necessa rily be faced with alterna tives at the
lower level. It may beco me unnecess ary for me to exclude the letter
K from the word "many." It will be correct to say that I know
positively how to spe ll "many"; not merely that I know how to avo id
making mistake s in spe lling that word.
It follows that Lewis Carro ll's joke abou t the theory of natural
selection is not entirely cogen t. If, in the communicational and
organiza tiona l processes of biolog ical evolut ion, there be something
like levels—it ems, patterns , and possibly patterns of patterns—then
it is logically possible for the evolution ary system to make
something like positive choices. Such levels and patterning might
conceivab ly be in or among genes or elsewher e.
The circuitry of the above mention ed monkey would be required
to recogniz e deviations from "pros e," and prose is chara cterized by
pattern or—as t he enginee rs call it—by redundanc y.
The occurren ce of the letter K in a given location in an English
prose message is not a purely rando m even t in the sense that there
was ever an equa l probab ility that any other of the twenty-five
letters might have occu rred in that location. Some letters are more
common in Engli sh than others , and certain combinations of letters
are more common than others. There is, thus, a species of patterning
413

which partly determines which letters shall occur in which slots. As
a result: if the receive r of the message had rece ived the entire rest of
the message but had not received the particular letter K which we
are discuss ing, he might have been able, with better than rando m
succe ss, to guess that the missing letter was, in fact, K. To the extent
that this was so, the let-ter K did not, for that receiv er, exclude the
other twent y-five letters because these were already partly excluded
by information which the recipien t received from the rest of the
message . This patterning or predic tability of particular events with in
a larger agg regate of even ts is techn ically called "redundanc y."
The concept of redundanc y is usual ly derived , as I have derived
it, by considering first the maximum of information which might be
carried by the given item and then considering how this total might
be reduced by knowledge of the surrounding patterns of which the
given item is a component part. There is, howeve r, a case for
looking at the whole matter the other way round . We might regard
patterning or predictability as the very essence and raison d'etre of
communication , and see the single letter unaccompanied by
collateral clues as a pecu liar and special case.
The idea that communicat ion is the creation of redundancy or
patterning can be applied to the simplest engineering examples. Let
us conside r an observer who is watch ing A send a message to B. The
purpose of the transac tion (from the point of view of A and B) is to
create in B's message pad a sequence of letters identical with the
sequenc e which formerly occu rred in A's pad. But from the point of
view of the observ er this is the creation of redund ancy. If he has
seen what A had on his pad, he will not get any new information
about the message itself from inspe cting B's p ad.
Eviden tly, the nature of "meaning," pattern, redundanc y,
information and the like. depends upon where we sit. In the usual
engine ers' discussion of a message sent from A to B, it is customary
to omit the observer and to say that B received information from A
which was measurab le in terms of the number of letters transmitted ,
reduc ed by such redundan cy in the t ext a s might have p ermitted B to
do some guessing . But in a wider universe , i.e., that defined by the
point of view of the observer, this no longe r appears as a
"transmission" of information but rather as a spreading of
redundan cy. The activities of A and B have combined to make the
414

universe of the observ er more predictable, more ordered, and more
redundant . We may say that the rules of the "game" played by A and
B explain (as "restraints") what would otherwise be a puzzl ing and
improbable coincidence in the observ er's unive rse, namely the
conformity between what i s written on the two m essage pads .
To guess, in essenc e, is to face a cut or slash in the sequen ce of
items and to predict across that slash what items might be on the
other side. The slash may be spatial or temporal (or both) and the
guessing may be either predic tive or retrospec tive. A pattern, in fact,
is def inable a s an aggr egate o f events o r objec ts which wi ll permit in
some degree such guesse s when the entire aggrega te is not available
for inspec tion.
But this sort of patterning is also a very, genera l pheno menon,
outside the realm of communication between organisms. The
recept ion of message materia l by one organism is not funda mentally
different from any other case of perception. If I see the top part of a
tree standing up, I can predict —wi th better than rando m succes s—
that the tree has roots in the ground. The percep t of the tree top is
redundant with (i.e., conta ins "information" about ) parts of the
system which I cannot perceive owing to the slash provided by the
opacity of the ground.
If then we say that a message has "meaning" or is "about" some
referent, what we mean is that there is a larger universe of relevance
consisting of message-plus -referen t, and that redundanc y or pattern
or predictability is introduced into this unive rse b y the message .
If I sa y to you "It is raining ," this message i ntrodu ces redundanc y
into the unive rse, message -plus-r aindrops, so that from the message
alone you could have guessed—wi th better than rando m success—
something of what you would see if you looked out of the window .
The universe , message -plus -referent, is given pattern or form—in
the Shakespea rean sense, the unive rse is informed by the message;
and t he "fo rm" of whi ch we are spe aking is not in t he m essage no r is
it in the referen t. It is a correspondenc e between message and
referent.
In loose talk, it seems simple to locate information. The letter K
in a given slot proposes that the letter in that particular slot is a K.
And, so long as all information is of this very direct kind, the
415

information can be "locat ed": the information about the letter K is
seemingl y in that slot.
The matter is not quite so simple if the text of the message is
redundan t but, if we are lucky and the redund ancy is of low order,
we may still be able to point to parts of the text which indicate
(carry some of the information) that the letter K is expe ctable in that
particular slot.
But if we are asked : Where are such items of information as that:
(a) "This message is in Engl ish"; and ( b) "In Eng lish, a letter K o ften
follows a letter C, excep t when the C begins a word"; we can only
say that such information is not localized in any part of the text but
is rather a statistical induct ion from the text as a whole (or perhaps
from an aggr egate of "similar" texts). This, after all, is metainfo rma-
tion and is of a basically different order—o f different logical type—
from the information t hat "the letter in this slot is K."
This matter of the localization of information has be-dev iled
communication theory and especi ally neurophysiology for many
years and it is, therefore , interesting to consider how the matter
looks if we start from redundanc y, pattern or form as the basic
concep t.
It is flatly obviou s that no variable of zero dimensions can be
truly located. "Info rmation" and "form" resemble contrast,
frequ ency, symmetry, correspondence , congru ence, confor mity, and
the like in being of zero dimensions and, therefor e, are not to be
located. The contr ast between this white paper and that black coffee
is not somewhere between the paper and the coffee and, even if we
bring the pape r and coffee into close juxtaposi tion, the contrast
between them is not thereb y located or pinch ed between them. Nor
is that contrast located between the two objects and my eye. It is not
even in my head; or, if it be, then it must also be in your head. But
you, the reade r, have not seen the pape r and the coffee to which I
was referring. I have in my head an image or transfor m or name of
the con trast between them; and y ou have i n your head a t ransform of
what I have in mine. But the confor mity between us is not localiz-
able. In fact, information and form are not items which can be
localized.
It is, however , possible to beg in (but perhaps not co mplete) a sor t
of mapping of formal relations within a system containing
416

redundanc y. Consider a finite aggregat e of objects or even ts (say a
sequence of letters, or a tree) and an observer who is already
informed about a ll the redundanc y rules which a re recogn izabl e (i.e.,
which have statistical signif icance ) within the aggregat e. It is then
possible to delimit regions of the aggregat e within which the
observer can achieve better than rando m guess ing. A further step
toward localization is accomplished by cutting across these regions
with slash marks, such that it is across these that the educat ed
observer can guess , from what is on one side of the slash, something
of what is on the other side.
Such a mapping of the distribution of patterns is, how-ev er, in
principle, incomplete becau se we have not considered the source s of
the observ er's prior knowl edge of the redundanc y rules. If, now, we
consider an observe r with no prior knowledge , it is clear that he
might discover some of the relevan t rules from his percep tion of
less than the whole aggrega te. He could then use his discove ry in
predic ting rules for the remainder— rules which would be correct
even though not exemplified. He might discove r that "H often
follows T" even though the remainder of the aggregat e contained no
example of this combination. For this order of pheno menon a
different order of slash m ark— metaslash es —wi ll be ne cessa ry.
It is interesting to note that metaslashe s which demarcate what is
necessa ry for the naive observer to discover a rule are, in principle,
displa ced relative to the slashes which would have appear ed on the
map prepared by an observer totally informed as to the rules of
redundanc y for t hat a ggrega te. (This p rinciple i s of so me importance
in aesthetics.
To the a esthe tic eye, the for m of a c rab with one c law bigger than
the other is not simply asymmetrical. It first pro-poses a rule of
symmetry and then subtly denies the rule by propos ing a more
complex combination of rules.)
When we exclude all things and all real dimensions from our
explanatory system, we are left regarding each step in a
communica tiona l sequence as a transform of the previou s step. If
we consider the passag e of an impulse along an axon, we shall
regard the even ts at each point along the pathwa y as a transform
(albeit identical or similar) of even ts at any previous point. Or if we
consider a series of neurons, each firing the next, then the firing of
417

each neuron is a transform of the firing of its predecesso r. We deal
with event sequen ces whi ch do not n ecessa rily imply a passing on of
the same energy.
Similarly, we can consider any network of neurons, and
arbitrarily transect the whole network at a series of different
positions , then we shall regard the even ts at each transection as a
transfor m of even ts at some previous transection.
In conside ring perception , we shall not say, for example, "I see a
tree," because the tree is not within our explanato ry system. At best,
it is on ly possible to se e an i mage which i s a co mplex but systematic
transfor m of the tree. This image, of course, is energized by my
metabol ism and the nature of the transfo rm is, in part, determined by
factors within my neural circuits: "I" make the image, under various
restraints, some of which are imposed by my neura l circuits, while
others are imposed by the external tree. An hallucination or dream
would be more truly "mine" insofar as it is produced withou t
immediate externa l restraints.
All that is not information, not redundanc y, not form and not
restraints— is no ise, the on ly possible sourc e of new patterns.
418

Redund ancy and Coding*
Discussion of the evolu tiona ry and other relationship s between
the communica tion systems of men and those of other animals has
made it very clear that the coding devic es characteristic of verbal
communica tion differ profoundl y from those of kinesi cs and
paralangu age. But the point has been made that there is a great deal
of resemblance between the codes of kines ics and paralangu age and
the cod es of nonhu man mammals.
We may, I think, state categor ically that man's verbal system is
not derived in any simple way from these preponderan tly iconic
codes. There is a genera l popu lar belief that in the evolution of man,
language replac ed the cruder systems of the other animals. I believe
this to be totally wrong and would a rgue a s follows:
In any complex functiona l system capab le of adaptiv e
evolutionar y change, when the performance of a given function is
taken over by some new and more efficien t method, the old method
falls into disuse and deca y. The techn ique of making weapons by the
knapp ing o f flint deteriorated when m etals came into use.
This decay of organs and skills under evolutionar y replacement
is a necessary and inevitable systemic pheno menon. If, therefor e,
verbal langu age were in any sense an evolutionary replacement of
communica tion by means of kinesics and paralanguage , we would
expect the old, preponderan tly iconic systems to have under gone
conspicuous decay. Clearly they have not. Rathe r, the kinesi cs of
men have become richer and more complex, and paralanguage has
blosso med side by side with the evolution of verba l language . Both
kinesi cs and paralanguage have been elabo rated into complex forms
of art, music, ballet, poetry, and the like, and, even in everyday life,
the intricacie s of human kines ic communication , facial express ion,
and vocal intona tion far exceed anything that any other animal is
known to produce. The logician's dream that men shou ld
* This essay appeared as Chapter 22 in Animal Communication: Techniques of
Study and Results of Research, edited by Thomas A. Sebeok. Copyright 1968 by
Indiana University Press. Reprinted by permission of the publisher .
419

communicate onl y by unambiguous digi tal sign als has not come true
and i s not likely to.
I suggest that this separate burgeoning evolut ion of kinesics and
paralanguage along side the evolu tion of verbal language indicates
that our iconic communication serves functions totally different
from those of language and, in-deed, performs function s which
verba l language is unsui ted to perfor m.
When boy says to girl, "I love you," he is using words to conve y
that which is more convinc ingly conve yed by his tone of voice and
his movements; and the girl, if she has any sense, will pay more
attention to those accompanying signs than to the words. There are
people—p rofes siona l actors, confidenc e tricksters, and others—who
are able to use kinesics and paralingui stic communication with a
degre e of volun tary control comparable to that volunt ary control
which we all think we have over the use of words . For these people
who can lie with kinesics, the special usefu lness of non-verba l
communication is reduced. It is a little more difficult for them to be
sincere and still more difficult for them to be believed to be sincere.
They are caught in a process of diminish ing return s such that, when
distrusted, they try to improve their skill in simulating paralingu istic
and kinesic sincer ity. But this is the very skill which led other s to
distrust them.
It seems that the discours e of nonverb al communication is
precisely conce rned with matters of relationship— love, hate,
respect, fear, dependenc y, etc.—between self and vis-à -vis or
between self and environment and that the nature of human society
is such that falsification of this discours e rapid ly beco mes
pathogen ic. From an adap tive poin t of view , it is therefore important
that this discou rse be carried on by techn iques which are relatively
unconsc ious and only imperfec tly subject to volunt ary control. In
the language of neuroph ysiolog y, the controls of this discou rse must
be placed in the brain caudad of the con trols of true language .
If this gen eral view of th e matter be cor rect, it must fol low that to
transla te kinesics or paralinguis tic messages into words is likely to
introduce gross falsification due not merely to the human propens ity
for trying to falsify statements about "feelings" and relationsh ip and
to the distortions which arise whenev er the product s of one system
of coding are dissected onto the premises of another, but especially
420

to the fact that all such translation must give to the more or less
uncon scious and involun tary iconic message the appe arance of
conscious intent.
As scientists, we are concerned to build a simulacru m of the
pheno menal universe in words. That is, our product is to be a verbal
transform of the pheno mena. It is necessar y, therefore, to examine
rather carefully the rules of this trans-formation and the differences
in coding between natural pheno mena, message pheno mena, and
words . I know that it is unusua l to presume a "coding " of nonliving
pheno mena and, to justify this phras e, I must expand somewhat on
the concep t of "redundanc y" as this word is used by the commu-
nications e nginee rs.
The engine ers and mathematicians have concentrated their
attention rigorous ly upon the internal structure of message material.
Typically, this m aterial consi sts of a sequence or col lection of even ts
or objects (commonly members of finite sets—phone mes and the
like). This sequence is differen tiated from irrelevan t even ts or
objects occurr ing in the same region of time-spac e by the
signal /nois e ratio and by other characteristics. The message materia l
is said to contain "redundanc y" if, when the sequence is received
with some items missing, the receive r is able to guess at the missing
items with better than rando m success. It has been pointed out that,
in fact, the term "redundanc y" so used becomes a synonym for "patterning ."1 It is importan t to note that this patterning of message
materia l alwa ys helps the receiver to differen tiate between signal
and noise. In fact, the regular ity called signal/noise ratio is really
only a special case of redundanc y. Camouflage (the opposi te of
communication) is achieved (1) by reducing the signa l/noise ratio,
(2) by breaking up the patterns and regula rities in the signal, or (3)
by introducing similar patterns into the no ise.
By confining their attention to the internal structure of the
message materia l, the engine ers believe that they can avoid the
complexit ies and difficulties introdu ced into communication theory
by the concep t of "meaning ." I would argue, howeve r, that the
concept "redundanc y" is at least a partial synonym of "meaning. " As
1 F. Attneave, Applications of Information Theory to Psychology , New York,
Henry Holt and Co., 1959.
421

I see it, if the receiv er can guess at missing parts of the message ,
then those parts which are received must, in fact, carry a meaning
which refers to the missing parts and is information about those
parts.
If now we turn away from the narrow universe of message
structu re and consider the outer world of natural phenomena, we
observ e at once that this outer world is similarly charac terized by
redundan cy, i.e., that when an observer perceiv es only certain parts
of a sequence or configur ation of pheno mena, he is in many cases
able to guess , with better than rando m success , at the parts which he
cannot immediately perceiv e. It is, indeed , a principa l goal of the
scientist to elucidate these redundanc ies or patternings of the phe-
nomenal wor ld.
If we now consid er that larger univers e of which these two
subunive rses are parts, i.e., the system: message plus external
pheno mena, we find that thi s larger system cont ains redundanc y of a
very speci al sort. The observer's ability to predict externa l
pheno mena is very much increased by his receipt of message
material. If I tell you that "it is raining" and you look out the
window , you will get less information from the perception of
raindrops than you would have got had you never received my
message . From my message you could have guessed that you would
see rain.
In sum, "redundancy" and "meaning" become synonymous
whenever both words are applied to the same universe of
discou rse. "Redundanc y" within the restricted universe of the
message sequence is not, of course, synonymous with "meaning" in
the wider unive rse that includes both message and ex ternal referent.
It will be noted that this way of thinking about communication
groups all methods of coding under the single rubric of part-for-
whole. The verbal message "It is raining" is to be seen as a part of a
larger unive rse within which that message creates redundan cy or
predictab ility. The "digital," the "analogic ," the "iconi c," the
"metaphor ic," and all other methods of coding are subsu med under
this single head ing. (What the grammarian s call "synecdoche" is the
metaphor ic use of the name of a part in place of the name of the
whole, as in the phras e "five head of cattle.")
422

This approach to the matter has certain advan tages : the analyst is
forced at all times to define the univers e of discourse within which
"redundanc y" or "meaning" is supposed to occur. He is forced to
examine the "logica l typing" of all message material. We shall see
that this broad view of the matter makes it easy to identify major
steps in the evolution of communication. Let us consid er the
scientist who is observing two animals in a physical environment.
The following co mponent s then m ust be conside red:
(1)The physical environment contains internal patterning or
redundancy, i.e., the perception of certain events or objects
makes other events or objects predictable for the animals and/or
for the observer.
(2)Sounds or other signals from one animal may con-tribute
redundancy to the system, environment plus signal; i.e., the signals
may be "about" the environment.
(3)The sequence of signals will certainly contain redundancy—
one signal from an animal making another signal from the same
animal more predictable.
(4)The signals may contribute redundancy to the universe; A's
signals plus B's signals, i.e., the signals may be about the
interaction of which they are component parts.
(5)If all rules or codes of animal communicat ion and
unde rstand ing were geno typically fixed, the list would end at this
point. But some animals are capab le of learning, e.g., the repetition
of sequences may lead to their becoming effective as patterns. In
logic, "ever y proposi tion propo ses its own truth," but in natural
history we deal always with a conver se of this genera lization . The
perceiv able event s which accompany a given percept propose that
that percept shall "mean" these events . By some such steps an
organis m may learn to use the information contained in patterned
sequences of external even ts. I can therefore predict with bet-ter
than rando m succe ss that in the unive rse, organism plus
environment, events will occu r to complete patterns or
configur ation s of learned adapta tion between organism and
environment.
(6)The behavioral "learning" which is usually studied in
psychological laboratories is of a different order. The redundancy
of that universe, which consists of the animal's actions plus
423

external events, is increased, from the animal's point of view,
when the animal regularly responds to certain events with certain
actions. Similarly, this universe gains redundancy when the
animal succeeds in producing those actions which function as
regular precursors (or causes) of specific external events.
(7)For every organism there are limitations and regularities
which define what will be learned and under what circumstances
this learning will occur. These regularities and patterns become
basic premises for the individual adaptation and social
organization of any species.
(8)Last but not least, there is the matter of phylogenetic
learning and phylogeny in general. There is redundancy in the
system, organism-plus-environment, such that from the
morphology and behavior of the organism a human observer can
guess with better than random success at the nature of the
environment. This "information" about the environment has
become lodged in the organism through a long phylogenetic
process, and its coding is of a very special kind. The observer who
would learn about the aquatic environment from the shape of a
shark must deduce the hydrodynamics from the adaptation which
copes with the water. The information contained in the phenotypic
shark is implicit in forms which are complementary to characteris-
tics of other parts of the universe, phenotype plus environment
whose redundancy is increased by the phenotype.
This very brief and incomplete surve y of some of the sorts of
redundan cy in biologica l systems and the universes of their
relevance indica tes that unde r the general rubric "part-for-whole" a
number of different sorts of relationsh ip between part and whole are
included. A listing of some of the characteristics of these formal
relations is in order. We consider some of the iconic cases:
(1)The events or objects which we here call the "part" or
"signal" may be real components of an existing sequence or
whole. A standing trunk of a tree indicates the probable presence
of invisible roots. A cloud may indicate the coming storm of
which it is a part. The bared fang of a dog may be part of a real
attack.
(2)The "part" may have only a conditional relationship to its
whole: the cloud may indicate that we shall get wet if we don't go
424

indoors; the bared fang may be the beginning of an attack which
will be completed unless certain conditions are met.
(3)The "part" may be completely split from the whole which is
its referent. The bared fang at the given instant may mention an
attack which, if and when it occurs, will include a new baring of
the fangs. The "part" has now became a true iconic signal.
(4)Once a true iconic signal has evolved—not necessarily
through steps 1, 2, or 3, above—a variety of other pathways of
evolution become possible:
(a)The "part" may become more or less digitalized, so that
magnitudes within it no longer refer to magnitudes within the
whole which is its referent but, for example, contribute to an
improvement of the signal/noise ratio.
(b)The "part" may take on special ritual or metaphoric
meanings in contexts where the original whole to which it once
referred is no longer relevant. The game of mutual mouthing
between mother dog and puppy which once followed her weaning
of the pup may become a ritual aggregation. The actions of
feeding a baby bird may become a ritual of courtship, etc.
Throughou t this series, whose branche s and varieties are here
only briefly indicated, it is notab le that animal communication is
confined to signa ls which are derived from actions of the animals
themselves , i.e., those which are parts of such action s. The externa l
universe is, as already noted, redundant in the sense that it is replete
with part-for-whole messages, and—pe rhaps for that reason—thi s
basic style of coding is characteristic of primitive animal communi-
cation . But in so far as animals can signal at all abou t the external
universe, they do so by means of actions which are parts of their
response to that universe. The jackdaws indicate to each other that
Lorenz is a "jackdaw-eater" not by simulating some part of the act
of eating jackdaws but by simulating part of their aggression vis-a-
vis such a creature. Occasionally actual pieces of the external
environment—scraps of potential nest-building material,
"trophies," and the like—are used for communication, and in these
cases again the messages usually contribute redundancy to the
universe message p lus the relationsh ip between t he organisms rather
than to the universe message plus externa l env ironment .
425

In terms of evolu tiona ry theory, it is not simple to ex-plain why
over and over again genot ypic controls have been evolv ed to
determine such iconic signaling. From the point of view of the
human observer such iconic signals are rather easy to interpr et, and
we might expec t iconic coding to be comparat ively easy for animals
to decod e—in so far as the animals must learn to do so. But the
geno me is presumed not capab le of learning in this sense , and we
might therefore expec t geno typicall y determined signal s to be
anicon ic or arbitrary rather than iconic.
Three possible explanations of the iconic nature of genotypic
signals can be offered:
(1)Even genotypically determined signals do not occur as
separate and isolated elements in the life of the phenotype but are
necessarily components in a complex matrix of behavior some, at
least, of which is learned. It is possible that the iconic coding of
genotypically determined signals renders these easy to assimilate
into this matrix. There may be an experiential "schoolmarm"
which acts selectively to favor those genotypic changes which
will give rise to iconic rather than arbitrary signaling.
(2)A signal of aggression which places the signaler in a
position of readiness to attack probably has more survival value
than would a more arbitrary signal.
(3)When the geno typically determined signal affects the
behavio r of another species—e .g., eye marks or postures which
have a warn ing effect, movements which facilitate camouflage or
apose matic mimicry—clear ly the signal must be iconic to the
perceptiv e system of that other species. Howeve r, an interesting
pheno menon arises in many instance s where what is achiev ed is a
seconda ry statistical iconicism. Labr oides dimidia tus, a small
Indo-Pa cific wrasse , which lives on the ectoparas ites of other
fishes, is strikingly colored and moves or "dance s" in a way which
is easily recogn ized. No doubt these characteristics attract other
fish and are par t of a signa ling s ystem which leads the othe r fish to
permit the approaches of the cleane r. But there is a mimic of this
species of Labr oides, a saber-toothed blenny (Aspidontus
426

taenia tus), whose similar color ing and movement permit the
mimic to approach— and b ite off pieces of the fins o f other fishes.2
Clearly the coloring and movements of the mimic are iconic and
"repres ent" the cleaner. But what of the coloring and movements of
the latter? All that is primaril y required is that the cleaner be
conspicuous or distinctive. It is not required that it represent
something else. But when we consider the statistical aspects of the
system, it beco mes clear that if the blenni es beco me too numerous,
the distinctive features of the wrasses will beco me iconic warn ings
and their hosts will avoid them. What is neces sary is that the signals
of the wrass e shall clearly and indubi tably represen t wrasse , i.e., the
signal s, though perhaps aniconi c in the first instan ce, must achieve
and maintain by multiple impact a sort of autoiconi cism. "When I
say it three times, it is true." But this necessity for autoicon icism
may also arise within the species. Genot ypic contro l of signal ing
ensures the neces sary repetitiveness (which might be only fortuitous
if the signals had to be learned) .
(4) There is a case for asserting that the geno typic determination
of adaptive characteristics is, in a special sense , more econo mical
than the achieve ment of similar characteristic by somatic change or
pheno typic learning . This matter has been argued elsewher e.3
Briefly it is asserted that the somatic adaptiv e flexib ility and/or
learning capac ity of any organis m is limited and that the demands
placed upon these capacities will be reduced by geno typic change in
any appropri ate direction. Such changes would therefore have
survival value because they set free precious adaptive or learning
capacity for other uses. This amounts to an argument for Baldwin
effects. An extention of this argument would suggest that the
iconic character of genotypically controlled signaling
characteristics may, in some cases, be explained by supposing that
these characteristics were once learned. (This hypothesis does not,
of course, imply any sort of Lamarckian inheritance. It is obvious
(1) that to fix the value of any variable in a homeostatic circuit by
2 J. E. Randall and H. S. Rand all, "Examples of Mimicry and Protective
Resemblance in Tropical Marine Fishes ," Bulletin of Marine Science of the Gulf
and Caribbean, 1960, 10: 444-80.3 G. Bateson, "The Role of Somatic Change in Evolution," Evolution, 1963,
17: 529-39.
427

such inheritance would soon gum up the homeostatic system of the
body, and (2) that no amount of modification of the dependent
variables in a homeostatic circuit will change the bias of the
circuit.)
(5) Last, it is unclear at what level genotypic determination of
behavior might act. It was suggested above that iconic codes are
easier for an organism to learn than more arbitrary codes. It is
possible that the genotypic contribution to such an organism might
take the form, not of fixing the given behavior, but rather of
making this behavior easier to learn—a change in specific learning
capacity rather than a change in genotypically determined
behavior. Such a contribution from the genotype would have
obvious advantages in that it would work along with ontogenetic
change instead of working possibly at cross-purposes with it.
To sum up the argument so far:
(1)But the evolution of aniconic verbal coding remains un-
explained. It is understandable that an early (in an evolutionary
sense) method of creating redundancy would be the use of iconic
part-for-whole coding. The external nonbiological universe
contains redundancy of this kind, and in evolving a code of
communication it is expectable that organisms would fall into the
same trick. We have noted that the "part" can be split from the
whole, so that a showing of the fangs can denote a possible but as
yet nonexistent fight. All this provides an explanatory background
for communication by means of "intention movements" and the
like.
(2)It it partly understandable that such tricks of coding by
iconic parts might become genotypically fixed.
(3)It has been suggested that the survival of such primitive (and
therefore involuntary) signalling in human communication about
personal relationship is explained by a need for honesty in such
matters.
We know from studies of apha sia, from Hocket t's enumeration at
this meeting of the characteristics of language and from elementar y
common sense that the component process es of creating and
unders tanding verbal communication are many and that language
fails when any one of thos e component proce sses is interrup ted. It is
possib le that each of these processe s shou ld be the focus of a
428

separa te study. Here , however , I shall consider only one aspect of
the m atter: the evo lution of simple i ndica tive assertion.
An interesting intermediate between the iconic coding of animals
and the verbal coding of human speech can be recogniz ed in human
dreaming and human myth. In psychoana lytic theory, the
product ions of dream proces s are said to be characterized by
"primary-process" thinking.4 Drea ms, whether verbal or not, are to
be conside red as metaphoric statements, i.e., the referen ts of dream
are relationships which the dreamer, consciou sly or uncon sciously,
perceiv es in his waking world . As in all metaphor , the relata remain
unmentioned and in their places appear other items such that the
relationsh ips between these substitute items shall be the same as
those b etween t he relata in the waking wor ld.
To identify the relata in the waking world to which the dream
refers would conver t the metaphor into a simile, and, in genera l,
dreams conta in no message materia l which overtly performs this
function . There is no signal in the dream which tells the dreamer that
this is metaphor or what the referent of the metaphor may be.
Similar ly, dream contains no tenses. Time is telescoped, and
repres entations of past event s in real or distorted forms may have
the presen t as their referent—or vice versa . The patterns of dream
are timeless.
In a theater, the audience is informed by the curtain and the
framing of the stage that the action on the stage is "only" a play.
From within that frame the producers and actors may attempt to
involve the audience in an illusion of reality as seemingl y direct as
the experience of dream. And, as in dream, the play has metaphor ic
reference to the out-side world. But in dream, unless the sleeper be
partly conscious of the fact of sleep , there is no curtain and no
framing of the action. The partial negative—"T his is only
metaphor"—is absen t.
I sugge st that this absence of metacommunicative frames and the
persistence in dream of pattern recogn ition are archaic
characteristics in an evolut ionar y sense. If this be correct, then an
unde rstand ing of dream should throw light both on how iconic
4 4 O. Fenichel , Psychoanalyti c Theory of Neu York, Norton, 1945.
429

communication operates among animals and on the mysteriou s
evolu tiona ry step from the iconic to the verba l.
Under the limitat ion imposed by the lack of a metaco mmu-
nicative frame, it is clearly impossib le for dream to make an
indicative statement, either positive or negat ive. As there can be no
frame which labels the content as "metaphor ic," so there can be no
frame to label the content as "literal." Drea m can imagine rain or
drought , but it can never assert "It is raining" or "It is not raining ."
Therefo re, as we have seen, the usefuln ess in imagining "rain" or
"drough t" is limited to their metaphor ic aspects.
Drea m can propose the applicability of pattern. It can never
assert or deny this applicability. Still less can it make an indicative
statement abou t any identified referen t, since no referent is
identified.
The pattern is the thing.
These characteristics of dream may be arch aic, but it is importan t
to remember that they are not obsol ete: that, as kinesic and
paralingu istic communication has been elabora ted into dance ,
music, and poetry, so also the logic of dream has been elabo rated
into theater and art. Still more astoni shing is that world of rigorou s
fantasy which we call mathematics, a world foreve r isolated by its
axioms and definitions from the possibility of making an indicative
statement abou t the "real" world. Only if a straight line is the
shortest distance between two points is the theore m of Pythagoras
asserted .
The banker manipulat es numerals accord ing to rules sup-plied by
the mathematician. These numerals are the names of numbers, and
the nu mbers are so mehow e mbodied in ( real or fi ctitious) dol lars. To
remember what he is doing , the bank er marks his numerals with
labels, such as the dollar sign, but these are nonmathematical and
no computer needs them. In the strictly mathematical procedure, as
in the process of dream, the pattern of relationships controls all
operations, but the relata are unidentified.
We return now to the contrast between the iconic method of
creating redund ancy in the universe, organis m plus other organism,
by the emission of parts of interactive pat-terns and the linguis tic
devic e of naming the relata. We noted above that the human
communication which creates redundan cy in the relationship s
430

between person s is still preponde rantly iconic and is achiev ed by
means of kinesics, paralingui stics, intention movements, actions,
and the like. It is in dealing with the unive rse, message plus
environment, that the evolution of verbal language has made the
greatest strides.
In animal discours e, redundanc y is introduced into this unive rse
by signa ls which are iconic parts of the signal er's probab le response .
The environmental items may serve an ostensive function but
cannot, in gene ral, be mentioned . Similarly, in iconic
communica tion about relation ship, the relata—the organisms
themselves—do not have to be identified because the subjec t of any
predica te in this iconic discourse is the emitter of the signal, who is
always ostensiv ely presen t.
It appears then that at least two steps were necessar y to get from
the iconic use of parts of patterns of own behavior to the naming of
entities in the extern al environment: there was both a chang e in
coding and a change in the centering of the subjec t-predica te frame.
To attempt to reconstruct these steps can only be speculative,
but some remarks may be offered:
(1)Imitation of environmental phenomena makes it possible. to
shift the subject-predicate frame from the self to some
environmental entity while still retaining the iconic code.
(2)A similar shifting of the subject-predicate frame from self to
other is latent in those interactions between animals in which A
proposes a pattern of interaction and B negates this with an iconic
or ostensive "don't." The subject of B's message here verbalized as
"don't" is A.
(3)It is possible that the paradigms of interaction which are
basic to iconic signaling about relationship could serve as
evolutionary models for the paradigms of verbal grammar. We
should not, I suggest, think of the earliest rudiments of verbal
communication as resembling what a man does with only a few
words of a foreign language and no knowledge of its grammar and
syntax. Surely, at all stages of the evolution of language, the
communication of our ancestors was structured and formed–
complete in itself, not made of broken pieces. The antecedents of
grammar must surely be as old or older than the antecedents of
words.
431

(4)For actions of the self, iconic abbreviations are readily
available, and these control the vis-à-vis by implicit reference to
interactional paradigms. But all such communication is necessarily
positive. To show the fangs is to mention combat, and to mention
combat is to propose it. There can be no simple iconic
representation of a negative: no simple way for an animal to say "I
will not bite you." It is easy, however, to imagine ways of
communicating negative commands if (and only if) the other
organism will first propose the pat-tern of action which is to be
forbidden. By threat, by inappropriate response and so on, "don't"
can be communicated. A pattern of interaction, offered by one
organism, is negated by the other, who disrupts the proposed
paradigm.
But "don't " is very different from "not." Commonly, the
important message "I will not bite you" is genera ted as an
agreement between two organisms following real or ritual combat.
That is, the opposi te of the final message is worked through to reach
a reduct io ad absur dum which can then be the basis of mutual
peace , hierarch ic precedence , or sexual relations . Many of the
curious interac tions of animals, called "play," which resemble (but
are not) combat are probably the testing and reaffirmation of such
negat ive agree ment.
But these are cumbersome and awkward methods of achieving
the negative.
(5)It was sugges ted above that the paradigms of verbal grammar
might somehow be derived from the paradig ms of interaction. We,
therefor e, look for the evolu tiona ry roots of the simple negative
among the paradigms of interaction . The matter, howeve r, is not
simple. What is known to occur at the animal level is the
simultaneous present ation of contradictory signals—pos tures which
mention both aggres sion and flight, and the like. These ambiguit ies
are, howeve r, quite different from the pheno menon familia r among
humans where the friendl iness of a man's words m ay be con tradicted
by the tension or aggres sivene ss of his voice or posture. The man is
engaging in a sort of deceit, an altoge ther more complex
achievement, while the ambivalent animal is offering positive
alterna tives. From neither of these patterns is it easy to derive a
simple "not."
432

(6)From these considerations it appears likely that the
evolution of the simple negative arose by introjection or imitation
of the vis-à-vis, so that "not" was somehow derived from "don't."
(7)This still leaves unexplained the shift from communication
about interaction patterns to communication about things and
other components of the external world. This is the shift which
determines that language would never make obsolete the iconic
communication about the contingency patterns of personal
relationship.
Further than that we cannot at present go. It is even possible
that the evolution of verbal naming preceded the evolution of the
simple negative. It is, however, important to note that evolution of
a simple negative would be a decisive step toward language as we
know it. This step would immediately endow the signals—be they
verbal or iconic–with a degree of separateness from their
referents, which would justify us in referring to the signals as
"names." The same step would make possible the use of negative
aspects of classification: those items which are not members of an
identified class would become identifiable as nonmembers. And,
lastly, simple affirmative indicative statements would become
possible.
433

Conscious Purpose versu s Nature*
Our civilization, which is on the block here for invest igation and
evalu ation , has its roots in three main ancient civilizations : the
Roman, the Hebrew and the Greek ; and it would seem that many of
our problems are related to the fact that we have an imperia list
civilization leavened or yeasted by a downtrodd en, exploited colon y
in Pales tine. In this confe rence , we a re again going to be fighting ou t
the conflict between t he Ro mans and the Palestinians .
You will remember tha t St. Paul boasted , "I was born free ." What
he meant was that he was born Roman, and that this had certain
legal advantage s.
We can engage in that old battle either by backing the
downtrodden o r by backing t he imperialists. If y ou are going t o figh t
that battle, you have to take sides in it. It's that simple.
On the other hand , of course, St. Paul's ambition , and the
ambition of the downtrodden , is alwa ys to get on the side of the
imperialists— to become middle-clas s imperia lists themselves— and
it is doub tful whether creating more members of the civilization
which we are here criticizing is a solution to the prob lem.
There is, therefore, anothe r more abstract proble m. We need to
unders tand the patho logies and peculiarities of the whole Romano-
Pales tinian system. It is this that I am interested in talking about. I
do not care, here, about defending the Romans or defending the
Pales tinians—th e uppe r dogs or the underdogs . I want to consider
the dynamics of the whol e traditional patholog y in which we are
caught , and in which we shall remain as long as we continue to
struggle within that old conflict. We just go round and round in
terms of the old premises.
Fortuna tely our civilization has a third root—in Greec e. Of
cours e Greece got caught up in a rather similar mess, but still there
was a lot of clean, cool thinking of a quite surpr ising kind which
was d ifferent.
* This lecture was given in August, 1968, to the London Conference on the
Dialectics of Liberation , and is here reprinted from Dialectics of Liberation by
permission of the publisher , Penguin Books Inc.
434

Let me approach the bigge r problem historically. From St.
Thomas Aquinas to the eighte enth century in Catholic count ries, and
to the Refo rmation among Protestan ts (be-cause we threw out a lot
of Greek sophis tication with the Reformation ), the structu re of our
religion was Greek . In mid-eighte enth century the biolog ical world
looked like this: there was a supreme mind at the top of the ladde r,
which was the basic explanation of everything downwards from that
—the supre me mind being, in Chris tianity, God; and having various
attribut es at various philosoph ic stages. The ladde r of explanation
went downwards dedu ctively from the Supreme to man to the apes,
and so on, down to the infusoria.
This hierarch y was a set of deduc tive steps from the most perfect
to the most crude or simple. And it was rigid. It was assumed that
every species was un changing .
Lamarck, probab ly the greate st biologis t in history, turned that
ladder of explana tion upside down. He was the man who said it
starts with the infuso ria and that there were chang es leading up to
man. His turning the taxono my upside down is one of the most
astoni shing feats that has ever occu rred. It was the equivalen t in
biolog y of the Copern ican revolution in astrono my.
The logica l outco me of turning the taxono my upside down was
that the study of evolut ion m ight p rovide an exp lanation of mind.
Up to Lamarck, mind was the explanation of the biological
world. But, hey presto, the question now arose : Is the biologica l
world the explanat ion of mind? That which was the expl anation now
became that which was to be explained. About three quart ers of
Lamarck's Philo sophie Zoolog ique (1809) is an attempt, very crude,
to build a comparative psycholog y. He achieved and formulated a
number of very modern ideas: that you canno t attribu te to any
creatu re psychologi cal capac ities for which it has no organs; that
mental process must always have physical represen tation; and that
the complexi ty of the nervous system is related to the complexi ty of
mind.
There the matter rested for 150 years, mainly because
evolutionar y theory was taken over, not by a Catho lic heresy but by
a Protestan t heresy, in the mid-nineteenth centu ry. Darwin's
opponen ts, you may remember, were not Aristotle and Aquinas , who
had some soph istication, but funda mentalist Christians whose
435

sophis tication stopped with the first chap ter of Genes is. The
quest ion of the nature of mind was something which the ninete enth-
centu ry evolu tionists tried to exclude from their theories, and the
matter did not come up again for serious considera tion until after
World War II. (I am doing some injustice to some heretics along the
road, notably to Sa muel But ler—and others.)
In World War II it was discov ered what sort of complexi ty entails
mind. And, since that discover y, we know that: wherever in the
Univers e we encounte r that sort of complexi ty, we are dealing with
mental pheno mena. It's as materialistic as that.
Let me try to describe for you that order of complexi ty, which is
in some degree a technica l matter. Russ el Wallac e sent a famous
essay to Darwin from Indones ia. In it he announced his discover y of
natural selection, which coincided with Darwin's . Part of his
description o f the struggl e for existence i s interesting:
The action of this principle [the struggle for existence] is
exactly like that of the steam engine, which checks and corrects
any irregularities almost before they become evident; and in like
manner no unbalanced deficiency in the animal kingdom can ever
reach any conspicuous magnitude, because it would make itself
felt at the very first step, by rendering existence difficult and
extinction almost sure to follow.
The steam engine with a governor is simply a circular train of
causal events, with somewhere a link in that chain such that the
more of something, the less of the next thing in the circuit. The
wider the balls of the governor diverge, the less the fuel supply. If
causal chains with that general characteristic are provided with
energy, the result will be (if you are lucky and things balance out)
a self-corrective system.
Wallace, in fact, proposed the first cybernetic model.
Nowada ys cybernetics deals with much more complex systems
of this gene ral kind; and we know that when we talk abou t the
proce sses of civilization, or evaluate human behavior, human
organization, or any biolog ical system, we are conce rned with self-
corrective systems. Basic ally these systems are always conservative
of something . As in the engine with a gove rnor, the fuel supp ly is
changed to conserv e—to keep constan t—the speed of the flywheel,
so alwa ys in such systems changes occu r to conserve the truth of
436

some descr iptive statement, some component of the status quo.
Wallace saw the matter correctly, and natural selection acts
primarily to keep the species unva rying; but it may act at higher
levels to keep constant that complex variable which we call
"surviva l."
Dr. Laing noted that the obvious can be very difficult for people
to see. That is because people are self-co rrective systems. They are
self-corr ective against disturb ance, and if the obvious is not of a
kind that they can easily assimilat e without internal disturbance ,
their self-cor rective m echanisms work t o side track i t, to hide it , even
to the extent of shut ting the eyes if necess ary, or shut ting off various
parts of the process of percep tion. Disturb ing information can be
framed like a pearl so that it doesn' t make a nuisance of itself; and
this will be done, acco rding to the unders tanding of the system itself
of what would be a nuisanc e. This too—the premise regarding what
would cause disturban ce—is something which is learned and then
becomes perpetuated o r con -served .
At this conferenc e, funda mentally, we deal with three of these
enormously complex systems or arrange ments of conservative
loops. One is the human individua l. Its physiolog y and neuro logy
conserve body temperatu re, blood chemistry, the leng th and size and
shape of organs during growth and embryology, and all the rest of
the body's characteristics. This is a system which conserves
descriptive statements about the human being, body or soul. For
the same is true of the psychology of the individual, where
learning occurs to conserve the opinions and components of the
status quo.
Second, we deal with the society in which that individual lives
—and that society is again a system of the same general kind.
And third, we deal with the ecosystem, the natural biological
surroundings of these human animals.
Let me start from the natural ecosystems around man. An
English oak wood, or a tropical forest, or a piece of desert, is a
community of creatures. In the oak wood perhaps 1000 species,
perhaps more; in the tropical forest perhaps ten times that number
of species live together.
I may say that very few of you here have ever seen such an
undisturbed system; there are not many of them left; they've
437

mostly been messed up by Homo sapiens who either exterminated
some species or introduced others which be-came weeds and pests,
or altered the water supply, etc., etc. We are rapidly, of course,
destroying all the natural systems in the world, the balanced
natural systems. We simply make them unbalanced—but still
natural.
Be that as it may, those creatures and plants live together in' a
combination of competition and mutual dependency, and it is that
combination that is the important thing to consider. Every species
has a primary Malthusian capacity. Any species that does not,
potentially, produce more young than the number of the population
of the parental generation is out. They're doomed. It is absolutely
necessary for every species and for every such system that its
components have a potential positive gain in the population curve.
But, if every species has potential gain, it is then quite a trick to
achieve equilibrium. All sorts of interactive balances and
dependencies come into play, and it is these processes that have the
sort of circuit structure that I have mentioned.
The Malthusian curve is exponential. It is the curve of
population growth and it is not inappropriate to call this the
population explosion.
You may regret that organis ms have this explosiv e characteristic,
but you may as well settle for it. The creatures that don 't are out.
On the other hand , in a balanced ecolog ical system whose
underp innings are of this nature, it is very clear that any monkeying
with the system is likely to disrup t the equilibrium. Then the
exponent ial curves will start to appear . Some plant will become a
weed, some creatures will be exterminated, and the system as a
balanced system is likely to fall to pieces.
What is true of the species that live together in a wood is also
true of the groupings and sorts of peopl e in a society, who are
similar ly in an uneas y balance of dependenc y and competit ion. And
the same truth holds right inside you, where there is an uneasy
physiologica l competit ion and mutual dependenc y among the
organs, tissues, cells, and so on. Without this competition and
dependenc y you would not be, because you canno t do without any
of the competing organs and parts. If any of the parts did not have
the expansiv e characteristics they would go out, and you would go
438

out, too. So tha t even i n the bod y you have a l iability. With imprope r
disturbance of the system, the exponent ial curves app ear.
In a society, the same is true.
I think you have to assume that all important physiological or
social change is in some degree a slipping of the system at some
point along an exponen tial curve . The slippage may not go far, or it
may go to disaster. But in princip le if, say, you kill off the thrushes
in a wood, certain components of the balance will run along
expon ential curves to a new stopp ing p lace.
In such slippage there is alwa ys dang er—the possibility that
some variable, e.g., popula tion density, may reach such a value that
further slippag e is controlled by factors which are inherently
harmful. If, for example, popul ation is finally contro lled by
available food supply, the surviving individuals will be half starved
and the food suppl y overgrazed, usually to a poin t of no return.
Now let me begin to talk abou t the individu al organism. This
entity is similar to the oak wood and its contro ls are represented in
the total mind, which is perhaps only a reflection of the total body.
But the system is segmented in various ways, so that the effects of
something in your food life, shall we say, do not totally alter your
sex life, and things in your sex life do not totally change your
kinesi c life, and so on. There is a certain amount of
compartmentalization, which is no doubt a necessary economy.
There is one compartmentalization which is in many ways
mysterious but certainly of crucial importance in man's life. I refer
to the "semipermeable" linkage between consciousness and the re-
mainder of the total mind. A certain limited amount of in-
formation about what's happening in this larger part of the mind
seems to be relayed to what we may call the screen of
consciousness. But what gets to consciousness is selected; it is a
systematic (not random) sampling of the rest.
Of course, the whole of the mind could not be reported in a part
of the mind. This follows logically from the relationship between
part and whole. The television screen does not give you total
coverage or report of the events which occur in the whole
television process; and this not merely because the viewers would
not be interested in such a re-port, but because to report on any
extra part of the total process would require extra circuitry.. But to
439

report on the events in this extra circuitry would require a still
further addition of more circuitry, and so on. Each additional step
toward increased consciousness will take the system farther from
total consciousness. To add a report on events in a given part of the
machine will actually decrease the percentage of total events
reported.
We therefore have to settle for very limited consciousness, and
the question arises: How is the selecting done? On what principles
does your mind select that which "you" will be aware of? And,
while not much is known of these principles, something is known,
though the principles at work are often not themselves accessible
to consciousness. First of all, much of the input is consciously
scanned, but only after it has been processed by the totally
unconscious process of perception. The sensory events are
packaged into images and these images are then "conscious."
I, the conscious I, see an unconsciously edited version of a
small percentage of what affects my retina. I am guided in my
perception by purposes. I see who is attending, who is not, who is
understanding, who is not, or at least I get a myth about this
subject, which may be quite correct. I am interested in getting that
myth as I talk. It is relevant to my purposes that you hear me.
What happens to the picture of a cyberneti c system—an oak
wood or an organism—when that picture is selectively drawn to
answer only questions of purpos e?
Consider the state of medicine today. It's called medical
science. What happens is that doctors think it would be nice to get
rid of polio, or typhoid, or cancer. So they devote re-search money
and effort to focusing on these "problems," or purposes. At a
certain point Dr. Salk and others "solve" the problem of polio.
They discover a solution of bugs which you can give to children so
that they don't get polio. This is the solution to the problem of
polio. At this point, they stop putting large quantities of effort and
money into the problem of polio and go on to the problem of
cancer, or whatever it may be.
Medicine ends up, therefore, as a total science, whose structure
is essentially that of a bag of tricks. Within this science there is
extraordinarily little knowledge of the sort of things I'm talking
about; that is, of the body as a systemically cybernetically
440

organized self-corrective system. Its internal interdependencies are
minimally understood. What has happened is that purpose has
determined what will come under the inspection or consciousness
of medical science.
If you allow purpose to organize that which comes under your
conscious inspection, what you will get is a bag of tricks—some of
them very valuable tricks. It is an extraordinary achievement that
these tricks have been discovered; all that I don't argue. But still
we do not know two-penn'orth, really, about the total network
system. Cannon wrote a book on The Wisdom of the Body, but
nobody has written a book on the wisdom of medical science,
because wisdom is precisely the thing which it lacks. Wisdom I
take to be the knowledge of the larger interactive system—that
system which, if disturbed, is likely to generate exponential curves
of change.
Consc iousness opera tes in the same way as medicine in its
sampling of the events and processe s of the body and of what goes
on in the total mind. It is organiz ed in terms of purpos e. It is a short –
cut device to enab le you to get quick ly at what you want ; not to act
with maximum wisdo m in order to live, but to follow the shortest
logica l or causal path to get what you next want , which may be
dinner ; it may be a Beethoven sonata; it may be sex. Above all, it
may be money or power .
But you may say: "Yes, but we have lived that way for a million
years." Consc iousnes s and purpose have been characteristic of man
for at least a million years, and may have been with us a great deal
longer than that. I am not prepa red to say that dogs and cats are not
conscious , still less that porpois es are no t conscious .
So you may say: "Why worr y abou t that?"
But what worries me is the addition of modern techno logy to the
old system. Today the purposes of consciou sness are implemented
by more and more effective machiner y, transportation systems,
airplanes , weaponr y, medicine , pesticides, and so forth. Conscious
purpose is now empowered to upset the balan ces of the body, of
society, and of the biologic al world around us. A patholog y—a loss
of balance—i s threatened .
I think that m uch of what brings us here toda y is basically related
to the thoughts that I have been putting b efore you. On the one hand,
441

we have the systemic nature of the individu al human being , the
systemic nature of the culture in which he lives, and the systemic
nature of the biologi cal, ecologic al system around him; and, on the
other hand, the curious twist in the systemic nature of the individua l
man wher eby consciou sness is, almost of necessity, blinded to the
systemic natur e of the man himself. Purposive consciou sness pulls
out, from the total mind, sequences which do not have the loop
structu re which is charac teristic of the whole systemic structure. If
you follow the "common-sense" dictates of consc iousness you
beco me, effectively, greedy and unwise—again I use "wisdo m" as a
word for recogni tion of and guidance by a knowl edge of the total
systemic c reatu re.
Lack of systemic wisdom is alwa ys punished . We may say that
the biologica l systems-the individu al, the culture, and the ecolog y—
are partly living sustaine rs of their component cells or organis ms.
But the systems are nonetheless punish ing of any specie s unwise
enough to quarrel with its ecolog y. Call the systemic forces "God" if
you will.
Let me offer you a myth.
There was once a Garden. It contained many hundred s of species
—probab ly in the subtropic s—liv ing in great fertility and balance ,
with plenty of humus, and so on. In that garden, there were two
anthropoids who were more intelligent than t he other animals.
On one of the trees there was a fruit, very high up, which the two
apes were unable to reach . So they began to think. That was the
mistake. They began to think purpos ively.
By and by, the he ape, whose name was Adam, went and got an
empty box and put it under the tree and stepped on it, but he found
he still couldn' t reach the fruit. So he got another box and put it on
top of the first. Then he climbed up on the two boxe s and finally he
got that app le.
Adam and Eve then became almost drunk with excitement. This
was t he wa y to do things. Make a plan, ABC and you get D.
They then began t o spec ialize in doing th ings th e planned wa y. In
effect, they cast out from the Garden the concep t of their own total
systemic na ture and o f its total systemic nature.
. After they had cast God out of the Gard en, they really went to
work on this purpo sive busine ss, and pretty soon the topso il
442

disappea red. After that, sever al species of plants became "weeds"
and some of the animals became "pests"; and Adam found that
gardening was much harder work. He had to get his bread by the
swea t of his brow and he said, "It's a vengeful God. I should never
have eaten that apple."
Moreove r, there occurr ed a qualitative change in the relationship
between Adam and Eve, after they had discard ed God from the
Garden . Eve began to resent the busines s of sex and reproduc tion.
Whenever these rather basic phenomena intruded upon her now
purposive way of living , she was reminded of the larger life which
had been kicked out of the Garden. So Eve began to resen t sex and
reproduc tion, and when it came to parturition she found this proce ss
very painful . She said this, too, was due to the vengefu l nature of
God. She even heard a Voice say "In pain shalt thou bring forth" and
"Thy desire sh all be unto t hy husb and, and he shall rule over thee."
The biblical version of this story, from which I have borrowed
extensively, does not explain the extraordin ary perversion of values ,
whereb y the woman's capacity for love comes to seem a curse
inflicted by the deity.
Be that as it may. Adam went on pursu ing his purposes and
finally inven ted the free-ente rprise system. Eve was not, for a long
time, allowed to participa te in this because she was a woman. But
she joined a bridge club and t here found an ou tlet for her hate.
In the nex t gen eration, they again had troub le with l ove. Cain , the
invento r and innova tor, was told by God that "His [Abel 's] desire
shall be unto t hee and t hou sh alt rule ov er him." So he killed Abel.
A parable, of course, is not data abou t human behavior . It is only
an explanato ry device. But I have built into it a pheno menon which
seems to be almost unive rsal when man commits the error of
purposive thinking and disregards the systemic nature of the world
with which he must deal. This pheno menon is called by the
psychologis ts "proje ction ." The man, after all, has acted according
to what he thought was co mmon sense and now he finds himself i n a
mess. He does not quite know what caused the mess and he feels
that what has happened is somehow unfair. He still does not see
him-self as par t of th e system in which the mess exists, and he ei ther
blames the rest of the system or he blames himself. In my parabl e
443

Adam combines two sorts of nons ense: the notion "I have sinned"
and t he no tion "God is venge ful."
If you look at the real situations in our world where the
systemic nature of the world has been ignored in favor of purpose
or common sense, you will find a rather similar reaction. President
Johnson is, no doubt, fully aware that he has a mess on his hands,
not only in Vietnam but in other parts of the national and
international ecosystems; and I am sure that from where he sits it
appears that he followed his purposes with common sense and that
the mess must be due either- to the wickedness of others or to his
own sin or to some combination of these, according to his
temperament.
And the te rrible thing abou t such si tuations is tha t inevitably they
shorten the time span of all planning . Emergenc y is presen t or only
just around the corner; and long-term wisdom must therefore be
sacrificed to expedienc y, even though there is a dim awareness that
expedi ency will never give a long-t erm solution .
Morev er, since we are engag ed in diagnosing the machine ry of
our own society, let me add one poin t: our poli ticians—bo th those in
a state of power and t hose i n a state of protest or hung er for power—
are alike utterly ignorant of the matters which I have been
discu ssing . You can search the Congre ssiona l Record for speech es
which show awarenes s that the problems of govern ment are
biologica l proble ms, and you will find very, very few that apply
biologica l insigh t. Extrao rdina ry!
In gene ral, govern mental decisions are made by person s who are
as igno rant o f thes e matters as pig eons. L ike the f amous Dr. Skinne r,
in The Way of All Flesh, they "combine the wisdo m of the dove with
the harmlessness of the serpen t."
But we are m et here not on ly for d iagnosi s of so me of the wor ld's
ills but also to think about remedies . I have al-ready sugges ted that
no simple remedy to what I called the Romano-Pale stinian problem
can be achieved by backing the Romans against the Palestinians or
vice versa. The proble m is systemic and the solution must surely
depend upon r ealizing this fact.
First, there is humility, and I propose this not as a moral
principle, distasteful to a large number of peop le, but simply as an
item of a scientific philosoph y. In the period of the Indust rial
444

Revolu tion, perhap s the most importan t disas ter was the enormous
increa se of scientific arrog ance. We had discovered how to make
trains and other machines . We knew how to put one box on top of
the other to get that apple, and Occiden tal man saw himself as an
autocra t with complete power over a universe which was made of
physics and chemistr y. And the biologic al pheno mena were in the
end to be controlled like processes in a test tube. Evolu tion was the
history of how organisms learned more tricks for controlling the
environment; and man had better tricks than any other creature.
But that arrogan t scientific philosoph y is now obso lete, and in its
place there is the discove ry that man is only a part of larger systems
and that the part can never control the whole.
Goebbe ls though t that he could control publi c opinion in
Germany with a vast communicat ion system, and our own public
relations men are perhaps liable to similar delusions . But in fact the
would -be controller must alwa ys have his spies out to tell him what
the peopl e are saying about his propag anda. He is therefore in the
position of being responsive to what they are saying. Therefore he
cannot have a simple lineal control. We do not live in the sort of
universe in which simple lineal control is possible. Life is not like
that.
Similarly, in the field of psychiatry, the family is a cybernetic
system of the sort which I am discussing and usually when
systemic pathology occurs, the members blame each other, or
sometimes themselves. But the truth of the matter is that both these
alternatives are fundamentally arrogant. Either alternative assumes
that the individual human being has total power over the system of
which he or she is a part.
Even within the individu al human being , control is limited. We
can in some degre e set ourselv es to learn even such abstract
characteristics as arroganc e or humility, but we are not by any
means the captains of our s ouls.
It is, howeve r, possible that the remedy for ills of conscious
purpose lies with the individua l. There is what Fr eud ca lled the royal
road to the unconsc ious. He was referring to dreams, but I think we
shou ld lump together dreams and the creativity of art, or the
percept ion of art, and poetry and such things . And I would include
with these the best of religion. These are all activities in which the
445

whole individual is involved. The artist may have a conscious
purpose to sell his picture, even perhaps a consc ious purpose to
make it. But in the making he must necessa rily relax that arrogance
in favor of a creative expe rience in which his consc ious mind plays
only a small pa rt.
We might say that in creative art man must exper ience himself—
his total self—as a cybernet ic model.
It is characteristic of the 1960 s that a large number of peop le are
looking to the psychedel ic drugs for some sort of wisdo m or some
sort of enlargement of consciou sness , and I think this symptom of
our epoch probabl y arises as an attempt to compensate for our
exces sive purpo sivene ss. But I am not sure that wisdom can be got
that way. What is required is not simply a relaxa tion of
consc iousness to let the uncon scious material gush out. To do this is
merely to exchange one partial view of the self for the other partial
view. I suspe ct that what is needed is the synthesis of the two views
and t his is more difficult.
My own slight exper ience of LSD led me to believe that
Prospe ro was wrong when he said, "We are such stuff as dreams are
made on." It seemed to me that pure dream was, like pure purpose ,
rather trivial. It was no t the stuff of which we are made, but only bits
and pieces of that stuff. Our consc ious purposes , similarly, are only
bits and pieces.
The systemic view i s something e lse again.
446

Effects of Conscious P urpose on Human
Adaptation*
"Prog ress," "learning, " "evolution," the similar ities and dif-
ference s between phylogenet ic and cultural evolution, and so on,
have been subjects for discussion for many years. These matters
become newl y invest igabl e in the light of cybernetics and systems
theory.
In this Wenner -Gren conferen ce, a particula r aspect of this wide
subject m atter wi ll be e xamined, n amely the ro le of con sciousne ss in
the ongo ing p rocess of hu man adapt ation .
Three cybernet ic or homeostatic systems will be considered: the
individu al human organism, the human society, and the larger
ecosystem. Consciousn ess will be conside red as an important
component in the coupl ing of these s ystems.
A question of great scientific interest and perhap s grave
importance is whethe r the information proce ssed through
consciousne ss is adequate and appropr iate for the task of human
adaptation. It may well be that consciou sness contains systematic
distortion s of view which , when imple mented by modern
technolog y, become destructive of the balanc es between man, his
society and his ecosystem.
To introduce this question the following considera tions are
offered :
(1)All biolog ical and evolv ing systems (i.e., indiv idual
organis ms, animal and human societies, ecosystems, and the like)
consist of complex cybernetic networks , and all such systems share
certain formal chara cteristics. Each system contains subsystems
which are potentially regener ative, i.e., which would go into
* This essay was prepared as the author's position paper for Wenner-Gren
Foundation Conference on "Effects of Conscious Purpose on Human Adaptation ."
The author was chairman of this conference, which was held in Burg Wartenstein ,
Austria , July 17-24, 1968 . The proceedings of the conference as a whole are to be
published by Knopf & Co. under the title Our Own Metaphor, edited by Mary
Catherine Bateson .
447

exponent ial "runawa y" if uncorre cted. (Exa mples of such
regen erative components are Malthusi an characteristics of
popula tion, schismogenic changes of person al interaction ,
armaments races, etc.) The regener ative potentialities of such
subsystems are typical ly kept in che ck by variou s sorts of gov erning
loops to achieve "steady state." Such systems are "conse rvative" in
the sense that they tend to conserve the truth of proposi tions about
the values of their component variables— especi ally they conse rve
the values of those variables which otherwise would show
exponent ial change . Such systems are homeostatic, i.e., the effects
of small changes of input will be nega ted and the steady state
maintained b y reversible adjustment.
(2)But "plus c'est la meme chose , plus ça change." This convers e
of the French aphori sm seems to be the more exact description of
biologica l and ecologica l systems. A constan cy of some variable is
maintained by chang ing other variables . This is characteristic of the
engine with a governor : the constancy of rate of rotation is
maintained by altering the fuel supp ly. Muta tis mutand is, the same
logic under lies e volutionary progres s: tho se mutational ch anges wil l
be perpetuated which contribut e to the constan cy of that complex
variable which we call "surviv al." The same logic also applies to
learning , social chang e, etc . The ongoing tru th of certain descriptive
propos itions is maintained by altering other propositions.
(3)In systems containing many interconnec ted ho meostatic loops,
the changes brough t abou t by an external impact may slowl y spread
through the system. To maintain a given variable (V1) at a given
value , the values of V2, V3, etc., undergo change. But V2 and V3
may themselves be subject to homeosta tic control or may be linked
to var iables (V4, V5, etc.) which are subjec t to control. This second-
order homeostasis may lead to change in V6, V7, etc. And so on .
(4)This pheno menon of spreading change is in the wides t sense
a sort of learning. Acclimation and addiction are special cases of
this process. Over time, the system becomes de-pendent upon the
continued presence of that original external impact whose
immediate effects were neutralized by the first order homeostasis.
448

Example: under the impact of Prohib ition, the American social
system reacted homeosta tically to maintain the constancy of the
supp ly of alcohol . A new profession, the bootlegger , was gene rated.
To control this profes sion, chang es occurred in the police system.
When the question of repea l was raised, it was expectab le that
certainly the bootleggers and possib ly the police would be in favor
of maintaining Proh ibition.
(5)In this ultimate sense, all biological change is conservative
and all learning is aversive. The rat, who is "re-warded" with food,
accepts that reward to neutralize the changes which hunger is
beginning to induce; and the conventionally drawn distinction
between "reward" and "punishment" depends upon a more or less
arbitrary line which we draw to delimit that subsystem which we
call the "individual." We call an external event "reward" if its
occurrence corrects an "internal" change which would be
punishing. And so on.
(6)Consciousness and the "self" are closely related ideas, but
the ideas (possibly related to genotypically determined premises
of territory) are crystallized by that more or less arbitrary line
which delimits the individual and defines a logical difference
between "reward" and "punishment." When we view the
individual as a servosystem coupled with its environment, or as a
part of the larger system which is individual + environment, the
whole appearance of adaptation and purpose changes.
(7)In extreme cases, change will precipitate or permit some
runawa y or slippage along the poten tially exponent ial curves of the
unde rlying regene rative circuits. This may occur without total
destruc tion of the system. The slippage along expon ential curves
will, of cours e, alwa ys be limited, in extreme cases , by breakdown
of the system. Shor t of this disaster, other factors may limit the
slippag e. It is importan t, howeve r, to note that there is a dange r of
reaching l evels at whi ch the l imit is imposed b y factors whi ch are i n
them-selv es deleterious . Wynne-Ed ward s has pointed out—what
every farmer knows— that a popu lation of healthy individua ls
cannot be directly limited by the available food suppl y. If starvation
449

is the method of getting rid of the excess popul ation, then the
survivors will suffer if not death at least severe dietary deficiency,
while t he food supp ly itself wi ll be r educed , perhaps i rrevers ibly, by
overgrazing . In princip le, the homeostatic controls of biologica l
systems must be activated by variables which are not in themselves
harmful. The re-flexes of respiration are activated not by oxygen
deficienc y but by relatively harmless CO2 exces s. The diver who
learns to ignore the signals of CO2 exces s and continues his dive to
approa ch ox ygen deficienc y runs serious r isks.
(8) The proble m of coup ling self-corrective systems together is
central in the adap tation of man to the societies and ecosystems in
which he lives. Lewis Carro ll long ago joked abou t the nature and
order of randomness created by the inappropr iate coup ling of
biologica l systems. The probl em, we may say, was to create a
"game" which should be rando m, not only in the restricted sense in
which "matching pennie s" is rando m, but meta-rando m. The
rando mness of the moves of the two players of "matching penni es"
is restricted to a finite set of known alterna tives, namely "heads" or
"tails" in any given play of the game. There is no possib ility of
going out-s ide this set, no meta-rando m choice among a finite or
infinite set of sets.
By imperfect coupl ing of biologica l systems in the famous game
of croque t, however , Carroll create s a meta-rando m game. Alice is
coupled with a flamingo, and the "ba ll" is a hedgehog.
The "purpose s" (if we may use the term) of these contrasting
biologica l systems are so discrepan t that the rando mness of play can
no longer be delimited with finite sets of alterna tives , known to the
players.
Alice 's difficulty arises from the fact that she does not
"unders tand" the flamingo, i.e., she does not have systemic
information abou t the "system" which confronts her. Similarly, the
flamingo does not unders tand Alice . They are at "cross-purpo ses."
The probl em of coupling man through consc iousness with his
biologica l environment is comparable . If consciou sness lacks
information abou t the nature of man and the environment, or if the
information is distorted and inapp ropriately selected, then the
coupling is likely to generate meta-random sequence of events.
450

(9)We presume that consciousness is not entirely with-out
effect—that it is not a mere collateral resonance without
feedback into the system, an observer behind a one-way mirror, a
TV monitor which does not itself affect the pro-gram. We believe
that consciousness has feedback into the remainder of mind and
so an effect upon action. But the effects of this feedback are
almost unknown and urgently need investigation and validation.
(10)It is surely true that the content of consciousness is no
random sample of reports on events occurring in the remainder of
mind. Rather, the content of the screen of consciousness is
systematically selected from the enormously great plethora of
mental events. But of the rules and preferences of this selection,
very little is known. The matter requires investigation. Similarly
the limitations of verbal language require consideration.
(11)It appears, however, that the system of selection of
information for the screen of consciousness is importantly related
to "purpose," "attention," and similar phenomena which are also
in need of definition, elucidation, etc.
(12)If consciousness has feedback upon the remainder of mind
(9, above), and if consciousness deals only with a skewed sample
of the events of the total mind, then there must exist a systematic
(i.e., nonrandom) difference between the conscious views of self
and the world, and the true nature of self and the world. Such a
difference must distort the processes of adaptation.
(13)In this connection, there is a profound difference between
the processes of cultural change and those of phylogenetic
evolution. In the latter, the Weismannian barrier between soma
and germ plasm is presumed to be totally opaque. There is no
coupling from environment to genome. In cultural evolution and
individual learning, the coupling through consciousness is present,
incomplete and probably distortive.
(14)It is suggest ed that the specific nature of this distortion is
such that the cybern etic natur e of self and the world tends to be
impe rceptible to consc iousness , insofa r as the contents of the
451

"screen" of consciousne ss are determined by conside rations of
purpose . The argument of purpose tends to take the form "D is
desirable; B leads to C; C leads to D; so D can be achieved by way
of B and C." But, if the total mind and the outer world do not, in
gener al, have this lineal structure, then by forcing this structure
upon them, we beco me blind to the cybernet ic circularities of the
self and the e xternal world . Our consc ious sa mpling of dat a will no t
disclose whole circuits but only arcs of circuits, cut off from their
matrix by our selective attention. Spec ifically, the at-tempt to
achieve a change in a given variable , located either in self or
environment, is likely to be unde rtaken withou t comprehension of
the homeosta tic network surrounding that variable. The
consid erations outlined in paragraphs 1 to 7 of this essay will then
be ignored. It may be essen tial for wisdom that the narrow
purpos ive v iew be somehow corrected .
(15)The function of consciousne ss in the coup ling between man
and the homeostat ic systems around him is, of course, no new
pheno menon. Three circumstances , however , make the
inves tigation of this pheno menon a n urgent m atter.
(16)First, there is man's habit of changing his environment rather
than chang ing himself. Faced with a changing variable (e.g.,
temperature) within itself which it should control, the organis m
may make chang es either within itself or in the external
environment. It may adapt to the environment or adap t the
environment to itself. In evolu tiona ry history, the great majorit y of
steps have been changes within the organis m itself; some steps
have been of an intermediate kind in which the organisms achieved
change of environ ment by- change of locale. In. a few cases
organisms other than man have achiev ed the creation of modified
microenv ironments around themselves , e.g., the nests of
hymenoptera and birds, concent rated forests of conif ers, fungal
colon ies, etc.
In all such cases, the logic of evolution ary progress is to-ward
ecosystems which sustain only the dominant , environment-
controlling s pecie s, and i ts symbionts and p arasites.
452

Man, the outstanding modifier of environment, similarly
achieves s ingle-spec ies ecosystems in his cities, bu t he goes one s tep
further, establishing special environments for his symbionts . These,
likewis e, beco me single -speci es ecosystems: fields of corn, cultures
of bacteria, batteries of fowls, colon ies of labora tory rats, and the
like.
(17)Secondl y, the power ratio between purpos ive consciousness
and the environment has changed rapidly in the last one hundred
years, and the rate of change in this ratio is certainly rapidly
increasing with technological advance. Conscious man, as a
changer of his environment, is now fully able to wreck himself and
that environment—with the very best of conscious intentions.
(18)Third, a peculiar sociological phenomenon has arisen in the
last one hundred years which perhaps threatens to isolate
conscious purpose from many corrective processes which might
come out of less conscious parts of the mind. The social scene is
nowadays characterized by the existence of a large number of self-
maximizing entities which, in law, have something like the status
of "persons"—trusts, companies, political parties, unions,
commercial and financial agencies, nations, and the like. In
biological fact, these entities are precisely not persons and are not
even aggregates of whole persons. They are aggregates of parts of
persons. When Mr. Smith enters the board room of his company,
he is expected to limit his thinking narrowly to the specific pur-
poses of the company or to those of that part of the company
which he "represents." Mercifully it is not entirely possible for
him to do this and some company decisions are influenced by
considerations which spring from wider and wiser parts of the
mind. But ideally, Mr. Smith is expected to act as a pure,
uncorrected consciousness—a dehumanized creature.
(19)Finally, it is appropriate to mention some of the factors
which may act as correctives—areas of human action which are
not limited by the narrow distortions of coupling through
conscious purpose and where wisdom can obtain.
(a) Of these, undoubtedly the most important is love. Martin
Buber has classified interpersonal relationships in a relevant
453

manner. He differentiates "I-Thou" relations from "I-It" relations,
defining the latter as the normal pattern of interaction between
man and inanimate objects. The "I-It" relationship he also regards
as characteristic of human relations wherever purpose is more
important than love. But if the complex cybernetic structure of
societies and ecosystems is in some degree analogous to
animation, then it would follow that an "I-Thou" relationship is
conceivable between man and his society or ecosystem. In this
connection, the formation of "sensi tivity groups" in many
deper sonal ized o rganiz ations i s of spe cial interest.
(b)The arts, poet ry, music, and the humanities similarly are areas
in which more of the mind is active than mere consc iousness would
admit. "Le coeur a ses raisons que la raison ne connaît point."
(c) Contac t between man and animals and between man and the
natural world breeds , perhaps—so metimes—wisdo m.
(d)There is religion .
(20) To conc lude, let us remember that job's narrow piety, his
purpos iveness , his common sense, and his worldly succes s are
finally stigmatized , in a marvelous totemic poem, by the Voice out
of the Whirlwind:
Who is this that darkeneth counsel by words
without understanding …
Dost thou know when the wild goats of the rock
bring forth?
Or canst thou tell when the hinds do calve?
454

Form, Subs tance, and Differe nce*
Let me say that it is an extraord inary hono r to be here tonight,
and a pleasu re. I am a little frightened of you all, becaus e I am sure
there are peop le he re who know every field of knowledge that I have
touched much better than I know it. It is true that I have touched a
number of fields, and I probab ly can face any one of you and say I
have touched a field that you have not touched . But I am sure that
for every field I have touched, there are peop le here who are much
more expert than I. I am not a wel l-read phi losophe r, and ph ilosoph y
is not my busines s. I am not a very well-read anthropologi st, and
anthropo logy is not exactly my business .
But I have tried to do something which Korz ybski was very
much concerned with doing , and with which the whole semantic
movement has been conc erned , namely, I have studied the area of
impact between very abstract and formal philosoph ic though t on the
one hand and the natural history of man and other creatures on the
other. This overlap between formal premises and actual behav ior is,
I assert, of quite dread ful importan ce today. We face a world which
is threaten ed not only with disorganiz ation of many kinds, but also
with the destruc tion of its environment, and we, today, are still
unable to think clearly about the relations between an organism
and its environment. What sort of a thing is this, which we call
"organism plus environment"?
Let us go back to the origina l statement for which Korz ybski is
most famous—the statement that the map is not the territory. This
statement came out of a very wide range of philosoph ic thinking,
going back to Greece , and wrigg ling through the history of
Europ ean thought over the last 2000 years. In this history, there has
been a sort of rough dicho tomy and often deep controve rsy. There
has been a violent enmity and blood shed. It all starts, I suppose ,
* This was the Nineteenth Annual Korzybski Mem orial Lecture, delivered
Januar y 9, 1970, under the auspices of the Institute of General Semantics. It is here
re-printed f rom the General Semantics Bulletin , No. 37, 1970 , by permission of the
Institute of General Semantics.
455

with the Pythagore ans versu s their predecesso rs, and the argument
took the shape of "Do you ask what it's made of—ear th, fire, wate r,
etc?" Or do you ask, "What is its pattern?" Pythagora s stood for
inquiry into pattern rather than inquiry into substance.1 That
controvers y has gone through the ages, and the Pythagore an half of
it has, until recen tly, been on the whole the submerged half. The
Gnostic s follow the Pythagor eans, and the alchemists follow the
Gnostic s, and so on. The argument reach ed a sort of climax at the
end of the eighteenth century when a Pythagorean evolutionar y
theory was built and then discarded—a theory which involved
Mind .5
The evolu tiona ry theory of the late eighte enth century, the
Lamarckian theory, which was the first organized transformist
theory of evolu tion, was built out of a curious historical backg round
which has been descr ibed by Lovejo y in The Great Chain of Being.
Before Lamarck, the organic world, the living world , was believed
to be hierarchic in structu re, with Mind at the top. The chain, or
ladde r, went down through the angels, through men, through the
apes, down to the infusor ia or protozoa , and below that to the plants
and s tones .
What Lamarck did was to turn that chain upsid e down. He
observ ed that animals changed under environmental pressure. He
was incorrect, of course , in believing that those changes were
inherited, but in any case, these chang es were for him the evidenc e
of evolut ion. When he turned the ladder upside down, what had
been the explanation, namely, the Mind at the top, now became
that which had to be explained. His problem was to explain Mind.
He was convinced about evolution, and there his interest in it
stopped. So that if you read the Philosophic Zoologique (1809), you
will find that the first third of it is devoted to solving the problem
of evolution and the turning upside down of the taxonomy, and the
rest of the book is really devoted to comparative psychology, a
science which he founded. Mind was what he was really interested
in. He had used habit as one of the axiomatic phenomena in his
5 R. G. Collin gwood has given a clear account of the Pythagorean position in
The Idea of Nature, Oxford, 1945.
456

theory of evolution, and this of course also took him into the
problem of comparative psychology.
Now mind and pattern as the explanato ry princip les whi ch, above
all, requir ed invest igation were pushed out of biolog ical thinking in
the later evolutionar y theories which were deve loped in the mid-
ninete enth century by Darwin , Huxle y, etc. There were still some
naugh ty boys, like Samuel Butler, who said that mind could not be
ignored in this way—but they were weak voice s, and inciden tally,
they never looked at organisms. I don't think Butler ever looked at
anything except his own cat, but he still knew more about evolution
than so me of the m ore conven tiona l thinke rs.
Now, at last, with the discover y of cybernet ics, systems theory,
information theory, and so on, we begin to have a formal base
enabling us to think about mind and enabling us to think abou t all
these problems in a way which was totally heterodox from about
1850 through to World War II. What I have to talk abou t is how the
great dichoto my of epistemology has shifted under the impact of
cybernet ics and i nformation t heory.
We can now say—or at any rate, can begin to say—what we
think a mind is. In the next twen ty years there will be other ways of
saying it and, becaus e the discover ies are new, I can only give you
my persona l version. The old versions are surely wrong, but which
of the revised pictures will survive , we do no t know .
Let us start from the evolut ionar y side. It is now empirically
clear that Darwinian evolu tiona ry theory contained a very great
error in its identification of the unit of survival under natural
selection. The unit which was believed to be crucial and around
which the theory was set up was either the breeding individual or
the family line or the sub-species or some similar homogeneous set
of conspecifics. Now I suggest that the last hundred years have
demonstrated empirically that if an organism or aggregate of
organisms sets to work with a focus on its own survival and thinks
that that is the way to select its adaptive moves, its "progress" ends
up with a destroyed environment. If the organism ends up
destroying its environment, it has in fact destroyed itself. And we
may very easily see this process carried to its ultimate reduc tio ad
absu rdum in the next twenty years. The unit of survival is not the
breeding organism, or the family line, or the society.
457

The old unit has already been partly corrected by the population
geneticists. They have insisted that the evolutionary unit is, in fact,
not homogeneous. A wild population of any species consists
always of individuals whose genetic constitution varies widely. In
other words, potentiality and readiness for change is already built
into the survival unit. The heterogeneity of the wild population is
already one-half of that trial-and-error system which is necessary
for dealing with environment.
The artificially homogenized populations of man's domestic
animals and plants are scarcely fit for survival.
And today a further correction of the unit is necessary. The
flexible environment must also be included along with the flexible
organism because, as I have already said, the organism which
destroys its environment destroys itself. The unit of survival is a
flexible organism-in-its-environment.
Now, let me leav e evolution for a moment to cons ider what is the
unit of mind. Let us go back to the map and the territory and ask:
"What is it in the territory that gets onto the map?" We know the
territory does not get onto the map. That is the central point about
which we here are all agreed. Now, if the territory were uniform,
nothing would get onto the map excep t its boundar ies, which are the
points at which it ceases to be uniform agains t some larger matrix.
What gets onto the map, in fact, is difference, be it a difference in
altitude, a difference in veget ation , a difference in popu lation
structu re, differenc e in surface , or what -ever . Differences are the
things t hat get onto a map.
But what is a differenc e? A differenc e is a very pecul iar and
obscur e conc ept. It is certainly not a thing or an even t. This piece of
paper is different from the wood of this lectern. There are many
differenc es between them—of color, texture, shape, etc. But if we
start to ask about the localization of those difference s, we get into
troub le. Obviousl y the difference between the paper and the wood is
not in the paper ; it is obvious ly not in the wood; it is obvious ly not
in the space between them, and it is obviou sly not in the time
between them. (Differenc e which occur s acros s time is what we call
"change .")
A differen ce, then, is an abs tract matter.
458

In the hard science s, effects are, in general, caused by rather
concrete cond itions or even ts—i mpacts, forces, and so forth. But
when you enter the world of communicat ion, organiz ation , etc., you
leave behind that whole world in which effects are brough t about by
forces and i mpacts and e nergy exchang e. You ent er a world i n which
"effects"—and I am not sure one should still use the same word—
are brought about by differences. That is, they are brought about by
the sort of "thing" that gets onto the map from the territory. This is
difference .
Difference travels from the wood and paper into my retina. It
then gets picked up and worked on by this fancy piece of
computing machinery in my head.
The whole energy relation is different. In the world of mind,
nothing—that which is not—can be a cause. In the hard sciences,
we ask for causes and we expect them to exist and be "real." But
remember that zero is different from one, and because zero is
different from one, zero can be a cause in the psychological world,
the world of communication. The letter which you do not write can
get an angry reply; and the income tax form which you do not fill
in can trigger the Internal Revenue boys into energetic action, be-
cause they, too, have their breakfast, lunch, tea, and dinner and can
react with energy which they derive from their metabolism. The
letter which never existed is no source of energy.
It follows, of course , that we must change our whole way of
thinking abou t mental and communicat ional proce ss. The ordinar y
analogi es of energy theory which people borrow from the hard
science s to provide a conc eptua l frame upon which they try to build
theories about psycholog y and behavior— that entire Procrus tean
structure— is non -sense . It is in error.
. I suggest to you, now, that the word "idea," in its most
elementary sense, is synonymous with "difference." Kant, in the
Critique of Judgmen t—if I understand him correctly—asserts that
the most elementary aesthetic act is the selection of a fact. He
argues that in a piece of chalk there are an infinite number of
potential facts. The Ding an sich, the piece of chalk, can never
enter into communication or mental process because of this
infinitude. The sensory receptors cannot accept it; they filter it out.
459

What they do is to select certain facts out of the piece of chalk,
which then become, in mod-ern terminology, information.
I suggest that Kant's statement can be modified to say that there
is an infinite number of differences around and within the piece of
chalk. There are differences between the chalk and the rest of the
universe, between the chalk and the sun or the moon. And within
the piece of chalk, there is for every molecule an infinite number
of differences between its location and the locations in which it
might have been. Of this infinitude, we select a very limited
number, which be-come information. In fact, what we mean by
information—the elementary unit of information—is a difference
which makes a difference, and it is able to make a difference
because the neural pathways along which it travels and is
continually transformed are themselves provided with energy. The
path-ways are ready to be triggered. We may even say that the
question is already implicit in them.
There is, however, an important contrast between most of the
pathways of information inside the body and most of the pathways
outside it. The differences between the paper and the wood are first
transformed into differences in the propagation of light or sound,
and travel in this form to my sensory end organs. The first part of
their journey is energized in the ordinary hard-science way, from
"behind." But when the differences enter my body by triggering an
end. organ, this type of travel is replaced by travel which is
energized at every step by the metabolic energy latent in the
protoplasm which receives the difference, recreates or transforms
it, and passes it on.
When I strike the head of a nail with a hammer, an impulse is
transmitted to its point . But it is a semantic error, a misleading
metapho r, to say that what travels in an axon is an "impulse." It
could correctly be called "news of a difference ."
Be that as it may, this contrast between internal and external
pathwa ys is not absolute. Except ions occu r on both sides of the line.
Some externa l chains of even ts are energized by relays, and some
chain s of even ts interna l to the body are energized from "behind ."
460

Notab ly, the mechanic al interaction of muscles can be used as a
computationa l model.6
In spite of these excep tions , it is still broadly true that the coding
and transmission of differen ces outside the body is very different
from the coding and transmission inside, and this difference must be
mentioned becau se it can lead us into error. We commonly think of
the external "physical world" as somehow separate from an internal
"mental world ." I believe that this division is based on the contrast
in cod ing and transmission inside and ou tside the bod y.
The mental world— the mind—the world of information
process ing—is not limited b y the skin.
Let us now go back to the notion that the transfo rm of a
difference traveling in a circuit is an elementar y idea. If this be
correct, let us ask what a mind is. We say the map is different from
the territory. But what is the territory? Operationally, somebody
went out with a retina or a measuring stick and made represen tations
which were then put upon paper . What is on the paper map is a
repres entation of what was in the retinal repres entation of the man
who made the map; and as you push the question back, what you
find is an infinite regress, an infinite series of maps. The territory
never gets in at all. The territory is Ding an sich and you can't do
anything with it. Alwa ys the proce ss of represen tation will filter it
out so that the mental world is only maps of maps of maps, ad
infinitum.7 All "phenomena" are literally appearances.
Or we can follow the chain forward. I receive various sorts of
mappings which I call data or information. Upon receipt of these I
act. But my actions , my muscular con-t ractions , are transforms of
differences in the input materia l. And I receive again data which are
transforms of my actions. We get thus a picture of the mental world
6 It is interesting to note that digital computers depend upon transmission of
energy "from behind " to send "news" along wire from one relay to the next. But
each relay has its own energy sourc e. Analogic computers, e.g., tide machines and
the like, are commonly entirel y driven by energy "from behind." Either type of
energization can be used for computational purposes.7 Or we may spell the matter out and say that at every step, as a difference is
transformed and propagated along its pathway , the embodiment of the difference be-
fore the step is a "territory" of which the embodiment after the step is a "map." The
map-territory relation obtains at every step.
461

which has some-how jumped loose from our conven tiona l picture of
the physical wor ld.
This is not new, and for historic background we go again to the
alchemists and Gnostics. Carl Jung once wrote a very curious little
book, which I recommend to all of you. It is called Septem Sermones
ad Mortuos , Seven Sermons to the Dead.8 In his Memoirs , Dreams
and Reflections, Jung tells us that his house was full of ghosts, and
they were noisy. They bothered him, they bothered his wife, and they
bothered the children. In the vulgar jargon of psychiatry, we might say
that everybody in the house was as psychotic as hooty owls, and for
quite good reason. If you get your epistemology confused, you go
psychotic, and Jung was going through an epistemological crisis. So
he sat down at his desk and picked up a pen and started to write.
When he started to write the ghosts all disappeared, and he wrote this
little book. From this he dates all his later insight. He signed it
"Basilides," who was a famous Gnostic in Alexandria in the second
century.
He points out that there are two world s. We might call them two
worlds of explanation. He names them the pleroma and the
creatura, these being Gnostic terms. The pleroma is the world in
which events are caused by forces and impacts and in which there
are no "distinctions." Or, as I would say, no "differences." In the
creatura, effects are brought about precisely by difference. In fact,
this is the same old dichotomy between mind and substance.
We can study and describe the pleroma, but always the
distinctions which we draw are attributed by us to the pleroma. The
pleroma knows nothing of difference and distinction; it contains no
"ideas" in the sense in which I am using the word . When we study
and describe the creatur a, we must correctly identify those
differenc es which are effective within it.
I sugge st that "pleroma" and "creatura" are word s which we
could usefully adopt , and it is therefore worthwhi le to look at the
8 Written in 1916 , translated by H. G. Baynes and privately circulated in 1925.
Republished by Stuart & Watkins, London, and by Random House, 1961 . In later
work, Jung seems to have lost the clarity of the Seven Sermons. In his "Answer to
Job," the archetypes are said to be "pleromatic." It is surely true, however, that
constellations of ideas may seem subjectively to resemble "forces" when their
ideational character is unrecognized.
462

bridges which exist between these two "world s." It is an
oversimplific ation to say that the "hard science s" deal only with the
pleroma and that the scien ces of the mind deal only with the
creatu ra. There is more to it than that.
First, consider the relation between energy and nega tive entropy.
The classical Carnot heat engin e consists of a cylinder of gas with a
piston . This cylinder is alternate ly placed in contact with a container
of hot gas and with a containe r of cold gas. The gas in the cylinder
alternate ly expands and contracts as it is heated or cooled by the hot
and cold sources. The piston is thus driven up and down.
But with each cycle of the engin e, the difference between the
temperatu re of the hot source and that of the cold source is reduced.
When th is difference beco mes zero, the engine will stop.
The physicist, describing the plero ma, will write equations to
translate the temperatu re differenc e into "available energy," which
he will call "nega tive entropy," and wi ll go on from there.
The analyst of the creatu ra will note that the whole system is a
sense organ which is trigge red by temperature difference. He will
call this difference which makes a differenc e "infor mation" or
"negative entropy." For him, this is only a special case in which the
effective difference happen s to be a matter of energetics. He is
equally interes ted in all differen ces which can activate some sense
organ. For h im, any such differenc e is "neg ative entropy."
Or consid er the phenomenon which the neuroph ysiologis ts call
"synaptic summation." What is observed is that in certain cases,
when two neuron s, A and B, have synaptic connection to a third
neuron, C, the firing of neither neuron by it-self is sufficient to fire
C; but that when both A and B fire simultaneousl y (or nearly so),
their combined "impulses" wi ll cause C to fire.
In plero matic language , this combining of events to surmount a
thresho ld is called "su mmation."
But from the point of view of the student of creatura (and the
neurophysiologis t must surely have one foot in the pleroma and the
other in creatura), this is not summation at all. What happen s is that
the system operates to create difference s. There are two
differentiated classes of firings by A: those firings which are
accompanied by B and those which are unaccompanied. Similarly
there are two c lasses o f firings b y B.
463

The so-called "summation," when both fire, is not an additive
proce ss from this point of view. It is the formation of a logica l
produc t—a p rocess of fractiona tion rather than summation.
The creatura is thus the world seen as mind, whereve r such a
view is appropr iate. And wherever this view is appropria te, there
arises a species of complexi ty which is absent from pleromatic
description: creatura l description is al-ways hierarchic .
I have said that what gets from territory to map is trans-forms of
differenc e and that these (somehow selected) differences are
elementar y ideas.
But there are differences between differenc es. Ever y effective
differenc e denote s a demarcation, a line of classification, and all
classification is hierarch ic. In other words, difference s are
themselves to be differentiated and classified. In this context I will
only touch lightly on the matter of classes of difference , becau se to
carry the matter further would land us in problems of Princ ipia
Math ematica .
Let me invite you to a psycholog ical expe rience , if only to
demonstra te the frailty of the human computer. First note that
differenc es in t exture are different (a) from differences in co lor. Now
note that differences in size are different (b) from differences in
shape . Similar ly ratios are different (c) from subtractive d ifferences .
Now let me invite you, as disciples of Korz ybski, to define the
differenc es between "different (a) ," "different (b)," and "differen t
(c) " in the above paragr aph. The computer in the human head
boggles at the task. But not all classes of difference are as awkward
to handle .
One such class you are all familiar with. Namely, the class of
differenc es which are created by the process of trans-fo rmation
whereb y the differences immanent in the territory become
differenc es immanent in the map. In the corne r of ever y serious map
you will find these rules of transformation spelled out—usu ally in
words. Within the human mind, it is absolu tely essen tial to
recogn ize the differenc es of this class, and, indeed, it is these that
form the central subjec t matter of "Scienc e and Sanity."
An hallucin ation or a dream image is surely a transfo rmation of
something . But of what? And by what rules of trans-formation?
464

Lastly there is that hierarch y of differences which biologists call
"level s." I mean such differenc es as that between a cell and a tissue,
between tissue and organ, organ and organism, and organism and
society.
These are the hierarch ies of units or Gestalten, in which each
subun it is a part of the unit of next larger scope . And, alwa ys in
biolog y, this difference or relationship which I call "part of" is such
that c ertain differen ces in t he par t have i nformational e ffect upon the
larger unit, and vice versa.
Having stated this relationsh ip between biologi cal part and
whole , I can now go on from the notion of creatur a as Mind in
general to the question of wha t is a mind.
What do I mean by "my" mind?
I suggest that the delimitation of an individual mind must
always depend upon what phenomena we wish to under-stand or
explain. Obviously there are lots of message path-ways outside the
skin, and these and the messages which they carry must be
included as part of the mental system whenever they are relevant.
Consid er a tree and a man and an axe. We observe that the axe
flies through the air and makes certain sorts of gashes in a pre-
existing cut in the side of the tree. If now we want to explain this set
of pheno mena, we shall be concerned with differences in the cut
face of the tree, differences in the retina of the man, differences in
his central nervous system, differences in his efferent neural
messages, differences in the behavior of his muscles, differences in
how the axe flies, to the differences which the axe then makes on
the face of the tree. Our explanation (for certain purposes) will go
round and round that circuit. In principle, if you want to explain or
understand anything in human behavior, you are always dealing
with total circuits, completed circuits. This is the elementary
cybernetic thought.
The elementary cybernetic system with its messages in circuit
is, in fact, the simplest unit of mind; and the trans-form of a
difference traveling in a circuit is the elementary idea. More
complicated systems are perhaps more worthy to be called mental
systems, but essentially this is what we are talking about. The unit
which shows the characteristic of trial and error will be
legitimately called a mental system.
465

But what about "me"? Suppose I am a blind man, and I use a
stick. I go tap, tap, tap. Where do I start? Is my mental system
bounded at the handle of the stick? Is it bounded by my skin? Does
it start halfway up the stick? Does it start at the tip of the stick?
But these are nonsense questions. The stick is a pathway along
which transforms of difference are being transmitted. The way to
delineate the system is to draw the limiting line in such a way that
you do not cut any of these pathways in ways which leave things
inexplicable. If what you are trying to explain is a given piece of
behavior, such as the locomotion of the blind man, then, for this
purpose, you will need the street, the stick, the man; the street, the
stick, and so on, round and round.
But when the blind man sits down to eat his lunch, his stick and
its messages will no longer be relevant—if it is his eating that you
want to understand.
And in addi tion t o what I have s aid to defin e the i ndividua l mind,
I think it necessa ry to include the relevant parts of memory and data
"banks ." After all, the simplest cybernet ic circuit can be said to have
memory of a dynamic kind—not based upon static storage but upon
the travel of information around the circuit. The behav ior of the
governo r of a steam engine at Time 2 is part ly determined by what it
did at Time 1—whe re the interval between Time 1 and Time 2 is
that time necessa ry for the information to complete t he circuit.
We get a picture, then, of mind as synonymous with cybernet ic
system—the relevan t total information-proc essing , trial-and-er ror
completing unit. And we know that within Mind in the widest sense
there will be a hierarchy of sub-s ystems, any one of which we can
call an ind ividua l mind.
But this picture is precisely the same as the picture which I
arrived at in discu ssing the unit of evolution. I believe that this
identity is the most importan t gene ralization which I have to offer
you tonight.
In conside ring units of evolution , I argued that you have at each
step to include the completed pathwa ys outside the protop lasmic
aggreg ate, be it DNA-in-the-cell, or cell-in-the-bod y, or body-in-the-
environment. The hierarchic structu re is not new. Formerly we
talked about the breeding individu al or the family line or the taxon,
and so on. Now each step of the hierarch y is to be thought of as a
466

system, instead of a chunk cut off and visualized as against the
surrounding matrix.
This identity between the unit of mind and the unit of
evolutionar y survival is of very great importance, not only
theore tical, but also ethica l.
It means, you see, that I now localize something which I am
calling "Mind" immanent in the large biologica l system—the
ecosystem. Or, if I draw the system boundari es at a different level,
then mind is immanent in the total evolut ionar y structur e. If this
identity between mental and evolutionar y units is broadly right, then
we face a number of shifts in ou r thinking.
First, let us conside r ecolog y. Ecolog y has currently two faces to
it: the face which is called bioen ergetics— the economics of energy
and materials within a coral reef, a red-wood forest, or a city—and,
second, an econo mics of information, of entropy, negen tropy, etc.
These two do not fit together very well precisely because the units
are differently bounded in the two sorts of ecolog y. In bioener getics
it is natural and appropriate to think of units bounded at the cell
membrane , or at the skin; or of units composed of sets of
conspeci fic individu als. These boundar ies are then the frontiers at
which measure ments can be made to determine the additive-
subtractive budg et of energy for the given unit. In contrast,
informational or entropic ecology deals with the budget ing of
pathwa ys and of probab ility. The resu lting bud-get s are fractiona ting
(not subtractive). The boundari es must enclose, not cut, the relevant
pathwa ys.
Moreove r, the very meaning of "surviva l" becomes different
when we stop talking about the survival of something bounded by
the skin and start to think of the surviva l of the system of ideas in
circuit. The conten ts of the skin are rando mized at death and the
pathwa ys within the skin are rando mized. But the ideas, under
further transformation , may go on out in the world in books or
works of art . Socrates as a bioene rgetic individual is dead . But m uch
of him still lives as a component in the contemporar y ecology of
ideas.9
9 For the phrase "e colog y of ideas," I am indebted to Sir Geo ffrey Vickers ' essa y
"The Ecolog y of Ideas" in Value Systems and Social Process, Basic Books, 1968.
467

It is also clear that theology beco mes changed and perhaps
renewed . The Medi terranean religion s for 5000 years have swung to
and fro b etween i mmanence and t ransc endence . In Bab ylon the gods
were transcendent on the tops of hills; in Egypt, there was god
immanent in Pharoah; and Christianity is a complex combinat ion of
these two beliefs.
The cybernet ic epistemology which I have offered you would
sugges t a new approach . The individua l mind is immanent but not
only in the body. It is immanent also in pathwa ys and messages
outside the body; and there is a larger Mind of which the individua l
mind is only a sub-s ystem. This larger Mind is comparable to God
and is perhaps what some people mean by "God ," but it is still im-
manent in the total interconnec ted social system and plane tary
ecolog y.
Freudi an psycholog y expand ed the conc ept of mind in-wards to
include the whol e communication system within the body—the
autono mic, the habitu al, and the vast range of unconsc ious process.
What I am saying expand s mind out-wards . And both of these
changes reduce the scope of the conscious self. A certain humilit y
beco mes approp riate, tempered by the digni ty or joy of being part of
something much bigg er. A part—if you will—of God.
If you put God outside and set him vis-à-vis his creation and if
you have the idea that you are created in his image, you will
logically and naturally see yourself as outside and against the things
around you. And as you arroga te all mind to yourself , you will see
the world around you as mindless and therefore not entitled to moral
or ethical considera tion. The environ ment will seem to be yours to
explo it. Your survival unit will be you and your folks or conspec ifics
again st the environment of other social units, other races and the
brutes and vege table s.
If this is your estimate of your relation to natu re and you have an
advanced technolog y, your likelihood of survival will be that of a
snowball in hell. You will die either of the toxic by-produc ts of your
own hate, or, simply, of over-popul ation and overgrazing . The raw
materials of the world are finite.
For a more formal discussion of the survival of ideas, see Gordon Pasks'
remarks in Wenner-Gren Conference on "Effects of Conscious Purpose on Human
Adaptation," 1968
468

If I am right, the whole of our thinking abou t what we are and
what other peop le are has got to be restru ctured. This is not funny,
and I do not know how long we have to do it in. If we continue to
operate on the premises that were fashionab le in the precybernetic
era, and which were especially underlined and streng thened during
the Indus-t rial Revo lution , which seemed to validate the Darwin ian
unit of surviva l, we may have twent y or thirty years befor e the
logica l reduct io ad absur dum of our old positions destroys us.
Nobod y knows how long we have, under the present system, before
some disaster strikes us, more serious than the destruction of any
group of nations. The most important task today is, perhap s, to learn
to think in the new way. Let me say that I don't know how to think
that way. Intellectually, I can stand here and I can give you a rea-
soned exposi tion of this matter; but if I am cutting down a tree, I
still think "Gregor y Bateson" is cutting down the tree. I am cutting
down the tree. "Myself" is to me still an excessive ly concr ete object,
different from the rest of what I have been calling "mind."
The step to realizing—to making habitual— the other way of
thinking— so that one naturally think s that way when one reaches
out for a glass of water or cuts down a tree— that step is not an easy
one.
And, quite seriou sly, I sugges t to you that we shou ld trust no
policy decision s which emanate from persons who do not yet have
that habit.
There are experi ences and discip lines which may help me to
imagine what it would be like to have this habit of correct thought .
Under LSD, I have exper ienced , as have many others, the
disappea rance of the division between self and the music to which I
was listening. The perceiv er and the thing perceived become
strange ly united into a single entity. This state is surely more correct
than the state in which it seems that "I hear the music." The sound,
after all, is Ding an Bich, but my percep tion of it is a part of mind.
It is told of Johann Sebas tian Bach that when somebody asked
him how he played so divine ly, he answered, "I play the notes , in
order, as they are written. It is God who makes the music." But not
many of us can claim Bach's correctnes s of epistemolog y—or that
of William Blake, who knew that the Poetic Imaginat ion was the
only reality. The poet s have known these things all through the ages ,
469

but the rest of us have gone astray into all sorts of false reification s
of the "self" and separations b etween t he "self" and "expe rience ."
For me anothe r clue another moment when the nature of mind
was for a moment clear—was provid ed by the famous expe riments
of Adelbe rt Ames, Jr. These are optical illusions in depth perc eption .
As Ames' guinea pig, you discover that those mental processe s by
which you create the world in three-dimensional perspe ctive are
within your mind but totally uncon scious and utterly beyond
volun tary control. Of course, we all know that this is so— that mind
creates the images which "we" then see. But still it is a pro-found
epistemological shock to have direct experien ce of this which we
alwa ys knew .
Pleas e do not misunders tand me. When I say that the poets have
alwa ys known these things or that most of mental process is
unconsc ious, I am not advoca ting a greate r use of emotion or a
lesser use of intellect. Of course, if what I am saying tonight is
approx imately true, then our ideas about the relation between
though t and emotion need to be revised. If the boundar ies of the
"ego" are wrongl y drawn or even totally fictitious, then it may be
nonsense to regard emotions or dreams or our unconsc ious
computat ions o f perspective as "ego-a lien."
We live in a strange epoch when many psychologis ts try to
"humanize" their science by preaching an anti-intellectual gospel .
They might, as sensibly, try to physicalize physics by discard ing the
tools of mathematics .
It is the attempt to separate intellect from emotion that is
monstrous , and I sugges t that it is equa lly monstrous—and
dangerou s—to attempt to separa te the external mind from the
interna l. Or to separate mind from body.
Blake noted that "A tear is an intellectu al thing," and Pascal
asserted that "The heart has its reasons of which the reason knows
nothing." We need not be put off by the fact that the reasoning s of
the heart (or of the hypothalamus) are accompanied by sensations of
joy or grief. These computations are concerned with matters which
are vital to mammals, namely, matters of relationship, by which I
mean love, hate, respect , dependenc y, spectatorship , performance,
dominance , and so on. These are central to the life of any mammal
and I see no objec tion to calling these computations "though t,"
470

though certainly the units of relationa l computation are differen t
from the units whi ch we us e to compute about isolable things .
But there are bridges between the one sort of thought and the
other, and it seems to me that the artists and poets are speci fically
concerned with these bridges . It is not that art is the express ion of
the unconsc ious, but rather that it is concerned with the relation
between the levels of mental process. From a work of art it may be
possible to analyze out some unconsc ious t houghts of the artist, but I
believe that, for example, Freud's analysis of Leonardo' s Virgin on
the Knees of St. Anne precisely misses the point of the whole exercise .
Artistic skill is the combining of many levels of mind —
uncon scious , consc ious, and external— to make a statement of their
combination . It is not a matter of express ing a single level.
Similar ly, Isadora Duncan , when she said, "If I could say it, I
would not have to dance it," was talking nonsense , be-cause her
dance was about combinat ions o f saying and moving.
Indeed, if what I have been saying is at all correct, the whole
base of aesthetics will need to be re-examined. It seems that we link
feelings not only to the computat ions of the heart but also to
computations in the externa l pathwa ys
of the mind. It is when we recogni ze the operations of creatu ra in
the external world that we are awar e of "beau ty" or "ugliness ." The
"primrose by the river's brim" is beau tiful because we are awar e that
the combination of differenc es which constitutes its appearanc e
could only be achieved by information proces sing, i.e., by thought.
We recognize an-other mind wi thin ou r own e xtern al mind.
And last, there is death. It is unders tandab le that, in a civilization
which separates mind from body, we should either try to forget
death or to make mythologies about the survival of transcendent
mind. But if mind is immanent not only in those pathwa ys of
information which are located in-side the body but also in extern al
pathwa ys, then death takes on a different aspect. The indiv idual
nexu s of pathwa ys which I call "me" is no longer so preciou s bec ause
that nexus is only part of a larger mind.
The ideas which seemed to be me can also beco me immanent in
you. Ma y they survive if true.
471

Comment on Par t V
In the final essay of this part, "Form, Subs tance and Difference,"
much of what has been said in earlier parts of the book falls into
place. In sum, what has been said amounts to this: that in addition to
( and always in confo rmity with) the familiar physical determinis m
which charac terises our univer se, there is a mental determinis m.
This mental determinism is in no sense superna tural. Rath er it is of
the very nature of the macroscopic* world that it exhibi t mental
characteristics. The mental determinis m is not transcenden t but
immanent and is especially complex and evident in those sections of
the unive rse which a re alive or which include l iving things.
But so much of occidenta l thinking is shaped on the premise of
transcenden t deity that it is difficult for many people to rethink their
theories in terms of immanence. Even Darwin from time to time
wrote abou t Natura l Selection in phrases which almost ascribed to
this process the characteristics of transcendenc e and purpose .
It may be worthwhil e, therefore , to give an extreme sketch of the
differenc e between the belief in transcend ence and that in
immanence.
Transcend ent mind or deity is imagined to be personal and
omnisci ent, and as receiving information by channe ls separate from
the earthly. He sees a species acting in ways which must disrupt its
ecolog y and, either in sorrow or in anger , He sends the wars , the
plague s, the pollution , and th e fallou t.
Immanent mind would achieve the same final result but without
either sorrow or anger. Immanent mind has no separ ate and un-
earthly channel s by which to know or act and , therefore, can have no
separate emotion or evalu ative comment. The immanent will differ
from the transcendent in greater determinism.
* I do not agree with Samuel Butler, Whitehead , or Teilhard de Chardin that it
follows from this mental character of the macroscopic world that the single atomies
must have mental character or potentialit y. I see the mental as a function only of
complex relationship .
472

St. Paul (Galatians VI) said that "God is not mocked," and
immanent mind similarly is neither vengefu l nor forgiving . It is of
no use to make excuse s; the immanent m ind is not "mocked."
But since our minds—and this includes our tools and actions—
are only parts of the larger mind, its computations can be con-fu sed
by our contrad ictions and confusions . Since it contains our insanity,
the immanent mind is inevitably subje ct to possible in-san ity. It is in
our powe r, with our technolog y, to create insani ty in the larger
system of which we are parts.
In the final section of the book, I shall consider some of these
mentall y pathogen ic process es.
473

Part V I: Crisis i n the Ecolo gy of
Mind

From Ver sailles to Cybernetics*
I have to talk abou t recent history as it appears to me in my
generation and to you in yours and, as I flew in this morning, words
began to echo in my mind. These were phrases more thund erous
than any I might be able to compose. One of these groups of word s
was, "The fathers have eaten bitter fruit and the children 's teeth are
set on edge ." Anoth er was the st atement of Jo yce that "hi story is tha t
nightmare from which there is no awaken ing." Another was, "The
sins of the fathers shall be visited on the children even to the third
and fourth genera tion of those that hate me." And lastly, not so
immediately relevan t, but still I think relevant to the proble m of
social mechanism, "He who would do good to another must do it in
Minute Particula rs. Gene ral Good is the plea of the scoundre l,
hypocrit e, and f latterer."
We are talking about serious things . I call this lecture "From
Versailles to Cybernet ics," naming the two historic even ts of the
twen tieth centur y. The word "cybernet ics" is familia r, is it not? But
how m any of you know wha t happened at Versailles in 1919?
The quest ion is, What is going to coun t as important in the
history of the last sixty years? I am sixty-two, and, as I began to
think about what I have seen of history in my lifetime, it seemed to
me that I had really only seen two moments that would rate as really
important from an anthropo logis t's point of view. One was the
events leading up to the Treaty of Versailles, and the other was the
cybernet ic breakthrough . You may be surprised or shock ed that I
have not mention ed the A-bo mb, or even World War II. I have not
mentioned the s pread o f the au tomobile , nor of t he rad io and TV, nor
many other thing s that have o ccurr ed in the last sixty years.
Let me state my criterion of historical importanc e:
Mammals in gener al, and we among them, care extremely, not
abou t episodes, but about the patterns of their relation-s hips. When
you open the refrige rator door and the cat comes up and makes
* Previously unpublished. This lecture was given April 21, 1966, to the "Two
Worlds Symposium " at Sacramento State College.
475

certain sounds, she is not talking abou t liver or milk, though you
may know very well that that is what she want s. You may be able to
guess correctly and give her that—if there is any in the refrigera tor.
What she actually says is something about the relationship between
her-self and you. If you transla ted her message into words, it would
be something like, "dependenc y, dependenc y, dependenc y." She is
talking , in fact, about a rather abstract pat-tern within a relationsh ip.
From that assertion of a pattern, you are expe cted to go from the
gener al to the specific— to deduc e "milk" or "liver."
This is crucial. This is what mammals are abou t. They are
conce rned with patterns of relationsh ip, with where they stand in
love, hate, respect , dependenc y, trust, and similar abstractions , vis-à-
vis somebod y else. This is wher e it hurts us to be put in the wrong.
If we trust and find that that which we have trusted was
untru stworth y; or if we distrust, and find that that which we
distrusted was in fact trust-worth y, we feel bad. The pain that
human beings and all other mammals can suffer from this type of
error is extreme. If, therefor e, we really want to know what are the
significant points in history, we have to ask which are the moments
in history when attitudes were changed. These are the moments
when peop le are hu rt because of their former "value s."
Think of the house thermostat in your home. The weather
changes outdoo rs, the temperatur e of the room falls, the
thermometer switch in the living room goes through its business and
switch es on the furnace ; and the furna ce warms the room and when
the room is hot, the thermometer switch turns it off again. The
system is what is called a homeostatic circuit or a servocircuit. But
there is also a little box in the living room on the wall by which
you set the thermostat. If the house has been too cold for the last
week, you must move it up from its present setting to make the
system now oscillate around a new level. No amount of weather,
heat or cold or whatever, will change that setting, which is called
the "bias" of the system. The temperature of the house will
oscillate, it will get hotter and cooler according to various
circumstances, but the setting of the mechanism will not be
changed by those changes. But when you go and you move that
bias, you will change what we may call the "attitude" of the
system.
476

Similarly, the important question about history is: Has the bias
or setting been changed? The episodic working out of events under
a single stationary setting is really trivial. It is with this thought in
mind that I have said that the two most important historic events in
my life were the Treaty of Versailles and the discovery of
cybernetics.
Most of you probabl y hardly know how the Treaty of Versailles
came into being. The story is very simple. World War I dragged on
and on; the Germans were rather obvious ly losing. At this point,
Geor ge Cree l, a public relations man—and I want you not to forget
that this man was a granddadd y of modern public relations—had an
idea: the idea was that maybe the Germans would surrender if we
offered them soft armistice terms. He therefore drew up a set of soft
terms, according to which there would be no punitive measures.
These terms were drawn up in fourteen points. These Fourteen
Points he passed on to President Wilson. If you are going to deceive
somebod y, you had better get an hones t man to carry the message.
Presiden t Wilson was an almost patho logica lly honest man and a
humanitarian. He elabora ted the points in a number of speeches :
there were to be "no annexations, no contribution s, no punitive
damages …" and so on . And t he Germans surrender ed.
We, British and Americans specially the British—continued of
course to blockade Germany becau se we didn' t want them to get
uppity before the Treaty was signed. So, for another year, they
continued to starve.
The Peace Conferenc e has been vividly described by aynard
Keynes in The Economic Consequence s of the Peac e (1919).
The Treaty was finally drawn up by four men: Clemenceau, "the
tiger," who wanted to crush Germany; Lloyd Geor ge, who felt it
would be politically expedien t to get a lot of repara tions out of
Germany, and some revenge; and Wilson, who had to be
bamboozled along . Wheneve r Wilson would wonder about those
Four teen Points of his, they took him out into the war cemeteries
and made him feel ashamed of not being angry with the Germans.
Who was t he other? Orl ando was t he other, an Italian.
This was one of the great sellouts in the history of our
civilization. A most extraordin ary event which led fairly directly and
inevitably into World War II. It also led (and this is perhaps more
477

interes ting than the fact of its leading to World War II) to the total
demoralization of German politics. If you promise your boy
something , and renege on him, framing the whole thing on a high
ethical plane, you will probabl y find that not only is he very angry
with you, but that his moral attitudes deteriorate as long as he feels
the unfair whip lash of what you are doing to him. It's not only that
World War II was the appropri ate response of a nation which had
been treated in this particular way; what i s more important is the fact
that the demoraliz ation of that nation was expec table from this sort
of treatment. From the demoraliz ation of Germany, we, too, became
demoralized. This is why I say that the Treaty of Versailles was an
attitudinal turning point .
I imagine that we have anoth er coupl e of gene rations of
aftereffects from that particula r sellout to work through . We are, in
fact, like members of the house of Atreus in Greek traged y. First
there was Thyestes' adultery, then Atreus ' killing of Thyestes' three
children , whom he served to Thyestes at a peace -making feast. Then
the murder of Atreu s' son, Agamemnon, by Thyestes' son,
Aegisth eus; and finally the murder of Aegis theus and Clytemnestr a
by Ores tes.
It goes on and on. The tragedy of osci llating and self-propaga ting
distrust, hate, and des truction down the generations .
I want you to imagine that you come into the middle of one of
these sequences of traged y. How is it for the middle gener ation of
the hous e of Atreu s? They are living in acrazy universe. From the
point of view of the people who started the mess, it's not so crazy;
they know what happened and how they got there. But the people
down the line, who were not there at the beginning, find
themselves living in a crazy universe, and find themselves crazy,
precisely because they do not know how they got that way.
To take a dose of LSD is all right, and you will have the
experience of being more or less crazy, but this will make quite
good sense because you know you took the dose of LSD. If, on the
other hand, you took the LSD by accident, and then find yourself
going crazy, not knowing how you got there, this is a terrifying and
horrible experience. This is a much more serious and terrible
experience, very different from the trip which you can enjoy if you
know you took the LSD.
478

Now consider the difference between my generation and you
who are under twenty-five. We all live in the same crazy universe
whose hate, distrust, and hypocrisy relates back (especially at the
international level)' to the Fourteen Points and the Treaty of
Versailles.
We older ones know how we got here. I can remember my father
reading the Fourt een Point s at the break fast table and saying, "By
golly, they're going to give them a decent armistice, a decen t peace ,"
or something of th e kind . And I can remember, but I will not at tempt
to verba lize, the sort of thing he said when the Treaty of Versailles
came out. It wasn' t printable . So I know more or less how we got
here.
But from your point of view , we are absolutely crazy, and you
don't know what sort of historic event led to this craziness. "The
fathers have eaten bitter fruit and the children 's teeth are set on
edge." It's all very well for the fathers, they know what they ate. The
children don' t know wha t was e aten.
Let us consider what is to be expec ted of peop le in the aftermath
of a major decept ion. Previous to World War 1, it was gene rally
assumed that compromise and a little hypocrisy are a very important
ingred ient in the ordinary comfortablen ess of life. If you read
Samuel Butle r's Erewhon Revisited, for example, you will see what I
mean. All the principal chara cters in the novel have got themselves
into an awfu l mess: some are due to be execu ted, and other s are due
for public scanda l, and the religious system of the nation is
threatened with collapse. These disasters and tangles are smoothed
out by Mrs. Ydgrun (or, as we would say, "Mrs . Grund y"), the
guardian of Erewhonian morals. She carefully reconstru cts history,
like a jigsaw puzzle , so that nobod y is really hurt and nobod y is
disgrac ed—st ill less is anybody execu ted. This was a very
comfortabl e philosoph y. A little hypocris y and a little compromise
oil the wheels of social life.
But after the great deception, this philosoph y is untenable. You
are perfec tly correct that something i s wrong ; and that the something
wrong is of the natur e of a deceit and a hypocrisy. You live in the
midst of corruption.
Of course , your natural responses are puritanical. Not sexual
puritanism, because it is not a sexua l deceit that lies in the
479

background . But an extreme puritanism against compromise, a
puritanism again st hypocris y, and this ends up as a reduct ion of life
to little pieces . It is the big integrated structures of life that seem to
have carried the lunac y, and so you try to focus down on the
smallest thing s. "He who would do good to anothe r must do it in
Minu te Particulars. Genera l Good is the plea of the scound rel,
hypocrite, and flatterer." The gene ral good s mells of hypocris y to the
rising genera tion.
I don't doubt that if you asked Geor ge Creel to justify the
Fourte en Points, he would urge the general good. It is possible that
that little operation of his saved a few thousand American lives in
1918. I don't know how many it cost in World War II, and since in
Korea and Vietna m. I recall that Hiroshi ma and Nagasak i were
justified by the gener al good and saving American lives. There was
a lot of talk abou t "uncondi tiona l surrender ," perhaps because we
could – not trust ourselv es to hono r a condi tiona l armistice . Was the
fate of Hirosh ima determined at Versailles?
Now I want to talk abou t the other significant historical even t
which has happ ened in my lifetime, approx imately in 1946-47. This
was the growing together of a number of idea s which had developed
in differen t places during World War II. We may call the aggregate
of these ideas cybernet ics, or communicat ion theory, or information
theory, or systems theory. The ideas were generated in many places:
in Vienna by Bertalanffy, in Harva rd by Wiener , in Prince ton by von
Neumann, in Bell Telephone labs by Shannon , in Cambridge by
raik, and so on. All these separate developments in different
intellectual centers dealt with communicational problems,
especially with the problem of what sort of a thing is an organized
system.
You will notice that everything I said abou t history and about
Versailles is a discu ssion of organiz ed systems and their proper ties.
Now I want to say that we are develop ing a certain amount of
rigorous scientific unders tanding of these very mysterious organized
systems. Our knowledge toda y is way ahead of an ything th at Geor ge
Creel c ould have s aid. He was an a pplied sci entist befo re the s cience
was r ipe to be app lied.
One of the roots of cyberne tics goes back to Whit ehead and
Russel l and what is called the Theory of Logic al Types. In princip le,
480

the na me is no t the t hing na med, and the na me of the na me is no t the
name, and so on. In terms of this powerful theory, a message about
war is not part of the war .
Let me put it this way: the message "Let's play chess" is not a
move in the game of chess. It is a message in a more abstract
language than the language of the game on the board. The message
"Let's make peace on such and such terms" is not within the same
ethical system as the deceits and tricks of battle. They say that all is
fair in love and war, and that may be true within love and war, but
outside and about love and war, the ethics are a little different. Men
have felt for centur ies that treach ery in a truce or peace -making is
worse than trickery in battle. Today this ethical principle receives
rigorous theoret ical and scientific suppor t. The ethics can now be
looked at with formalit y, rigor, logic, mathematics, and all that, and
stands on a different sort of basis from mere invoca tiona l
preach ments. We do not have to feel our way; we can sometimes
know right from wrong.
I included cybernetics as the second historic even t of importance
in my lifetime becaus e I have at least a dim hope that we can bring
ourselve s to use this new unders tanding with some hones ty. If we
unde rstand a little bit of what were doing , maybe it will help us to
find our way out of the maze of hallucinations that we have created
around ou r-selves.
Cybernetic s is, at any rate, a contribu tion to change—not simply
a change in a ttitude , but ev en a change i n the unde r-standing o f what
an attitude i s.
The stance that I have taken in choosing what is importan t in
history—saying that the important things are the moments at which
attitude is determined, the moments at which the bias of the
thermostat is changed—th is stance is deriv ed directly from
cybernet ics. These are thought s shaped by even ts from 1946 and
after.
But pigs do not go around ready-roasted. We now have a lot of
cybernet ics, a lot of games theory, and the beginnings of
unde rstand ing of complex systems. But any understanding can be
used in destructive ways.
I think that cyberneti cs is the bigges t bite out of the fruit of the
Tree of Knowledge that mankind has taken in the last 2000 years.
481

But most of such bites out of the apple have proved to be rather
indigest ible—usu ally for cyberneti c reasons.
Cybernet ics has integrity within itself, to help us to not be
seduced by it into more lunacy, but we cannot trust it to keep us
from sin.
For example, the state depar tments of severa l nations are today
using games theory, backed up by computers, as a way of decid ing
interna tiona l policy. They identify first what seem to be the rules of
the game of international interaction; they then consid er the
distribu tion of strength , weapons, strategic points, grievances , etc.,
over the geography and the identified nations. They then ask the
computers to compute what shou ld be our next move to minimize
the chances of our losing the game. The computer then crank s and
heaves and gives an answer , and there is some temptation to obey
the computer . After all, if you follow the computer you are a little
less responsible than i f you made up your own m ind.
But if y ou do what t he co mputer advis es, you assert b y that m ove
that you support the rules of the game which you fed into the
computer . You have a ffirmed the rules of that game.
No doub t nations of the other side also have computers and are
playing similar games, and are affirming the rules of the game that
they are feeding to their computers . The result is a system in which
the rules of intern ation al interaction beco me more and m ore rigid.
I submit to you that what is wrong with the international field is
that the rules need chang ing. The question is not that is the best
thing to do within the rules as they are at the moment. The
question is how can we get away from the rules within which we
have been operating for the last ten or twenty years, or since the
Treaty of Versailles. The problem is to change the rules, and
insofar as we let our cybernetic inventions—the computers—lead
us into more and more rigid situations, we shall in fact be
maltreating and abusing the first hopeful advance since 1918.
And, of course, there are other dangers latent in cybernetics and
many of these are still unidentified. We do not know, for example,
what effects may follow from the computerization of all
government dossiers.
But this much is sure, that there is also latent in cybernetics the
means of achieving a new and perhaps more human outlook, a
482

means of changing our philosophy of control and a means of
seeing our own follies in wider perspective.
483

Pathologies of Epis temology*
First, I would like you to join me in a little exper iment. Let me
ask you for a show of hands. How many of you will agree that you
see me? I see a number of hands—so I guess insanity loves
company. Of course, you don't "really" see me. What you "see" is a
bunch of pieces of information about me, which you synthesi ze into
a picture image of me. You make th at image. It's that simple.
The propos ition "I see you" or "You see me" is a proposition
which contains within it what I am calling "epis temology." It
conta ins within it assumptions about how we get in-formation , what
sort of stuff information is, and so forth. When you say you "see"
me and put up your hand in an innoc ent way, you are, in fact,
agreeing to certain propo sitions about the nature of knowing and the
nature of the universe in which we live and how we know abou t it.
I shall argue that many of these propo sitions happen to be false,
even though we all share them. In the case of such epistemological
propos itions, error is not easily detected and is not very quickly
punished . You and I are able to get along in the world and fly to
Hawaii and read papers on psychiat ry and find our places around
these tables and in genera l function reasonably like human beings
in spite of very deep error. The erroneous premises, in fact, work.
On the other hand, the premises work only up to a certain limit,
and, at some stage or under certain circumstances, if you are
carrying serious epistemological errors, you will find that they do
not work any more. At this point you discover to your horror that it
is exceedingly difficult to get rid of the error, that it's sticky. It is as
if you had touched honey. As with honey, the falsification gets
around; and each thing you try to wipe it off on gets sticky, and
your hands still remain sticky.
Long ago I knew intellectu ally, and you, no doub t, all know
intellectua lly, that you do not see me; but I did not really encounte r
* This paper was given at the Second Conference on Mental Health in Asia and
the Pacific, 1969 , at the East-West Cent er, Hawaii. Copyright © 1972 by the East-
West Center Press. It will also appear in the report of that conference and is here
reprinted b y permission of the East- West Center Press, Hawaii
484

this truth until I went through the Adelber t Ames expe riments and
encounter ed circumstances under which my epistemologica l error
led to errors of action.
Let me describe a typical Ames exper iment with a pack of Luck y
Strike cigarettes and a book of matches. The Luck y Strikes are
placed about three feet from the subjec t of experiment suppo rted on
a spike above the table and the matches are on a similar spike six
feet from the subject. Ames had the subject look at the table and say
how big the objects are and where they are. The subje ct will agree
that they are wher e they are, and that they are as big as they are, and
there is no apparen t epistemologica l error. Ames then says, "I want
you to lean down and look through this plank here." The plank
stands vertically at the end of the table. It is just a piece of wood
with a round hole in it, and you look through the hole. Now, of
course, you have lost use of one eye, and you have been brough t
down so that you no longe r have a crow's-e ye view. But you still see
the Luck y Strikes where they are and of the size which they are.
Ames then said, "Why don't you get a parallax effect by sliding the
plank?" You slide the plank sidewa ys and suddenly your image
changes. You see a little tiny book of matches about half the size of
the origin al and placed three feet from you; while the pack of Luck y
Strikes appear s to be twice its origina l size, and is now six feet
away.
This effect is accomplished very simply. When you slid the
plank, y ou in fac t opera ted a le ver unde r the t able which y ou had no t
seen. The lever reversed the parallax effect; that is, the lever caused
the thing which was closer to you to travel with you, and that which
was far from you to get left behind.
Your mind has been trained or genot ypical ly determined —and
there is much evidence in favor of training— to do the mathematics
necessa ry to use parallax to create an image in depth. It performs
this feat without volition and without your consciousn ess. You
cannot c ontro l it.
I want to use t his example as a p aradig m of the sor t of er ror that I
intend to talk about . The case is simple; it has experimental back ing;
it illustrates the intangib le nature of epistemological error and the
difficulty of changing epistemological habit.
485

In my everyday thinking, I see you, even though I know
intellectua lly that I don't. Since abou t 1943 when I saw the
exper iment, I have worked to practice living in the world of truth
instead of the world of epistemologica l fantasy; but I don't think I've
succe eded. Insan ity, after all, takes psycho-th erapy to change it, or
some very great new exper ience . Just one experien ce which ends in
the laboratory is in-sufficient.
This morning, when we were discu ssing Dr . Jung's pape r, I raised
the question which nobod y was willing to treat seriousl y, perhaps
becaus e my tone of voice encouraged them to smile. The question
was whethe r there are true ideologies . We find that differen t peop les
of the world have different ideolog ies, different epistemologies ,
differen t ideas of the relation ship between man and nature, differen t
ideas about the nature of man himself, the nature of his knowledge,
his feelings , and his will. But if there were a truth about these
matters, then only those social groups which thought according to
that truth could reasonab ly be stable. And if no culture in the world
thinks a ccording to that truth, then there would be no stable culture.
Notice again that we face the question of how long it takes to
come up agains t trouble . Epistemologica l error is often reinforced
and therefore self-validating. You can get along all right in spite of
the fact that you entertain at rather deep levels of the mind premises
which a re simply false.
I think perhaps the most interesting— though still incomplete—
scientific discove ry of the twen tieth century is the discover y of the
nature of mind. Let me outline some of the ideas which have
contribut ed to this discover y. Immanuel Kant , in the Critique of
Judgment , states that the primary act of aesthetic judgment is
selection of a fact. There are, in a sense , no facts in natur e; or if you
like, there are an infinite number of potential facts in nature, out of
which the judgment selects a few which beco me truly facts by that
act of selection. Now, put beside that idea of Kant Jung's insigh t in
Seven Sermons to the Dead , a strange docu ment in which he point s
out that there are two worlds of explan ation or worlds of
unders tanding , the pleroma and the creatura. In the pleroma there
are only forces and impacts. In the creatu ra, there is difference. In
other words , the pleroma is the world of the hard science s, while the
creatura is the world of communication and organization. A
486