Motivation and Emotion, Vol. 20, No. 2, 1996 [609739]

Motivation and Emotion, Vol. 20, No. 2, 1996
Facial Feedback Hypotheses: Evidence,
Implications, and Directions I
Daniel N. McIntosh 2
This review evaluates four facial feedback hypotheses, each proposing a certain
relation between the face and emotions. It addresses criticisms of the data,
considers implications for emotional and social processes, and advises
directions for future research. The current data support the following: Facial
actions are sensitive to social context, yet correspond to the affective dimension
of emotions; matches with specific emotions are unlikely. They modulate
ongoing emotions, and initiate them. These two claims have received
substantially improved support, in part due to studies controlling for effects of
experimental demand and task difficulty. Facial action may influence the
occurrence of specific emotions, not simply their valence and intensity. Facial
action is not necessary for emotions. There are multiple and nonmutually
exclusive plausible mechanisms for facial effects on emotions. Future work
must focus on determining the relative contributions of these mechanisms, and
the parameters of their effects on emotions.
Smile when you read this! Proponents of facial feedback hypotheses claim
that doing so cause you to like this article more, and make your mood
more positive. Skeptics disagree. Despite a century of discussion of this
question, research has been strongly evident for less than 25 years, and our
knowledge regarding facial effects on emotion is changing rapidly. The
most recent comprehensive reviews concluded that facial actions influence
emotions (Adelmann & Zajonc, 1989; Manstead, 1988). Nonetheless, ques-
tions about the interpretation of the data and the importance of the effect
1I appreciate the helpful comments of Harry Gollob, Greg McHugo, Catherine Reed, Craig
Smith, and R. B. Zajonc on earlier drafts of this paper.
2Address all correspondence to Daniel N. Mclntosh, Department of Psychology, University
of Denver, Denver, Colorado 80208.
121
0146-7239/96/060~0121509.50/~ © 1996 Plenum Publishing Corporation

122 Mclntosh
keep the idea controversial (Buck, 1984; Fridlund, 1994; Matsumoto, 1987).
Further, the nature and mechanisms of the phenomenon are debated even
among supporters. Less discussed are reasons why psychologists should care
about this phenomenon.
EVOLUTION OF FACIAL FEEDBACK HYPOTHESES
The idea that the face influences emotions has adapted and mutated
over the years, providing variation and viability to its descendants. Psycholo-
gists studying the effects of the face on emotions trace their inspiration to
Charles Darwin and William James. Darwin (1872/1955) noted that enhanc-
ing or inhibiting the expression of an emotion alters the intensity of that
emotion. He wrote this in order to explain why expression is an important
topic for, and explicable from, a naturalist perspective; he also noted that
emotional expressions serve a social-communicative function. However, so-
matic influence on emotion is neither a primary concern, nor integral to
his theorizing. James (1890/1950) went further, proposing that physiological
changes, including facial action, are the emotions. The body changes, and
the subjective experience of emotion follows. Somatic input is essential to
the theory. This contrast is one root of the debate surrounding facial feed-
back. For the "Jamesian" emotions theorists, a small effect is important,
as it pertains to understanding details of emotions processes. Others typi-
cally consider the effect interesting only if it plays a role in "everyday phe-
nomena" (Fridlund, 1994, p. 182).
Although a strong critique by Cannon (1927) muted consideration of
feedback, two ideas revived interest in the effects of facial movement in
the 1960s. First, Tomkins (1962) gave the face a primary role in his emo-
tions theory: "The face expresses affect, both to others, and to the self, via
feedback, which is more rapid and more complex than any stimulation of
which the slower moving visceral organs are capable" (p. 205). This avoided
Cannon's criticisms by making visceral changes unnecessary, and by placing
the focus on a part of the body relatively untouched by Cannon's analysis
(see also Gellhorn, 1964). Building on this, Izard (1971) indicated that dis-
crete facial configurations are a primary determinant of the quality of the
felt emotion. As with James, facial action is a fundamental component in
this approach.
The second idea is based on the two-factor theory of emotion, in which
intensity of emotion is associated with generalized physiological arousal,
and the specific emotion is mediated by cognitive processes (Schachter &
Singer, 1962). Bern (1967) built on this notion by discussing how emotional
experience may be based in part on cognitive self-perception processes.
This theory suggests that facial configuration may be one source of infor-

Facial Feedback Hypotheses 123
mation used in making self-attributions of emotions (Laird, 1974), but facial
effects on emotion are not fundamental to it.
A decade later, these two theoretical branches bore empirical fruit.
Laird (1974) developed the self-perception view of emotional experience
and tested the role facial action plays. He indicated that if people are in-
duced to express an emotion, they may then feel that emotion. Lanzetta
and colleagues found that individuals who feigned not experiencing pain
during electric shock showed fewer physiological and subjective pain re-
sponses than those who pretended to be experiencing unbearable pain
(Lanzetta, Cartwright-Smith, & Kleck, 1976; see also Kleck, Vaughan, Cart-
wright-Smith, Vaughan, Colby, & Lanzetta, 1976).
By the late 1970s, this theorizing and data gave currency to the idea
that facial movement could influence emotions. The term facial feedback
hypothesis was coined because the effect had become somewhat separated
from its theoretical parents (Ellsworth & Tourangeau, 1981; Tourangeau &
Ellsworth, 1979). The label carries connotations regarding the nature of
the process. For consistency's sake, I use the term to denote the general
area of analysis, but note that not all proposed mechanisms assume that
facial effects involve feedback. Further, there are variations on what rela-
tion is suggested by the term. The diversity of hypotheses labeled "facial
feedback" has confused the debate about such effects, making the associa-
tion between facial patterns and emotions less clear. However, careful con-
sideration of the data supports the existence of some specific relations, and
provides information relevant to emotions (and other) processes. It is time
to move past demonstrations of an effect, and to use the data as a basis
for addressing more sophisticated questions and applications. The first step
in this process is clarifying what is known.
EVALUATION OF CONTEMPORARY FACIAL FEEDBACK
HYPOTHESES
There are four common general proposals related to facial feedback.
They may not be equally valid. To best grasp what we know about this
process, it is important to specify which hypothesis is being tested. I have
organized this review around the four questions. Three are based on the
scheme used in Adelmann and Zajonc's (1989) review, although the cate-
gorization of studies is slightly different. The first does not imply causality:
(1) Does facial configuration correspond to emotions? The next two hold
facial actions as able to influence emotions: (2) Does facial movement
modulate emotions in the presence of other emotional stimuli; and (3) can
facial action initiate emotions, even with no (other) emotional stimuli pre-

124 Mclntosh
sent? The last question follows Tourangeau and Ellsworth (1979): (4) Is
facial action necessary for the presence of emotions? Note that causality is
not unavoidably implied in the necessity hypothesis. It may be that par-
titular facial configurations are an essential component of emotions, but
serve no causal function. Cutting across all four questions is the issue of
what emotional outcome is supported. Is the face associated with the de-
gree of positive or negative affect only (dimensional versions), or is there
a link to specific emotions–anger rather than sadness, for example (cate-
gorical versions)? Tests of the hypotheses pertained mostly to the dimen-
sional question through the mid-1980s (Winton, 1986). One question for
the present review, then, is to consider the degree of support for the cate-
gorical versions evident since that time.
Instead of noting every study, I emphasize studies from the late 1980's
onwards, as the last comprehensive reviews were published then. Further,
research since that time has addressed primary methodolic criticisms and
begun to fill some conceptual gaps; I will focus on what has been learned.
I spend little space on between-subjects versions of the hypotheses. These
hold that those individuals who are more expressive should experience
more intense emotions. There is no such positive association. People who
tend to express more do not feel more, in fact, their physiology tends to
be less reactive (Buck, 1980; e.g., Lanzetta & Kleck, 1970). Although re-
lated to the phenomenon at hand, this area deserves a separate review.
The within-subjects versions are the hypotheses presently at issue. Within
individuals, does facial action correspond to or cause changes in experience
and physiology?
Do Facial Actions Correspond to Emotions?
This hypothesis simply states that facial patterns covary with emotions.
Supportive studies find that emotion associates with visually observed facial
activity (Buck, Savin, Miller, & Caul, 1972; Ekman, Friesen, & Ancoli,
1980) and facial electromyography (EMG; Cacioppo, Petty, Losch, & Kim,
1986; Dimberg 1988). EMG not only changes as a result of manipulation
of emotions (i.e., the probability of a particular EMG pattern given the
associated affective state is high), but also predicts self-reported emotions
(i.e., the probability of an affective state is high given a particular EMG
state) (Cacioppo, Martzke, Petty, & Tassinary, 1988).
These studies support strongly the dimensional version of the corre-
spondence hypothesis. Although many studies simply have not checked for
categorical effects (i.e., within-valence differences), most that did have been
largely unsuccessful or inconsistent (Camras, Holland, & Patterson, 1993;

Facial Feedback Hypotheses 125
e.g., Brown & Schwartz, 1980; Fridlund, Schwartz, & Fowler, 1984;
Tourangeau & Ellsworth, 1979). Note that the lack of correspondence be-
tween specific facial patterns and particular emotional experiences does not
contradict the data showing that people can identify archetypal facial pat-
terns associated with categorized emotions (see, e.g., Ekman, 1994). Know-
ing that a certain pattern is typical of anger, for example, is different than
consistently possessing that pattern when anger is felt.
The difficulty in finding categorical correspondence suggests a funda-
mental problem. There may be too many antecedents and contexts for fa-
cial movement for one-to-one relations between muscle patterns and
particular emotions to be the rule ('lhssinary & Cacioppo, 1992). Note that
this analysis leaves room for a modulating or initiating function of facial
action. Many variables may influence facial action, reducing correspon-
dence. Nonetheless, particular facial patterns may still sway emotional state.
The lack of correspondence simply suggests nonfacial influences on emo-
tion, and nonemotional influences on facial movement.
The social context of the actor is one variable that affects facial dis-
plays. A likely primary function of facial motion is communication of inner
state or intent to others. Buck (1980, 1994), for example, does a nice job
portraying how facial actions are influenced by both internal states and
social factors. Going further, some (e.g., Fridlund, 1994) argue that facial
configurations are a function of social context or communicative intent
only, not emotional state. Consistent with the notion that there is strong
social effect on facial activity, audience effects on smiles are found as early
as 10 months of age (Jones, Collins, & Hong, 1991; for adult data, see
Chovil, 1991, Kraut & Johnston, 1979). There are two problems with the
position that social considerations are the sole influence on facial pattern-
ing, however. To begin, facial action can vary as a function of affective
state even when the movements are too small for observers to notice (Ca-
cioppo, Bush, & Tassinary, 1992). Thus, studies finding no visible changes
in connection with emotional stimuli or state are not evidence that facial
action and emotions are independent. More important, social modulation
does not preclude a connection to emotional state. Recent work demon-
strates that facial movement is influenced by both factors. Cacioppo et al.
(1992) found that:(1) EMG varied with the affective content of slides even
when participants viewed them privately;(2) facial responses were similar
during social vs. asocial emotional stimuli; and (3) movement varied with
affective tone of stimuli even when individuals were told to inhibit expres-
sion to avoid communicating their emotional responses. Hess, Banse, &
Kappas (1995) compared women's reactions to films varying in emotional
strength under conditions varying in sociality and interpersonal relationship.
Facial action was influenced by both social factors and emotional intensity

126 Mclntosh
(see also Kraut, 1982). The data show a clear correspondence in valence
between affective state and facial patterns. A categorical association be-
tween specific emotions and patterns is difficult to find, perhaps because
of other (e.g., social) influences on facial motion.
Given the above findings, what are some directions for future work?
One is to follow the lead of Smith (1989), and search for correspondences
between specific facial movements (e.g., the eyebrow frown), and distinct
components of emotions (e.g., appraisals and physical changes). This level
of analysis may allow for more precise descriptions of what factors are as-
sociated with which facial actions, and which aspects of emotions. The di-
mensional findings are consistent with this approach; extending this level
of analysis to other dimensions and facial movements has potential to un-
tangle the mutual influences affecting the correspondence between facial
actions and emotional processes.
Does Facial Movement Modulate Emotions?
The modulation hypothesis states that facial action can affect emotions
during ongoing emotional experiences. This may refer to intensity (e.g.,
frowning when already sad might make one sadder) or quality of experience
(e.g., smiling when sad may make the sadness bittersweet). However, tests
of this hypothesis have focused on changes in intensity. The emotion theo-
ries that formed the early basis for facial feedback focused on the role of
the face in ongoing emotional experience (Izard, 1981; Tomkins, 1981).
Thus, demonstrations of modulation are important for evaluating these
views. Additionally, this is a potential causal function, and the issue of cau-
sality is implied in folk theories, in how people interpreted the initial theo-
ries (Ellsworth & Tourangeau, 1981), by Izard's (1981, 1990) propositions,
and by the earliest studies. Further, if a causal role is established, this in-
creases the practical implications of the phenomenon (Tourangeau &
Ellsworth, 1979). Whether it be by modulation or initiation, this is the cru-
cial question regarding facial feedback: Does the face do more than express
emotions; does it also change them?
When participants are instructed how to alter their faces muscle by
muscle, their responses to stimuli are sometimes influenced by the pattern
in which their faces are set (Laird, 1974; Rutledge & Hupka, 1985), but
not always (Tourangeau & Ellsworth, 1979). Exaggeration or inhibition of
expressions while experiencing stimuli also provide uneven support for the
modulation hypothesis (Bush, Barr, McHugo, & Lanzetta, 1989; McCanne
& Anderson, 1987; Ohira & Kurono, 1993). Concerns about demand effects

Facial Feedback Hypotheses 127
(Buck, 1980; Manstead, 1988) and questions about the support for the cate-
gorical hypothesis (Winton, 1986) plague this work.
A recent study provides more solid support for the modulation hy-
pothesis, while also addressing persistent concerns. Using imagery to gen-
erate emotions, Hess, Kappas, McHugo, Lanzetta, and Kleck (1992)
compared the time it took individuals to reach four emotional states during
three conditions. In the "Feel" condition participants generated an emo-
tion. For the "Feel-and-Show" condition they generated the target emotion
and displayed the facial configuration consistent with it. In the "Show" con-
dition they posed an expression. In the Feel condition, subjects took longer
to achieve the emotion state than in the Feel-and-Show condition. The ad-
dition of facial action facilitated the self-generation of the emotions. These
data address several concerns. First, demand is an unlikely explanation for
the effects, as the measure is indirect, and participants did not know that
response latency was being assessed. Second, perceived task difficulty was
not associated with stronger negative and weaker positive emotional feel-
ings. Finally, because two positive and two negative emotional states were
used, and there were differences in self-reported emotions consistent with
facial expression, it provides evidence for the categorical version of the hy-
pothesis.
When physiological measures are used, it is implausible that partici-
pants are simply conforming to experimental demand (Buck, 1980). Lan-
zetta and colleagues (Colby, Lanzetta, & Kleck, 1977; Kleck et al., 1976;
Lanzetta et al., 1976) report that alterations of pain expressions during
shock influence both electrodermal response (EDR) and subjective report.
These studies show that facial and bodily action influence physiological and
subjective status. Although pain is not a clearly emotional outcome (Buck,
1980; Tourangeau & Ellsworth, 1979), the studies reported by Lanzetta et
al. included EDR assessment during anticipation of electric shock. Facial
posing during this anxiety-provoking interval influenced EDR in the same
way. Other research also finds effects on emotional responses, although
the support is uneven (Ianni, Stettner, & Freedman, 1986; Zuckerman,
Klorman, & Larrance, 1981).
Some recent studies address the issue of experimental demand by al-
tering facial action by having people engage in emotion-analogous facial
movement. Effects of facial configuration on responses provide support for
the modulating function. Strack, Martin, & Stepper (1988) instructed par-
ticipants to hold a pen in their teeth, lips, or nondominant hand. The first
produces a smile-like pattern, the second inhibits it, and the third is a con-
trol condition. Individuals found cartoons more humorous during the simu-
lated smile, and less humorous when their smiles were suppressed. Zajonc,
Murphy, and Inglehart (1989) created facial actions analogous to positive

128 Mclntosh
and negative configurations by having participants pronounce vowels. In
one study, speakers read aloud stories with either a high frequency of the
vowel ~/, or none of that vowel. Voicing g/results in facial activity similar
to a frown or scowl. If facial action alters the affective impact of the stories,
then the participants should experience more negative feelings during the
high-ii story than during the no-t~ one. They did. Larsen, Kasimatis, and
Frey (1992) attached golf tees to participants' brows. Contraction of the
corrugator supercilii causes the tips of the two tees to touch, and produces
an action generally associated with sad facial expressions, among others.
When individuals kept the tees together, they reported that sad photo-
graphs made them sadder than when they were instructed to keep the tees
apart. They found no differences between the tasks on rated difficulty, sug-
gesting that this is not the reason for the different emotion rating.
In summary, reviews completed in the late 1980s concluded that facial
action modulated ongoing emotions (Adelmann & Zajonc, 1989; Manstead,
1988). Since these reviews, additional studies have addressed this issue, and
the majority have been supportive. The greater number of scientists and
methods that find the effect increases confidence that it is not artifactual.
Further, the more recent positive efforts have responded to the methodo-
logical criticisms of the earlier ones, and address directly issues of experi-
mental demand, effort effects, and categorical effects. Only work on the
latter, however, begins to address the issue of qualitative (vs. intensity)
modulation; future work should address this more directly. Despite the
stronger empirical basis, however, the number of studies with nonsignificant
results suggests that the effect is not large, the outcome may be moderated
by unmeasured variables, some research is using unreliable measures or
invalid manipulations, or some combination of these. One task is to estab-
lish clearly conditions under which facial action modulates emotional ex-
perience.
Can Facial Action Initiate Emotions
The second causal facial feedback hypothesis holds that people can
initiate emotions through facial movement. This has often been considered
a stronger hypothesis than modulation; however, initiating an emotional
state when none is present is not necessarily a more powerful effect than
altering an extant state elicited by other stimuli.
Several studies demonstrate initiation (e.g., Duncan & Laird, 1977,
1980), and more recently, categorical effects have been addressed. In one
study, participants contracted muscles involved in fear, anger, sadness, and
disgust, in turn, and rated their emotions subsequent to each contraction

Facial Feedback Hypotheses 129
on these feelings, plus happiness, surprise, agreeableness, and interest (Du-
clos, Laird, Schneider, Sexter, Stem, & Van Lighten, 1989). Fear ratings
were significantly higher during the fear expression than in any of the other
three, and sad was higher during the sad pattern. Anger and disgust were
not significantly differentiated from each other, but were from fear and
sadness. These data suggest that facial patterns can sometimes initiate emo-
tion-specific experience (see also Laird, Cuniff, Sheehan, Shulman, &
Strum, 1989). However, this work is vulnerable to potential experimental
demand effects.
Physiological measures are less likely to be products of demand, and
there is some evidence that physiology is altered by facial manipulations.
An early demonstration of autonomic outcomes of facial patterns was re-
ported by Ekman, Levenson, and Friesen (1983). With replication, however,
the number of physiologic changes that reliably differ with facial expression
has diminished (Zajonc & Mclntosh, 1992). Most consistently produced are
lower heart rate (HR) increases during the disgust facial pattern than dur-
ing anger, fear, and sadness, and larger increases in finger temperature
during the anger configuration than in fear (Levenson, Ekman, & Friesen,
1990; see Shortt, Bush, McCabe, Gottman, & Katz, 1994, for a partial rep-
lication; cf Tourangeau & Ellsworth, 1979).
Levenson et al. (1990) collected information on participants' subjective
experiences during facial expressions. Facial configurations produced self-
reports of the associated emotion on a significant proportion of trials. Al-
though the muscle-by-muscle manipulation may cause some experimental
demand to report the associated emotion, the physiological differentiation
they found makes this less plausible. The connection of these outcomes to
facial patterning is bolstered by the finding that when participants' faces
best matched prototypical emotional expressions, the physiological differ-
ences were strongest, and self-reported emotions most evident. In one ex-
periment, Levenson et al. address the concern that physiological differences
are related to variations in the difficulty required to pose the expression,
rather than emotion produced as a result of the expression. No correlations
between difficulty and the magnitude of physiological change were signifi-
cant, and sorting the emotional expressions by difficulty did not produce
groups that matched sets found when grouping by autonomic nervous sys-
tem (ANS) data. Differential effort is thus not a likely explanation for the
ANS results.
Another example of facial actions initiating emotions is reported by
Zajonc et al. (1989). They had individuals pronounce /i (analogous to a
scowl) or o (relaxed face) for one minute. The phoneme i/ caused a sig-
nificant increase in forehead temperature, which has been associated with
negative feelings (e.g., Mclntosh, Zajonc, Vig, & Emerick, in press; Zajonc,

130 Mclntosh
Murphy, & Mclntosh, 1993; see discussion of the Vascular Theory of Emo-
tional Efference, below). Participants liked the o sound better and rated it
as more pleasant than a. In another study, participants repeated several
vowel sounds. Voicing/i increased forehead temperature, ah, and e (which
produces a smile analog) decreased it. The least liked sound was ti; ah and
e were liked most. The findings have been replicated, with the effect of
voicing the vowels on mood remaining when level of difficulty of speaking
them is statistically controlled (Mclntosh et al., in press).
Further evidence for an initiatory function of facial action is provided
by Hess et al. (1992). Participants' heart-rate patterns were consistent
across conditions in which they were asked to simply feel an emotion, sim-
ply display it, or do both. Individuals physiologically responded in the dis-
play-only condition in ways similar to the conditions in which they were to
feel the emotion. Also, they reported feeling emotions when they were
asked to display an emotion, but not to feel it. This study gives evidence
of within-valence differentiation, providing support for the categorical ver-
sion of the initiation hypothesis.
A new type of data supporting the initiating hypothesis has recently
emerged. Based on brain lateralization of emotions (Davidson, 1992), Schiff
and Lamon (1989; Schiff, Esses, & Lamon, 1992) examined the influence
of unilateral facial action. Coders blind to condition rated the number and
valence of emotion propositions used by participants in telling stories based
on mood sensitive pictures from the Thematic Apperception Test (TAT);
there is little chance that demand characteristics influenced these data. Par-
ticipants tended to feel sad or depressed when contracting the left comer
of their mouths. Following a right facial half-smile, they felt positive and
energetic. Although Wissing and Wessels (1992) replicated these results,
the effect may not occur among women.
In summary, support for the initiation hypothesis has been slower to
obtain than for correspondence and modulation. Recent work, however,
provides stronger evidence for categorical and dimensional versions using
a variety of manipulations, and both self-report and physiological outcomes.
Is Facial Action Necessary for Emotions?
The necessity hypothesis states that emotions cannot occur without fa-
cial activation. Although it has tended to be thought of as indicating that
facial movement is a necessary cause, it may also refer to facial changes
as being an inextricable component of emotions, without such expressions
themselves being causal. The latter is essentially a strong correspondence
hypothesis. Thus, the low association between categories of emotions and

Facial Feedback Hypotheses 131
facial patterns argues against this position. More generally, people experi-
ence emotions during times of facial paralysis, most commonly in REM
dreaming when there is striate muscle paralysis (see Fridlund, 1994, for
expanded examples). These occurrences make it difficult to maintain this
hypothesis.
However, one version has not been empirically refuted, largely because
it is difficult to test. Perhaps central nervous system (CNS) representations
of facial expressions can meet the hypothesized need for such patterns with-
out actual facial motion (Izard, 1977, 1990; Tomkins, 1962; cf Damasio's,
1994, somatic-marker hypothesis). Facial action or a neurocognitive repre-
sentation of facial action may be necessary for emotions. It is not clear
how this is functionally different than a completely central approach, except
that it may predict different consequences of damage to particular brain
regions, as alternate structures might be necessary for emotional experience
if representations are needed than if they are not.
Recurring Criticisms of Facial Feedback Studies
Fridlund (1994) has organized several of the frequent concerns regard-
ing facial feedback data. First, studies often contain implicit suggestions to
participants about how they should act (Buck, 1980). As detailed above,
recent work makes it unlikely that participants know what is expected and
respond in ways consistent with the hypotheses. Regarding dependent vari-
ables, the physiological data would be difficult for subjects to generate in
response to experimental demand, and the individual differences in re-
sponse to the manipulations do not follow a pattern expected if experi-
mental demand were the cause (Laird, 1984). Further, the latency measure
(Hess et al., 1992) and TAT procedure (Schiff & Lamon, 1989) are indi-
cators of effects not sensitive to demand. On the manipulation side, several
studies have used tasks that are unlikely to cue the participants as to the
expected response (e.g., Strack et al., 1988; Zajonc et al., 1989).
Second, Fridlund (1994) suggests that these just mentioned facial tasks
(e.g., holding a pen in one's teeth) are silly or embarrassing, and this is
what causes the changes in emotion. However, that both negative and posi-
tive affective changes result from the tasks makes this confound less plau-
sible. For example, why does voicing e make one feel silly, and voicing a
make one embarrassed, and not the reverse; why is reading a story with
many//'s more aversive than humorous?
The third concern is the confounding of negative faces with difficulty.
Because people can smile using only one muscle, the zygomaticus major
which pulls the corners of the mouth up, and negative patterns need acti-

132 Mclntosh
vation and coordination of more than one muscle, participants may like
the smile conditions better simply because they are easier (Fridlund, 1994).
This criticism is refuted by the studies that find reports of effort or difficulty
to be unrelated to the manipulation or emotion. Further, smiles utilizing
two muscles (adding the orbicularis oculi, which crinkles the corners of the
eyes) are more associated with positive affect than those using only the
one (Ekman, 1992). This should not be the case if effort is the cause of
more negative feelings.
Fourth, specific to those studies in which people modify their expres-
sions during ongoing emotional experiences, those told to modify their ex-
pressions may regulate their facial configuration by doing things (e.g.,
producing mental images) that also modify feelings. As there are several
affirmative studies that do not use this manipulation, it is not a strong cri-
tique of the hypotheses.
Summary: The Status of Facial Feedback Research
Recent studies address long-standing criticisms of facial feedback stud-
ies, and provide a firmer foundation on which to evaluate the propositions.
Especially in the case of the initiation hypothesis, recent work has dramati-
cally increased the supportive evidence. Further, additional support for
categorical effects is emerging. Nonetheless, the phenomena remain elu-
sive. Despite the stronger empirical bases for the effects, research on facial
feedback hypotheses leave a number of significant questions unanswered.
Scholars need to evaluate conditions under which facial movement alters
emotions, the mechanisms by which they do so, and the size of the effect
under different conditions and by varying mechanisms. Further, the impli-
cations of these phenomena need exploration. These issues constitute the
necessary next area of scientific focus.
POSSIBLE MECHANISMS FOR FACIAL EFFECTS ON EMOTIONS
The primary research question should now be how facial feedback in-
fluences emotions. One way to categorize mechanisms is by whether they
are part of an emotions theory, or whether the effects are due to a more
general process. The Jamesian camp holds that the influence of facial ac-
tion on emotions is the result of the emotions system (e.g., Ekman, 1992;
Izard, 1971; Tomkins, 1962). For the others, the theoretical focus is on more
general processes (e.g., Buck, 1980; Fridlund; 1994; Laird, 1974). With the
exception of Laird, those supporting the latter mechanisms view the effects

Facial Feedback Hypotheses 133
as minor. The Jamesian group views even minor effects as important, due
to their connection to understanding human emotional processes.
"Jamesian" Mechanisms: Facial Feedback as Part of Emotions Theories
CNS Connections. The dominant emotions-theory view of facial feed-
back effects sees facial movements as part of the package of events that
make up each emotion, or emotion program. These hypothesized programs
are viewed as coordinating people's responses to emotion-eliciting situ-
ations; they suppose a connection between the motor cortex and other parts
of the brain (especially the limbic region; Kelley & Stinus, 1984) involved
in physiological changes during emotions (Ekman, 1992; Levenson et al.,
1990). Activating one part of the response set is thought to stimulate the
others. If voluntary expressive actions simulate the innate emotional con-
figuration, then the expression-specific feeling state may be initiated (Izard,
1990). Neural tracts may lead from facial muscles or skin to an integration
center after the expression is produced, or the feedback may occur earlier,
for example in the firing of the muscles (Matsumoto & Lee, 1993). Leven-
son et al. (1990) think it likely that commands from the motor cortex to
the facial nucleus are accompanied by parallel signals to the ANS. Damasio
(1994) describes a neural network by which bodily states provide the input
for the subjective feelings associated with emotions.
Obviously, there is opportunity for much empirical specification. Stud-
ies of people with spinal lesions, lesions in the distal facial nerve tracts, or
temporary neuromuscular blocks would address these possibilities (Matsu-
moto & Lee, 1993). For example, people with spinal damage may be more
subjectively emotionally excitable since the injury, yet experience less so-
matic emotional responses (Bemlond, Nieuwenhuyse, Fasotti, & Schuer-
man, 1991). Emotions are multicomponential phenomena, involving
subjective state and a variety of physical events. Examining the effects of
damage to part of the system provides data on the organization of the
system. The report of Bermond et al. (1991) shows that aspects of emotions
can be altered by physical damage, but that the changes are not always
coordinated among components of emotions.
One concern with many CNS approaches is that they are based heavily
on the evidence for facial action creating specific, categorical, emotion-re-
lated patterns of ANS response. If facial action is conceived as starting
neurophysiological states corresponding to emotions, this presupposes the
existence of such patterns. However, it is difficult to say whether the ANS
changes caused by facial action are emotional. Not all methods of gener-
ating facial actions and of inducing emotions consistently show ANS pat-

134 Mclntosh
terns (Cacioppo, Klein, Berntson, & Hatfield, 1993; Zajonc & McIntosh,
1992)
Some theorists may not expect a high degree of consistency, however.
One theory indicates that only facial muscle patterns that closely match
highly specific configurations should induce emotions (Hager & Ekman,
1981; Levenson et al., 1990). Inexact or partial replications of emotional
expressions should not, thus, induce the emotion or cause the physiological
changes. In Izard's (1990) scheme, partial activation may cause the feeling
state, but only if other areas of the face are not sending signals for other
emotions. Note, however, that it may be that central processes are the key
to facial feedback effects, even if current categorical theories of emotions
are incorrect; it would simply require some modification of some CNS theo-
ries.
Work on the effects of unilateral muscular contractions is also based
on a CNS connection between facial action and emotion. This research
focuses on the subjective experience of emotion, rather than physiological
changes. Results are consistent with studies of brain lateralization of emo-
tion (Davidson, 1992). Integrating these methods will provide information
on brain functioning in emotions and on the parameters of CNS-face-emo-
tion connections.
Note that Izard's (1971, 1977) view differs from the purely central
views (Ekman, 1992; Levenson et al., 1990), as peripheral feedback plays
an important role. Izard (1990) notes that mechanoreceptors in facial skin
respond vigorously to facial muscle movement and are sensitive to the
slightest movement. It is this patterning that provides distinctions among
emotions. The next proposed mechanism also goes beyond purely central
mechanisms.
The Vascular Theory of Emotional Efference (I/TEE). Zajonc and col-
leagues (Zajonc & McIntosh, 1992; Zajonc et al., 1989, 1993) present a
manner by which affect can be altered via facial action that does not rely
on self-perception, cognitive appraisal, or activation of hypothesized CNS
emotion programs. Brain temperature is influenced by the cavernous sinus,
a venous configuration that envelops the internal carotid just before the
carotid enters the brain. Air inhaled through the nose cools the blood of
the cavernous sinus during normal breathing. Thus, nasal breathing cools
the brain (Kluger & D'Alecy, 1975). VTEE assumes that cooling of the
brain is hedonically positive (perhaps due to temperature effects on neuro-
chemical processes). Indeed, direct cooling of the hypothalamus in rats
leads to the same behavior demonstrated by rats receiving pleasurable elec-
trical stimulation (Berridge & Zajonc, 1991).
In humans, there is evidence for the postulated chain of events. Be-
cause facial actions that resemble emotional patterns influence the amount

Facial Feedback Hypotheses 135
of air inhaled via the nose (Mclntosh et al., in press), they may have af-
fective consequences. Repeatedly voicing//(producing a scowl) decreases
the amount of air individuals inhale through the nose, leads to increased
temperature at specific forehead points (a measure of brain temperature
increase; see Mclntosh et al., in press for discussion), and generates nega-
tive affect (Mclntosh et al., in press; Zajonc et al., 1989). Direct manipu-
lation of breathing (e.g., pinching the nose shut) alters forehead
temperature and affect (Mclntosh et al., in press), and changing the tem-
perature of air inhaled changes forehead temperature and affect (Zajonc
et al., 1989). This mechanism explains dimensional effects of facial move-
ment only, and does not predict ANS changes following facial action.
"Darwinian" Mechanisms: Facial Feedback as By-Product
Conditioning. Specific facial expressions may become conditioned to
certain subjective or physiological states. The typical presence of a smile
when one is happy, for example, may cause the smile to become a classically
conditioned stimulus for happiness. Involuntary expressions are likely to be
more strongly connected to emotional responses than voluntary ones, as
the former will more consistently be tied to the emotion; voluntary expres-
sions may be formed, exaggerated, and suppressed based on social context
(Buck, 1980). This view can account for facial effects on subjective expe-
rience and most physiological changes, both dimensional and categorical
effects, and modulation and initiation. Further, it is also consistent with
the differences found in between-subjects vs. within-subjects tests of the
hypothesis (Buck, 1980). Two findings suggest that conditioning is not the
sole explanation for facial feedback data, however. First, it is not clear using
this explanation why a muscle contraction on the left side of the body
evokes a different subjective state than the same movement on the right
side, given that movements on both sides of the face are consistently paired
with the state. Second, it is difficult to consider pairing of emotion and
expression to be the sole reason for the newer data supporting categorical
effects of facial action when there is such variable correspondence between
expression and discrete emotions.
Self-Perception. A second approach that predicts emotional effects of
facial action is self-perception theory (Bem, 1967). From this perspective,
the face provides data to the person regarding what he or she is feeling.
Findings for both initiation and modulation are consistent with this model.
Because this model predicts categorical effects, those results are especially
important (Duclos et al., 1989); previous difficulty in finding them has
caused some to doubt this mechanism (Winton, 1986). The effects of uni-

136 Mclntosh
lateral contractions and emotion-analog tasks (e.g., pen holding) are some-
what inconsistent with this theory. For example, Strack et al. (1988) point
out that their findings using oral pen holding counter a self-perception in-
terpretation, because subjects' attention is drawn away from the face, and
the facial configuration is constrained which should lead to participants dis-
counting its effect.
Dramaturgy. Fridlund (1994) describes a dramaturgical route by which
facial action could lead to altered emotions. People may use sensations
from their faces to assume a role. They then might recruit memories or
images to induce the emotion consistent with that role. This shares with
the self-perception mechanism the idea of the face as data used by the
person, but instead of the information being related to how one is feeling,
it is instead relevant to how one should be acting. For example using an
example from Laird's (1974) study, when a person is "instructed to move
facial muscles to form a smile at a Ku Klux Klan slide [this person] is
being asked implicitly to take an amused stance toward them" (Fridlund,
1994, pp. 178-179). If one forms an angry face, it follows, one will take on
an angry role, which may involve acting angry by making a fist, yelling, and
perhaps thinking angry thoughts.
Note that Fridlund (1994) separates the behavior from the emotion;
"acting" angry is not "is" angry. This may be a problematic distinction, in
that the action may lead to experience, even if experience incongruent with
the socially-expected action is suggested. In one study, subjects asked under
hypnosis to display one emotion and experience another reported being
unable to experience one incongruent with their expressions (Pasquarelli
& Bull, 1951). Nonetheless, this proposal suggests that emotional behavior,
at least, can be influenced by facial patterning.
This analysis points to the necessity of understanding the social context
in which the facial action occurs. Unresolved here are results of conflicting
social requests. It would be interesting to see the facial expressions, sub-
jective reports, and ANS changes in people asked to display anger, yet be
sad. How would role taking influence their responses? What effect would
the inner sadness have on the behavior? What effect would the anger dis-
play have on the felt sadness?
Multiple Causality
These different proposals should not be viewed as mutually exclusive.
The influence of the face on emotions is likely multiply determined. No
single proposed mechanism easily accounts for the variety of findings. For
example, VTEE does not address asymmetry in brain activity associated

Facial Feedback Hypotheses 137
with certain smiles, but does uniquely predict differences in temperature
and breathing associated with affective changes (see Mclntosh et al., in
press). Part of the confusion in the literature is probably due to different
mechanisms contributing differentially to varying findings. There may be
different mechanisms for categorical vs. dimensional effects, for physiologi-
cal vs. subjective effects, and for the combinations of these. For example,
perhaps only certain highly precise facial patterns activate the CNS mecha-
nisms producing categorical physiological effects, yet less exact emotion-
analogous configurations cause changes in subjective valence due to
breathing-related temperature changes (VTEE). Some mechanisms might
be more powerful, or more general than others. Possibly the primary cause
varies by emotion, with disgust expressions inducing associated feelings via
activation of a central emotions program, and smiles creating positive feel-
ings due to classical conditioning, for example. Further investigation into
all the proposed mechanisms is necessary, as is careful consideration of
which effects are connected with which methods and outcomes. Learning
more about the mechanisms of facial feedback will help us to understand
emotions processes, point to influences on emotions (including sources of
individual differences), and will assist in considering any applications of
the effect. However, for some purposes, the specific mechanism is not
highly important if the effect is large enough (Capella, 1993; Izard, 1990).
Even if the influence of the face is due to classical conditioning or role
playing, the effect exists.
IMPLICATIONS OF FACIAL EFFECTS ON EMOTION
Implications for Emotions Theories
Much of the drive in investigating facial feedback comes from those
in the Jamesian camp who view it as an integral factor in human emotional
processes. The findings regarding correspondence are important here, as
they demonstrate that the face is associated with emotion, not just social
communication. Indeed, this may be all that is needed for the data to be
consistent with some theories (see Tomkins, 1962, 1981).
The evidence in support of the modulation hypothesis is quite impor-
tant to several emotions theories, specifically that of Izard (1977, p. 64),
in which facial expression can modulate experience only when the expres-
sion is "sufficiently veridical." These results thus maintain the plausibility
of Izard's theory. The CNS theories also remain plausible, especially as the
evidence for the initiation function grows. The work supporting categorical
effects is especially important, as these are predicted by these theories. As

138 Mclntosh
these findings are historically weaker and more controversial, further work
should address clearly the existence of categorical effects. As our knowl-
edge of neuropsychological processes related to emotions grows, it is likely
the CNS theories will be modified. Data from the facial feedback studies
are useful in that they might point to neural systems that should be exam-
ined for their emotional function. Given recent advances in neuropsy-
ehological methods, it is an opportune time to do the empirical work
necessary to specify the CNS theories. The ambiguity of ANS specificity
may be less relevant if CNS patterns associated with emotions can be iso-
lated.
Facial feedback findings are also relevant to a central issue in emotions
theories: Is cognition necessary for emotions? Facial feedback may be an
example of noncognitive processes influencing emotion, but this depends
on one's definition of cognition, and on which mechanisms are responsible
for the effect. VTEE may not require cognition. Some conditioning and
CNS theories may require only the simplest kind, and then only if cognition
is defined to include immediate perceptual knowledge (see Buck, 1980,
1994). The dramaturgical and self-perception routes may need the complex
information processing of higher-level cognition. The most recent work
(e.g., Mclntosh et al., in press; Schiff & Lamon, 1989, Strack et al., 1988,
Zajonc et al., 1989) casts doubt on the necessity of cognitive appraisal.
Physical processes appear enough to generate feelings. Just as evidence of
more central processes and cognitive and social influences makes a com-
pletely peripheral/facial theory of emotion untenable, so the facial feedback
data makes a completely cognitive theory less plausible (Berkowitz &
Devine, 1995).
Finally, the work in facial feedback is relevant to the question of the
existence of basic emotions (see Duclos et al., 1986). Work on the cate-
gorical versions of the hypotheses often presumes the existence of basic
emotions (e.g., Ekman, 1992). Evidence regarding categorical effects should
be taken into account in the discussions about the nature of emotions.
The relevance of these findings for understanding basic processes of
emotions notwithstanding, facial effects on emotion are more likely impor-
tant the larger their effects are. Matsumoto (1987) reported a small to mod-
erate effect size (.343), and both proponents and critics of the hypotheses
consider the results as consistent with their views. More affirmative studies
have been published since. However, data from modulation studies indicate
that the consequences of facial patterning are much smaller than the in-
fluence of the stimuli (e.g., Kraut, 1982; Laird, 1974; Lanzetta et al., 1976;
Tourangeau & Ellsworth, 1979). "Turning your frown upside-down" may
make you feel better than not doing so, but isn't likely to make you feel
happy if your new car has just been rear-ended. However, when the context

Facial Feedback Hypotheses 139
is less powerful than a car accident, the initiation function of facial move-
ment may play an important role. Once an emotion state is, even weakly,
initiated, memories, appraisals, and behaviors may be influenced to be con-
sistent with this state, and amplify the original weak state caused by the
facial pattern. Below, implications beyond emotion theories are considered.
Implications for Emotion Regulation
Attention toward how individuals regulate or control their emotions
has increased recently (see Wegner & Pennebaker, 1993). The ability of
voluntary facial action to modulate and initiate emotions indicates that in-
dividuals can regulate their emotions by controlling their facial movements.
Athletes who "put on a game face" of anger, may increase the aggression
with which they perform.
This has both psychological and social applications. Wikan (1989), for
example, describes the role of expression as personal and social emotion
management in North Bali. There is social pressure to smile during tragedy.
Explicit in this practice is the understanding of expressions as causes of
inner feelings. If one is socially pressured to smile in the aftermath of
trauma, it may decrease the feelings of depression that follow. On the other
hand, being coerced to display one emotion while feeling another creates
conflict (Hochschild, 1983); facial feedback research may shed light on what
occurs in these circumstances. Understanding the individual and social uses
of facial feedback to regulate emotions has importance for both everyday
emotion regulation and psychotherapy (Gellhorn, 1964; see Izard, 1971, for
an extended discussion). This may be one method, for example, for people
to begin to alter affective responses to certain circumstances.
Implications for Social Processes
One active area of interest in facial feedback is in research on emo-
tional contagion, or socially-induced affect (Hatfield, Cacioppo, & Rapson,
1994; Levenson, in press, Mclntosh, Druckman, & Zajonc, 1994). This is
fitting. Strong early evidence for facial feedback came from studies of Lan-
zetta and colleagues (e.g., Lanzetta et al., 1976), whose initial work was
based in part on research demonstrating socially-induced affect among
monkeys (see Mclntosh et al., 1994; McHugo & Smith, this issue). Facial
mimicry is one way in which emotions may be socially shared.
The early data led to exploration of the role of the face in the vicarious
instigation of affect. Vaughan and Lanzetta (1981) found that participants
who amplified their facial responses to the other's pain showed increased

140 Mclntosh
EDR and HR compared to those who inhibited facial movement, or were
given no instructions. Vicarious emotional experience appeared mediated
by facial actions of the participant. Combined with the evidence that people
mimic the emotional expressions of others, facial feedback provides a
mechanism for one person's emotions to influence another's (e.g., Bush et
al., 1989; Hatfield et al., 1994; Mclntosh et al., 1994). Forming one's face,
either automatically or intentionally, into the same pattern as another per-
son's may, via facial feedback, lead to overlap of subjective or physical emo-
tional experience. Dyadic rapport is another interpersonal process that may
rely in part on facial feedback (Capella, 1993). Mutual mimicry, and resul-
tant shared emotions may smooth social interaction. Rapport and empa-
thy-and therefore facial feedback–are important for a wide variety of
normal social functions (Hatfield et al., 1994; Levenson, in press; Mclntosh
et al., 1994).
Difficulties in such processes may be associated with pathology. Early
problems in mimicking others may be associated with autism, and may lead
to some of the social deficits autistic individuals experience (Rogers & Pen-
nington, 1991). Facial feedback is involved in this, as, via mimicry, it could
provide the subjective experience of others' emotions that allows normal
individuals to develop understandings of others. Neurological problems
with the somatic feedback system may be associated with sociopathy
(Damasio, 1994).
Finally, social context probably influences facial feedback processes in
at least two ways. First, how much individuals express, and what they ex-
press is strongly influenced by who they are with (Fridlund, 1994); the facial
feedback hypotheses suggest that this will affect the emotions people have
in such circumstances. For example, when subjects are being observed dur-
ing electric shock, their facial movements, self-reported feelings, and physi-
cal arousal are all lower than when alone (Kleck et al., 1976). Second, social
cues are another influence on emotions; when they are strong they may
outweigh effects of facial feedback.
DIRECTIONS AND QUESTIONS
Research has demonstrated that the face can influence emotions. And,
there are a number of potentially important implications of this phenome-
non. However, the area is in need of development and refinement. One
issue for subsequent work is discerning the circumstances under which the
effects are most powerful. First, overall effect size should be reevaluated,
adding the recent findings to those used in Matsumoto's (1987) metaana-
lysis. In doing this work, the notion of one study, one vote should be

Facial Feedback Hypotheses 141
avoided; certain manipulations, measures, and procedures are more con-
vincing than others. Further, an examination of the effects for the different
hypotheses would be useful. The power of facial action may vary with in-
dividual differences, the facial manipulation used (muscle-by-muscle, ana-
log patterns, amplification), varying outcomes (memories, ANS, CNS,
subjective experience), and the social context.
Some individual differences have been found relating to sensitivity to
internal cues (Laird, 1984). Katkin (1985), for example, has found clear
individual differences in the ability to consciously detect heartbeats. People
who possess more accurate autonomic self-perception may be more likely
to experience facial feedback effects–physical cues may be more powerful
for them. A quite consistent finding is that when environmental cues are
minimized, men detect physiological changes more readily than women
(Katkin, 1985; Pennebaker & Roberts, 1992). This suggests that facial feed-
back effects would be more powerful among men than women, especially
to the extent self-perception of somatic cues is responsible for the effects.
Compatible with this is Wissing and Wessels' (1992) finding that unilateral
facial contractions worked only for men.
However, much research finds effects of facial action in mixed-gender
samples, and Ohira and Kurono (1993) found effects in a sample of Japa-
nese women. Two possibilities are suggested. First, only some mechanisms
may be gender related; work should attempt to determine if mechanisms
of facial feedback differ across gender. Second, any gender difference may
be cultural, and by implication, so might some facial feedback effects (see
Pennebaker & Roberts, 1992 for discussion of possible causes of the gender
difference). In general, finding individual, gender, and cultural differences
in facial feedback and related effects may help untangle the mechanisms
responsible for the phenomena. If there are cultural differences, for exam-
ple, the mechanisms should be evaluated with respect to how well they can
account for these.
Katkin's (1985) work also suggests a situational effect on visceral self-
perception. Accuracy of heartbeat detection increases with physically or
psychologically-induced arousal. This suggests that in situations of high
arousal, somatic feedback effects (including facial) may be enhanced.
One potentially relevant difference in manipulation is whether the fa-
cial motion is voluntary or posed. Facial actions are under dual neurological
control. Voluntary and involuntary (or, perhaps, constructed vs. over-
leamed, see Fridlund, 1994) facial actions involve different neural paths
(Matsumoto & Lee, 1993). Izard (1990) claims that this is one reason why
studies in which the experimenter manipulates participants' faces show
lower effect sizes than those using spontaneous faces. He (Izard, 1981,
1990) suggests that for voluntary facial movements to elicit or regulate emo-

142 Mclntosh
tion strongly, the movements must be congruent with ongoing cognitive
processes.
How different are voluntary and involuntary patterns? Smith, McHugo,
and Lanzetta (1986) demonstrated that the pattern of muscle activation in
posed happiness, anger, and sadness were highly similar to that generated
during imagery of past experiences of the respective emotions; however,
during sadness, muscle activity over the brow was more evident in the im-
agery task than in the posed expression. Is this difference (and potential
ones of similar magnitude) enough to affect facial feedback effects? Some
say so. Hager and Ekman (1981) indicate that only faces that very closely
match basic emotion patterns will elicit emotions; effects of precision of
expression are discussed by Levenson et al. (1990). The differences in in-
nervation and muscle action between voluntary and manipulated facial ac-
tions may influence the extent of facial feedback. Future work needs to
continue to determine the degree differences in facial patterns related to
this manipulation, and evaluate whether this difference is significant for
facial feedback effects.
One important direction would be to further link facial feedback ef-
fects with data on more general somatic feedback effects (see Buck, 1984;
Fridlund, 1994). One example of this is Duclos et al. (1989), who examined
both facial and posture feedback. Another is Schiff and Lamon (1994), who
extended their work on unilateral facial contractions to squeezing a ball in
either the right or left hand. Their findings were parallel with those of the
facial contraction studies reported above. Most of the proposed mecha-
nisms for facial feedback are relevant to postural, and perhaps vocal feed-
back effects, for instance (Hatfield et al., 1994). For example, can breathing
as one does while angry cause feelings of anger by activating the emotions
program also initiated by an angry facial pattern? Expanding the research
more completely to the body will provide information on both the mecha-
nism and the power of facial feedback effects.
CONCLUSION
Recent evidence bolsters previous work demonstrating that facial ac-
tion corresponds to emotions, and can modulate and initiate emotional ex-
perience. There is still no evidence that facial action is necessary for
emotions. Evidence for categorical effects has grown. Support for the phe-
nomena is not always obtained, however. This is likely due to causes of
facial movement beyond emotional state (e.g., communicative intent), and
to influences on emotion in addition to facial feedback (e.g., cognitive ap-
praisal). Further, a lack of knowledge regarding the parameters of feedback

Facial Feedback Hypotheses 143
processes makes finding the effect more difficult. Future work may isolate
a more constrained process that is more reliable.
A focus on determining the contribution of various mechanisms to fa-
cial feedback effects is necessary. Knowledge of how facial actions influence
emotions will advance understanding of emotions processes in general, and
will clarify the role of the face in emotional and social domains. Such in-
formation will also assist in defining the parameters within which feedback
effects occur. Determining mediators and moderators of facial feedback
will provide a more solid foundation on which to examine both basic ques-
tions and applications of this phenomenon.
Even if the Jamesian mechanisms (i.e., those that include facial feed-
back as integral to emotional processes) are not supported, the existence
of facial feedback effects has significant implications. Facial influence on
emotions may play a particularly important role in an array of intrapersonal
and interpersonal dynamics (e.g., emotion regulation and socially-induced
affect). Research on these applications of facial feedback effects should
proceed in parallel to development of investigation of the basic phenomena.
A quarter-century of rapidly expanding research has laid the ground-
work for exploring complex and difficult questions about facial feedback
processes. To progress, the field must move past replication of demonstra-
tions of such phenomena to focused work on discerning their mechanisms,
parameters, and applications. This work has started, with several programs
of research now addressing some of these points. There is much more to
be done; the next 25 years should be an exciting time to investigate this
element of emotional and social processes.
REFERENCES
Adelmann, P. K., & Zajonc, R. B. (1989). Facial efference and the experience of emotion.
Annual Review of Psychology, 40, 249-280.
Bern, D. J. (1967). Self-perception: An alternative interpretation of cognitive dissonance
phenomena. Psychological Review, 74, 183-200.
Berkowitz, L., & Devine, P. G. (1995). Has social psychology always been cognitive? What is
"cognitive" anyhow? Personality and Social Psychology Bulletin, 21, 696-703.
Bermond, B., Nieuwenhuyse, B., Fasotti, L., & Schuerman, J. (1991). Spinal cord lesions,
peripheral feedback, and intensities of emotional feelings. Cognition and Emotion, 5,
201-220.
Berridge, K. C., & Zajonc, R B. (1991). Hypothalamic cooling elicits eating: Differential effects
on motivation and pleasure. Psychological Science, 2, 184-189.
Brown, S., & Schwartz, G. E. (1980). Relationships between facial electromyography and
subjective experience during affective imagery. Biological Psychology, ll, 49-62.
Buck, R. (1980). Nonverbal behavior and the theory of emotion: The facial feedback
hypothesis. Journal of Personality and Social Psychology, 38, 811-824.
Buck, R. (1984). The communication of emotion. New York: Guilford.

144 Mclntosh
Buck, R. (1994). Social and emotional functions in facial expression and communication: the
readout hypothesis. Biological Psychology, 38, 95-115.
Buck, R., Savin, V. J., Miller, R. E., & Caul, W. F. (1972). Nonverbal communication of
affect in humans. Journal of Personality and Social Psychology, 23, 362-371.
Bush, L. K., Barr, C. L., McHugo, G. J., & Lanzetta, J. T. (1989). The effects of facial control
and facial mimicry on subjective reactions to comedy routines. Motivation and Emotion,
13, 31-52.
Cacioppo, J. T., Bush, L. K., & Tassinary, L. G. (1992). Microexpressive facial actions as a
function of affective stimuli: Replication and extension. Personality and Social Psychology
Bulletin, 18, 515-526.
Cacioppo, J. T., Klein, D. J., Berntson, G. G., & Hatfield, E. (1993). The psychophysiology
of emotion. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotions (pp. 119-142).
New York: Guilford.
Cacioppo, J. T., Martzke, J. S., Petty, R. E., & Tassinary, L. G. (1988). Specific forms of
facial EMG response index emotions during an interview: From Darwin to the continuous
flow hypothesis of affect-laden information processing. Journal of Personality and Social
Psychology, 54, 592-604.
Cacioppo, J. T., Petty, R. E., Losch, M. E., & Kim, H. S. (1986). Electromyographic activity
over facial muscle regions can differentiate the valence and intensity of affective reactions.
Journal of Personality and Social Psychology, 50, 260-268.
Camras, L. A., Holland, E. A., & Patterson, M. J. (1993). Facial expression. In M. Lewis &
J. M. Haviland (Eds.), Handbook of emotions (pp. 199-208). New York: Guilford.
Cannon, W. B. (1927). The James-Lange theory of emotions: A critical examination and an
alternative theory. American Journal of Psychology, 39, 106-112.
Capella, J. H. (1993). The facial feedback hypothesis in human interaction: Review and
speculation. Journal of Language and Social Psychology, 12, 13-29.
Chovil, N. (1991). Social determinants of facial displays. Journal of Nonverbal Behavior, 5,
141-154.
Colby, C. A., Lanzetta, J. T., & Kleck, R. E. (1977). Effects of the expression of pain on
autonomic and pain tolerance responses to subject-controlled pain. Psychophysiology, 14,
537-540.
Damasio, A. R. (1994). Descartes' error:. Emotion, reason and the human brain. New York."
Putnam.
Darwin, C. (1955). The expression of the emotions in man and animals. New York: Philosophical
Library (original work published 1872).
Davidson, R. J. (1992). Anterior cerebral asymmetry and the nature of emotion. Brain and
Cognition, 6, 245-268.
Dimberg, U. (1988). Facial electromyography and the experience of emotion. Journal of
Psychophysiology, 2, 277-282.
Duclos, S. E., Laird, J. D., Schneider, E., Sextet, M., Stern, L., & Van Lighten, O. (1989).
Emotion-specific effects of facial expressions and postures on emotional experience.
Journal of Personality and Social Psychology, 57, 100-108.
Duncan, J., & Laird, J. D. (1977). Cross-modality consistencies in individual differences in
self-attribution. Journal of Personality, 45, 191-206.
Duncan, J., & Laird, J. D. (1980). Positive and reverse placebo effects as a function of
differences in cues used in self-perception. Journal of Personality and Social Psychology,
39, 1024-1036.
Ekman, P. (1992). Facial expressions of emotion: New findings, new questions. Psychological
Science, 3, 34-38.
Ekman, P. (1994). Strong evidence for universals in facial expression: A reply to Russell's
mistaken critique. Psychological Bulletin, 115, 268-287.
Ekman, P. Friesen, W. V., & Ancoli, S. (1980). Facial signs of emotional experience. Journal
of Personality and Social Psychology, 39, 1125-1134.
Ekman, P., Levenson, R. W., & Friesen, W. V. (1983). Autonomic nervous system activity
distinguishes between emotions. Science, 22L 1208-1210.

Facial Feedback Hypotheses 145
Ellsworth, P. C., & Tourangeau, R. (1981). On our failure to disconfirm what nobody ever
said. Journal of Personality and Social Psychology, 40, 363-369.
Fridlund, A. J. (1994). Human facial expression; An evolutionary view. New York: Academic
Press.
Fridlund, A. J., Schwartz, (3. E., & Fowler, S. C. (1984). Pattern recognition of self-reported
emotional state from multiple-site facial EMG activity during affective imagery.
Psychophysiology, 21, 622-637.
Gellhorn, E. (1964). Motion and emotion: the role of proprioception in the physiology and
pathology of the emotions. Psychological Review, 71, 457-472.
Hager, J. C. & Ekman, P. (1981). Methodological problems in Tourangeau and EIIsworth's
study of facial expression and the experience of emotion. Journal of Personality and Social
Psychology, 40, 358-362.
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. Cambridge:
Cambridge University Press.
Hess, U., Banse, R., & Kappas, A. (1995). The intensity of facial expression is determined
by underlying affective state and social situation. Journal of Personality and Social
Psychology, 69, 280-288.
Hess, U., Kappas, A., McHugo, G. J., Lanzetta, J. T, & Kleck, R. E. (1992). The facilitative
effect of facial expression on the self-generation of emotion. International Journal of
Psychophysiology, 12, 251-265.
Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. Berkeley,
CA: University of California Press.
Ianni, P., Stettner, L., & Freedman, R. R. (1986). Voluntary facial actions modulate
cardiovascular responses during angry emotion. Psychophysiology, 23, 443.
Izard, C. E. (1971). The face of emotion. New York: Appleton-Centrny-Crofts.
Izard, C. E. (1977). Human emotions. New York: Plenum.
Izard, C. E. (1981). Differential emotions theory and the facial feedback hypothesis of emotion
activation: Comments on Tourangean and Ellsworth's "The role of facial response in the
experience of emotion." Journal of Personality and Social Psychology, 40, 350-354.
Izard, C. E. (1990). Facial expressions and the regulation of emotions. Journal of Personality
and Social Psychology, 58, 487-498.
James, W. (1950). The principles of psychology. New York: Holt (original work published 1890).
Jones, S. S., Collins, K., & Hong, H. (1991). An audience effect on smile production in
10-month old infants. Psychological Science, 2, 45-49.
Katkin, E. S. (1985). Blood, sweat, and tears: Individual differences in autonomic
self-perception. Psychophysiology, 22, 125-137.
Kelley, A. E., & Stinus, L. (1984). Neuroanatomical and neurochemical substrates of affective
behavior. In N. A. Fox & R. J. Davidson (Eds.), The psychobiology of affective development.
Hiilsdale, N J: Eribaum.
Kleck, R. E., Vaughan, R. C., Cartwright-Smith, J., Vaughan, K. B., Colby, C. Z., & Lanzetta,
J. T. (1976). Effects of being observed on expressive, subjective, and physiological
responses to painful stimuli. Journal of Personality and Social Psychology, 34, 1211-1218.
Kluger, M. J., & D'Alecy, L. B. (1975). Brain temperature during reversible upper respiratory
bypass. Journal of Applied Physiology, 38, 268-271.
Kraut, R. E. (1982). Social presence, facial feedback, and emotion. Journal of Personality and
Social Psychology, 42, 853-863.
Kraut, R. E., & Johnston, R. E. (1979). Social and emotional messages of smiling: An
ethological approach. Journal of Personality and Social Psychology, 37, 1539-1553.
Laird, J. D. (1974). Self-attribution of emotion: the effects of expressive behavior on the quality
of emotional experience. Journal of Personality and Social Psychology, 29, 475-486.
Laird, J. D. (1984) The real role of facial response in the experience of emotion: A reply to
Tourangeau and Ellsworth, and others. Journal of Personality and Social Psychology, 47,
909-917.
Laird, J. D., Cuniff, M., Sheehan, K., Shulman, D., & Strum, G. (1989). Emotion specific
effects of facial expressions on memory for life events. In D. Kuiken (Ed.), Mood and

146 Mclntosh
memory: Theory, research and applications (Special Issue). Journal of Social Behavior
and Personality, 4, 87-98.
Lanzetta, J. T., Cartwright-Smith, J., & Kleck, R. E. (1976). Effects of nonverbal dissimulation
on emotional experience and autonomic arousal. Journal of Personality and Social
Psychology, 33, 354-370.
Lanzetta, J. T., & Kleck, R. (1970). Encoding and decoding of nonverbal affect in humans.
Journal of Personality and Social Psychology, 16, 354-370.
Larsen, R. J., Kasimatis, M., & Frey, K. (1992). Facilitating the furrowed brow: An unobtrusive
test of the facial feedback hypothesis applied to unpleasant affect. Cognition and Emotion.
6, 321-338.
Levenson, R. W., Ekman, P., & Friesen, W. V. (1990). Voluntary facial action generates
emotion-specific autonomic nervous system activity. Psychophysiology, 27, 363-384.
Manstead, A. S. R. (1988). The role of facial movement in emotion. In H. L. Wagner (Ed.),
Social Psychophysiology: Theory and Clinical Applications (pp. 105-129). New York: Wiley.
Matsumoto, D. (1987). The role of facial response in the experience of emotion: More
methodological problems and a recta-analysis. Journal of Personality and Social Psychology,
52, 769-774.
Matsumoto, D., & Lee. M. (1993). Consciousness, volition, and the neuropsychology of facial
expressions of emotion. Consciousness and Cognition, 2 237-254.
McCanne, T. R., & Anderson, J. A. (1987). Emotional responding following experimental
manipulation of facial electromyographic activity. Journal of Personality and Social
Psychology, 52, 759-768.
Mclntosh, D. N., Druckman, D., & Zajonc, R. B. (1994). Socially-lnduced Affect. In D.
Druckman & R. A. Bjork (Eds.), Learning, rememberin~ believing: Enhancing human
performance (pp. 251-276, 364-371). Washington, D.C.: National Academy Press.
Mclntosh, D. N., Zajonc, R. B., Vig, P. S., & Emerick, S. W. (in press). Facial movement,
breathing, temperature, and affect: Implications of the vascular theory of emotional
efferenee. Cognition and Emotion.
Ohira, H., & Kurono, K. (1993). Facial feedback effects on impression formation. Perceptual
and Motor Skills, 77, 1251-1258.
Pasquarelli, B., & Bull N. (1951). Experimental investigation of the mind-body continuum in
affective states. Journal of Nervous and Mental Disease, 113, 512-521.
Pennebaker, J. W., & Roberts, T. (1992). Toward a his and hers theory of emotion: Gender
differences in visceral perception. Journal of Social and Clinical Psychology, 11, 199-212.
Rogers, S. J., & Pennington, B. F. (1991). A theoretical approach to the deficits in infantile
autism. Development and Psychopathology, 3, 137-162.
Rutledge, L. L., & Hupka, R B. (1985). The facial feedback hypothesis: methodological
concerns and new supporting evidence. Motivation and Emotion, 9, 219-240.
Schachter, S., & Singer, J. E. (1962). Cognitive, social, and physiological determinants of
emotional state. Psychological Review, 69, 379-399.
Schiff, B. B., Esses, V. M., & Lamon, M. (1992). Unilateral facial contractions produce mood
effects on social cognitive judgments. Cognition and Emotion, 6, 357-368.
Schiff, B. B., & Lamon, M. (1989). Inducing emotions by unilateral contraction of facial
muscles: A new look at hemispheric specialization and the experience of emotion.
Neuropsychologia, 27, 923-935.
Schiff, B. B., & Lamon, M. (1994). Inducing emotions by unilateral contraction of hand
muscles. Cortex, 30, 247-254.
Shortt, J. W., Bush, L. K., McCabe, J. L. R., Gottman, J. M., & Katz, L. F. (1994). Children's
physiological responses while producing facial expressions of emotions. Merrill-Palmer
Quarterly, 40, 40-59.
Smith, C. A. (1989). Dimensions of appraisal and physiological response in emotion. Journal
of Personality and Social Psychology, 56, 339-353.
Smith, C. A., McHugo, G. J., & Lanzetta, J. T. (1986). The facial muscle patterning of posed
and imagery-induced expressions of emotion by expressive and nonexpressive posers.
Motivation and Emotion, 10, 133-157.

Facial Feedback Hypotheses 147
Strack, R., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of facial
expressions: A non-obtrusive test of the facial feedback hypothesis. Journal of Personality
and Social Psychology, 54, 768-777.
Tassinary, L. G., & Cacioppo, J. T. (1992). Unobservable facial actions and emotion.
Psychological Science, 3, 28-33.
Tomkins, S. S. (1962). Affect, imagery, consciousness: Vol. 1. The positive affects. New York:
Springer.
Tomkins, S. S. (1981). The role of facial response in the experience of emotion: A reply to
Tourangeau and Ellsworth. Journal of Personality and Social Psychology, 40, 355-357.
Tourangeau, R., & EIIsworth, P. C. (1979). The role of facial response in the experience of
emotion. Journal of Personality and Social Psychology, 37, 1519-1531.
Vaughan, FL B., & Lanzetta, J. T., (1981). The effect of modification of expressive displays
on vicarious emotional arousal. Journal of Experimental Social Psychology, 17, 16-30.
Wegner, D. M., & Pennebaker, J. W. (Eds.) (1993). Handbook of mental control New Jersey:
Prentice-Hall.
Wikan, U. (1989). Managing the heart to brighten face and soul: Emotions in Balinese
morality and health care. American Ethnologist 16, 294-312.
Winton, W. M. (1986). The role of facial response in self-reports of emotion: A critique of
Laird. Journal of Personality and Social Psychology, 50, 808-812.
Wissing, M., & Wessels, S. (1992). Unilateral Lower Facial Contractions and Emotional
Experience: An Affirmative Study. 25th International Congress of Psychology Brussels,
Belgium, July 19-24.
Zajonc, R. B., & McIntosh, D. N. (1992). Emotions research: Some promising questions and
some questionable promises. Psychological Science, 3, 70-74.
Zajonc, R. B., Murphy, S. T., & Inglehart, M. (1989). Feeling and facial efference: Implications
of the vascular theory of emotions. Psychological Review, 96, 395-416.
Zajonc, R. B., Murphy, S. T., & McIntosh, D. N. (1993). Brain temperature and subjective
emotional experience. In M. Lewis & J. M. Haviland (Eds.) Handbook of emotions (pp.
209-220). New York: Guilford.
Zuckerman, M., Klorman, R., & Larrance, D. T. (1981). Facial, autonomic, and subjective
components of emotion: The facial feedback hypothesis versus the
externalizer-internalizer distinction. Journal of Personality and Social Psychology, 41,
929-944