The Community of Practice:Theories and methodologies [609150]

The Community of Practice:Theories and methodologies
in language and genderresearch
JANET HOLMES
School of Linguistics andApplied Language Studies
Victoria University of Wellington
Wellington, New Zealand
[anonimizat]
MIRIAM MEYERHOFF
Department of Linguistics
University of Hawaii
Honolulu, HI 96822
[anonimizat]
ABSTRACT
Thisarticleprovidesanintroductiontothisissueof LanguageinSociety by
exploring the relationship of the concept of Community of Practice (CofP)torelatedtermsandtheoreticalframeworks.Thecriterialcharacteristicsandconstitutive features of a CofP are examined; the article points out how aCofPframeworkisdistinguishedfromothersociolinguisticandsocialpsy-chologicalframeworks,includingsocialidentitytheory,speechcommunity,socialnetworkandsocialconstructionistapproaches.(CommunityofPrac-tice,speechcommunity,gender,sex,socialpractice,ethnographicsociolin-guistics, discourse analysis)
The term “Community of Practice” (CofP) has recently shouldered its way into
thesociolinguisticlexicon.Thepurposeofthisissueof LanguageinSociety isto
provideanalysesoflanguagevariation,discourse,andlanguageusethatillustratethepotential(andalsothelimits)ofthisconceptasatheoreticalandmethodolog-ical basis for inquiry.
It is not generally helpful to add a term to one’s field unless it is intended to
serve some demonstrably useful purpose. The term “Community of Practice”bearsastrongsimilaritytotheexistingterm“speechcommunity ”–aconceptthat
has proved to be a productive and useful tool for research into the orderly het-erogeneity of language in its social setting; thus it must be shown how the CofPinsomewaytakesusfarthertowardourgoalofunderstandingtheconstraintsonnatural language variation.
Inaddition,somesociolinguistsmayseeintheCofPatoolforthedescription
of language variation that bears a strong resemblance to fundamental principlesLanguage in Society 28, 173–183. Printed in the United States ofAmerica
© 1999 Cambridge University Press 0047-4045 /99 $9.50 173
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

of social identity theory. The distinction between intergroup and interpersonal
identitieshasbeenthebasisforsocialpsychologicalresearchformorethantwodecades.Itmust,therefore,bedemonstratedhowthenotionoftheCofPbuysussomethingmorethansocialidentitytheorydoes,andhowitcanbeofdirecthelpin understanding human behavior, and particularly linguistic behavior.
The contributors to this issue all present work that can be characterized as
researchintotherelationshipbetweenlanguageandgender,andallofthemworkwithconversationaldata;however,thegroupswithwhichtheyworkdifferqual-itatively.ThesequalitativedifferenceswillilluminatethecontributionthataCofPanalysiscanmaketothestudyoflanguageandsociety.Thechiefpurposesofthisintroductory article are to exemplify communities of practice, and to provide aprincipled basis for distinguishing the CofP from similar concepts such as thespeech community, social networks, and social identity .
The term “Community of Practice” was introduced to language and gender
research by Eckert & McConnell-Ginet 1992. Following Lave &Wenger 1991,they defined a CofPas follows:
An aggregate of people who come together around mutual engagement in an
endeavor. Ways of doing things, ways of talking, beliefs, values, power rela-tions–inshort,practices–emergeinthecourseofthismutualendeavor.Asasocialconstruct,aCofPisdifferentfromthetraditionalcommunity,primarilybecauseitisdefinedsimultaneouslybyitsmembershipandbythepracticeinwhich that membership engages. (1992:464)
This definition suggests that the concept of a CofPis a dynamic, rich, and com-
plexone.Itemphasizesthenotionof“practice”ascentraltoanunderstandingofwhy the concept offers something different to researchers than the traditionalterm “community” – or, in the context of sociolinguistic research, more thanconcepts like “speech community” and “social network.”
For Lave & Wenger 1991, the CofP is one component of a social theory of
learning,andWenger1998usesittocritiquetraditionalmodelsoflearning.These,heargues,abstractlearnersfromtheirnormalinteractionalcontexts;theyrequirelearners to assimilate material that the teachers have selected in an artificial
environment, the classroom. Wenger suggests instead that learning is a naturalandinevitableaspectoflife,andafundamentallysocialprocess.HeregardstheconceptofCofPasameansofexaminingonenaturalmethodoflearningwhich,inmanyrespects,resemblesanapprenticeship.Theprocessofbecomingamem-berofaCofP–aswhenwejoinanewworkplace,abookgroup,oranewfamily(e.g.throughmarriage)–involveslearning.WelearntoperformappropriatelyinaCofPasbefitsourmembershipstatus:initiallyasa“peripheralmember,”laterperhaps as a “core member” (or perhaps not – one may choose to remain a pe-ripheral member). In other words, a CofPinevitably involves the acquisition ofsociolinguistic competence.
1JANET HOLMES & MIRIAM MEYERHOFF
174 Language in Society 28:2 (1999)
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

The CofPis one way of focusing on what members do: the practice or activ-
ities that indicate that they belong to the group, and the extent to which theybelong. The practice or activities typically involve many aspects of behavior,including global or specific aspects of language structure, discourse, and inter-actionpatterns.Theobviousappealofthisapproachisthatitoffersthesociolin-guist a framework of definitions within which to examine the ways in whichbecomingamemberofaCofPinteractswiththeprocessofgainingcontrolofthediscourse appropriate to it. By emphasizing a process in which apprentices ab-sorb attitudes to situations and interlocutors – and in which they learn how tomodifytheirlinguisticandotherbehaviors,insuchawayastofeedperceptionsofselfandother–theCofPalsohasobviousattractionsforsocialpsychologists.
DIMENSIONS OF A COMMUNITY OF PRACTICE
Wenger (1998:76) identifies three crucial dimensions of a CofP:
(1) a. Mutual engagement.
b. Ajoint negotiated enterprise.c. Ashared repertoire of negotiable resources accumulated over time.
Wewillexemplifyeachofthesecriteria,drawingonrecentresearchonlanguage
in New Zealand government policy units (Holmes 1997, Holmes et al. 1999).
Mutual engagement
Thistypicallyinvolvesregularinteraction:Itisthebasisfortherelationshipsthat
make the CofPpossible. People who work together in policy units typically in-teract regularly: casually, as they pass in the corridor or share morning tea; in-tensively, in pairs or small groups to discuss particular projects; and, in a unit,comprehensively, as a large group which meets once a week to discuss moregeneral issues.
Joint enterprise
Thisreferstoaprocess:Thejointenterpriseisnotjustastatedsharedgoal,buta
negotiatedenterprise,involvingthecomplexrelationshipsofmutualaccountabil-ity that become part of the practice of the community (Wenger 1998:80). Thestated goal of the policy units we researched was summarized in their missionstatements,whichspecifiedtheirroleincontributingtothedevelopmentofgov-ernmentpolicy.However,inWenger’stermstheyconstitutedaCofP,sincemem-bers of the policy units were engaged in an ongoing process of negotiating andbuilding their contributions toward the larger enterprise. These negotiations re-flectedmembers’understandingoftheirpersonalroleswithintheinstitution,andthey are one characteristic that defines them as a CofP.
Fromasociolinguist’sperspective,thischaracterizationofajointenterpriseor
shared goal appears very general. While it may be satisfactory for a theory oflearning, it seems likely that what precisely constitutes a shared goal or jointenterprisewillhavetobespecifiedmorefullyinorderforthenotionoftheCofPTHE COMMUNITY OF PRACTICE
Language in Society 28:2 (1999) 175
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

to be useful to a wider range of research programs in the social sciences. Mey-
erhoff’s article in this collection explores in more detail the need for a specificshared goal.
Shared repertoire
Overtime,thejointpursuitofanenterpriseresultsinasharedrepertoireofjoint
resources for negotiating meaning (Wenger 1998:85). This includes linguisticresources such as specialized terminology and linguistic routines, but also re-sources like pictures, regular meals, and gestures that have become part of thecommunity’s practice. In the New Zealand policy units, for example, we ob-servedregulargreetingritualsbetweenmembers,understandingsabouthowmuchsocialtalkwastolerableinvaryingcontexts(Holmes1998b),andpreferredwaysof coming to decisions in meetings in different units. These linguistic manifes-tationsofasharedrepertoireprovideanespeciallyfruitfulsourceofinsightsforthe sociolinguist and discourse analyst.
Asnotedabove,theprogressivenatureofaCofPmeansthatindividualmem-
bership in a CofPwill differ. Some people will be core members, and some pe-ripheralmembers.Thebasisofthisvariationliesinhowsuccessfullyanindividualhas acquired the shared repertoire, or assimilated the goal(s) of the joint enter-prise, or established patterns of engagement with other members. In a similarvein,Wengerproposes(1998:130–31)thatthecriterialcharacteristicsofaCofPare instantiated through a number of more specific features:
• Sustained mutual relationships – harmonious or conflictual.
• Shared ways of engaging in doing things together.• The rapid flow of information and propagation of innovation.• Absenceofintroductorypreambles,asifconversationsandinteractionswere
merely the continuation of an ongoing process.
• Very quick setup of a problem to be discussed.• Substantial overlap in participants’descriptions of who belongs.• Knowingwhatothersknow,whattheycando,andhowtheycancontribute
to an enterprise.
• Mutually defining identities.• The ability to assess the appropriateness of actions and products.• Specific tools, representations, and other artifacts.• Local lore, shared stories, inside jokes, knowing laughter.• Jargonandshortcutstocommunicationaswellastheeaseofproducingnew
ones.
• Certain styles recognized as displaying membership.• Ashared discourse that reflects a certain perspective on the world.
These features present a wealth of opportunity to interested researchers. With
appropriate operationalization, they lend themselves to developing an index ofthe distinctiveness of different communities of practice.The articles in this col-JANET HOLMES & MIRIAM MEYERHOFF
176 Language in Society 28:2 (1999)
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

lection that deal with clearly defined Communities of Practice begin to specify
the relative and absolute importance of these features in the theory. Thus, forresearchers interested in the relationship between language and society, thesefeaturesprovideabasisforexploringtheutilityoftheCofPmodelinrelationtoparticular communities.
This may generate very practical outcomes. For instance, a consistent set of
features forming the basis for comparison between Communities of Practice –allowing us to specify the degree to which they are similar or different – showsconsiderablepromiseforresearchoninteractionintheworkplace.Theextenttowhichpracticesatoneworkplacedifferfromthoseatanotherhasimplicationsforpeople who join these workplaces, and also for outsiders who want to interactwith those members effectively.
SomeoftheNewZealandpolicyunitsfromwhichwecollecteddataprovided
evidenceofmeetingalargenumberofthesefeaturesofaCofP.Forexample,thefeaturesoflackofpreamblesandrapidsetting-upofproblemsweredisplayedinthe way a manager could give instructions to an administrative assistant.
(2) Senior policy analyst Greg enters office of administrative assistant Jo.
G: can you ring these people for me Jo 1set up a meeting forTuesday afternoon
J: sure no problem [ pause] what time
G: 2 o’clock
J: fine
G: oh and and book the committee room
ItisimportanttokeepthethreecriterialcharacteristicsofaCofP(1a–c)distinct
from the 14 constitutive features listed above. CofPs will display the latter todifferentextents.Forexample,notallgovernmentpolicyunitsinourdatasharedthe property of having “inside jokes.”
2
CONTRASTING COFPWITH OTHER MODELS
AND THEORETICAL FRAMEWORKS
HavingoutlinedthepropertiesandprovenanceoftheCofP,letusnowattemptto
distinguish the CofP more precisely from other sociolinguistic and social psy-chological frameworks.
Social identity theory
The notion of social identity was first articulated by Henri Tajfel, and it has
subsequentlybeentestedinnumerousexperimentalandqualitativestudiesinthefieldofsocialpsychology.AsTajfelproposedit(1978:44),socialidentitytheoryholdsthatindividuals’socialbehaviorisajointfunctionof(a)theiraffiliationtoa particular group identity that is salient at that moment in the interaction, and(b)theirinterpretationoftherelationshipofone’singrouptosalientoutgroups.InTajfel’s theory, an individual’s social – or intergroup – categorizations are cog-nitive tools, the function of which is to help an individual make sense of andfacilitatesocialaction(Tajfel&Turner1986:15).Tajfel(1978:43)sawinterper-THE COMMUNITY OF PRACTICE
Language in Society 28:2 (1999) 177
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

sonalandintergroupidentitiesasacontinuum,althoughotherwaysofmodeling
therelationshipbetweenthemhavesubsequentlybeenproposed.Giles&Coupland1991 suggest that personal and social identities are independent of each other.Meyerhoff & Niedzielski 1994 represent them as interdependent, but they char-acterize the dependence in a non-scalar fashion.
These identities are, however, taken to be highly abstract representations
(Abrams1996:147)whichmustbeconstructedthroughsocialprocessesandovertime.Tajfelsuggested(Tajfel&Turner1986:16–17)thattheprimaryprocessbywhich a positive social identity is established is through comparison with othergroups. Naturally, many social behaviors are the basis for this process of com-parison, and language is just one of the ways in which an individual develops astrongsocialidentity.Socialidentitytheoryisakeyprinciplebehindcommuni-cativeaccommodationtheory(Gilesetal.1987),andtothisextentithasbecomea dimension regularly employed in the investigation and interpretation of lan-guage variation.
The speech community
Thenotionofthespeechcommunityisafundamentaloneinlinguistics.Itisthe
unspoken basis of most linguistics research, and its relevance has been articu-lated with the greatest precision for the study of language variation and change.But despite its significance to the study of language, there is no single, agreed-upon definition of the speech community (see Santa Ana & Parodi 1998 for arecent discussion).
Labov(1972:121)definedthespeechcommunityasagroupofspeakerswho
participate in a shared set of norms, where “these norms are observed in overtevaluativebehavior,andbytheuniformityofabstractpatternsofvariation.”
3In
this way, the notion of a speech community is compatible with inter-individualvariation.Foranylinguisticvariable,itmaybethecasethatnoindividual’sper-formancerealizestherangeofpossiblevariants,butratherthattheheterogeneityofallindividuals’linguisticbehaviorshowsconsistenteffectsacrossanorderedset of domains.When it can be shown that this kind of orderliness is unlikely toresult from chance, then it can be inferred that the speakers share underlyingevaluations of the social or stylistic significance of the possible variants.
Gumperz 1971 provided a more interactional definition of the speech com-
munity, focusing on the frequency and quality of interactions among members,wherethequalityofinteractionisdefinedpartlyintermsofcontrastswithothers.Preston1989takesthisqualitativeslantevenfurther;insomeofhisresearch,theboundaries of speech communities are described according to whether speakersshare the same beliefs about their own language and the language of outgroups.
Eventhoughthesedefinitionsofthespeechcommunityaresomewhatdiverse,
thereisacommonthreadinthem:thesensethataspeechcommunityisawayofbeing.One’smembershipinaspeechcommunitydependsonsocialorbehavioralproperties that one possesses. The dimensions along which a CofPdiffers fromJANET HOLMES & MIRIAM MEYERHOFF
178 Language in Society 28:2 (1999)
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

a speech community are discussed in much greater detail in the articles by
Bucholtz and by Eckert & McConnell-Ginet in the present collection. Table 1summarizesthedifferencesandsimilaritiesamongthethreeconstructs–speechcommunity,socialidentity,andCofP–allofwhichhavevariouslyprovedusefulin the study of socially stratified language variation.
Social networks
AnanalysisoflanguageinuseemployingtheCofPframeworkalsohasfeatures
incommonwithsocialnetworkanalysis(cf.Bortoni-Ricardo1985,Milroy1987,Lippi-Green 1989, Kerswill 1994). But again, the two frameworks can usefullybe distinguished. Both include some distinction between core membership andperipheralmembership.Theideasofmeasuringanindividual’stieswithinanet-work (multiplex and uniplex), and of the density of a network as a whole, aresimilartotheideathatmembershipinaCofPisacquiredastheresultofaprocessof learning. Such measures provide an escape from unhelpful dichotomies – apoint elaborated in Bergvall’s article in this collection.
Bycontrast,aCofPoffersadifferentperspectivefromasocialnetworkonthe
study of language in society:ACofPrequires regular and mutually defining in-teraction. In a social network, by contrast, weak ties exist even among peopleTABLE1. Different assumptions and predictions that can be used to distinguish research
based on the speech community, social identity theory, and communities of practice.
Speech Community Social IdentityCommunity of
Practice
Shared norms and evalua-
tions of norms arerequired.Shared identifications are
required.Sharedpracticesarerequired.
Sharedmembershipmaybe
definedexternally.Membershipisconstructed
internallyorexternally.Membershipisinternally
constructed.
Nothingtosayaboutrela-
tionshipbetweenanindividual’sgroupandpersonalidentities.Relationbetweengroupand
personalidentitiesisun-clear:continuum?orthog-onal?Activelyconstructeddepen-
denceofpersonalandgroupidentities.
Non-teleological. Non-teleological:anyout-
comesareincidental.Sharedsocialorinstrumental
goal.
Nothingtosayaboutmainte-
nanceor(de)constructionofboundariesbetweencategories.Groupidentityisdefined
throughcomparisonandcompetitionwithoutgroups.Boundariesaremaintained
butnotnecessarilydefinedincontrastswithout-groups.
Acquisitionofnorms. Learningincidental. Socialprocessoflearning.THE COMMUNITY OF PRACTICE
Language in Society 28:2 (1999) 179
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

whohavelimitedorinfrequentcontact.Inshort,asocialnetworkandaCofPcan
bedifferentiatedbythenatureofthecontactthatdefinesthem.Asocialnetworkrequires quantity of interaction; a CofPrequires quality of interaction.
Notwithstanding these differences, it is possible to imagine developing an
index of an individual’s degree of integration into a CofP – one that might becomparable with the measures that have been used to account for different de-greesofintegrationintosocialnetworks.Thiscross-comparisonofmetricsmightbe enlightening because both networks and CofPfocus on diversity and variety,allowing for variation over time; both concepts provide a means of measuringchange, both linguistic and social.
Social constructionist approaches
Recentresearchontherelationshipbetweenlanguageandgenderhasbeendom-
inated by approaches that examine the ways in which gender is socially con-structed in interaction, rather than existing as a fixed social category to whichindividuals are assigned at birth (e.g. Crawford 1995, Hall & Bucholtz 1995,Bergvall et al. 1996, Bucholtz et al. 1996).The concept of CofPis clearly muchmorecompatiblewiththiskindofsocial-constructionistapproachthanareotherlessdynamicoractivity-focusedconcepts.ThustheCofPbeenwelcomedinlan-guage and gender research as a corrective to unsatisfactory essentialist ap-proachestotheanalysisofgender.
4InCameron’swords(1992:13),itencourages
adifferentfocus:“notgenderdifferencesbutthedifferencegendermakes.”Eck-ert & McConnell-Ginet (1992:466) point out that, rather than emphasizinggen-der differences that result from differing patterns of early socialization, genderresearchers can more fruitfully focus on “people’s active engagement in the re-production of or resistance to gender arrangements in their communities.” In-stead of abstracting gender from social practice, they note the need to focus on“gender in its full complexity: how gender is constructed in social practice, andhow this construction intertwines with that of other components of identity anddifference, and of language” (472). The concept of CofP, they suggest, offers afruitful way forward. In the present collection, Eckert & McConnell-Ginet pur-suethesepointsfurther,illustratingwaysinwhichtheconcepthasprovedusefulin their own research. Similarly, Bergvall examines the broader implications ofthe CofPfor theory and methodology in language and gender research.
Thelinguisticbehaviorsandsocialcharacteristicsofspecificcommunitiesof
practice are sketched in the articles by Bucholtz and Ehrlich. Bucholtz’s dataprovide useful evidence of the role played by peripheral or apprentice membersofaCofP,andthelevelofworkthatisexpendedintheongoingprocessofmain-tainingtheboundariesofaCofP.Ehrlichexploresthelinksbetweentheconstruc-tion of individual identities and communities of practice. She examines howdiscoursepatterns,sharedbytheCofPconstitutedbyadisciplinarytribunal,de-construct the identity of the victims whose case they are hearing.JANET HOLMES & MIRIAM MEYERHOFF
180 Language in Society 28:2 (1999)
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

ThecontributionsbyMeyerhoffandFreedillustratethatthebehaviorofsome
socialgroupsmaynotbemostinformativelydescribedintermsofaCofP. Mey-erhoffconcludesthatthenotionsofaspeechcommunityandintergroupdistinc-tivenessaccountmostappropriatelyforthedistributionofapologiesinwomen’sand men’s speech in Vanuatu (South Pacific). Freed argues that, although preg-nantwomenareoftenreferredtoinwaysthatsuggestthatsocietyperceivesthemasaCofP,theoperativecommunitiesofpracticeintheirlivesarereallythedoc-tors, health professionals, and family members with whom they interact. Thesegroups, in concert, produce a “master narrative” of pregnancy.
Finally, the CofP concept offers a potentially productive means of linking
micro-levelandmacro-levelanalyses.TheCofPinevitablyinvolvesmicro-levelanalysis of the kind encouraged by a social constructionist approach. It requiresdetailedethnographicanalysisofdiscourseincontext–toidentifysignificantorrepresentative social interactions, to characterize the processes of negotiatingshared goals, and to describe the practices that identify the CofP.ACofPmust,however, also be described within a wider context which gives it meaning anddistinctiveness.Inotherwords,“thepatterns,generalizations,andnormsofspeechusagewhichemergefromquantitativeanalysesprovideacrucialframeworkwhichinforms and illuminates the ways in which individual speakers use language”(Holmes 1998a:325). Just as quantification depends on preliminary analysis inordertoestablishvalidunitsandcategoriesofanalysis,sodetailedethnographicanalysiscannotaccountforindividuals’linguisticchoicesinavacuum.Thereisalimittothekindsofpatternsthatareevidentatthemicro-leveloftheindividual.
Itishere,inthelinkingofmicro-analysisandmacro-analysis,thattheconcept
of CofPand the associated methodologies perhaps have most to offer, as Eckertand McConnell-Ginet illustrate in their contribution. The level of analysis in-volvedintheCofPapproachencouragesafocusonsocialdiversity,whilesimul-taneouslyfacilitatingtheperceptionofsubtleyetmeaningfulpatterns.ThusEckert1988 showed that some patterns of variation are simply overlooked without theinformation provided by her long-term and highly local studies of Detroit teen-agers.ShedemonstratedthatinnovativeformsoftheNorthernCitiesphonolog-icalshiftwerebeingtransferredalongpathsfromtheso-called“burnouts”tothe“jocks” in the high school she studied. Moreover, looking in detail at the mostinnovative speakers among the burnouts, she made the earliest academic obser-vationofextensionstotheongoingchainshiftofvowelsintheNorthernCities.Byincorporatingintoherstudyprinciplesandmethodsthatarealsorequiredbythe CofP framework, she was able to relate phonetic details in the speech ofindividuals to larger patterns of sociolinguistically stratified variation. In addi-tion, she was able to relate the patterns of innovations and conservatisms, asconstructedbyindividualsintheirspeech,tothesocialpatternsthatthespeakerswere also actively involved in maintaining or building.
The CofP thus offers a fruitful concept to those interested in exploring the
relationship between language and society. With its criterial characteristics, itTHE COMMUNITY OF PRACTICE
Language in Society 28:2 (1999) 181
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

provides an ideal framework for exploring the process by which individuals ac-
quiremembershipinacommunitywhosegoalstheyshare;itprovidesameansofstudyingtheacquisitionofsociolinguisticcompetence,asindividualslocatethem-selves in relation to other community members; and, for similar reasons, it pro-vides a framework for examining language change. A precise definition of thedistinction between core and peripheral members in a CofPpromises further tofleshoutourunderstandingofthepathsbywhichlinguisticchangesspreadthrougha speech community.The specification of constitutive features permits compar-ison among different communities of practice along a range of diverse dimen-sions. Thus, while the CofP shares some characteristics with the concepts ofspeechcommunity,socialidentity,andsocialnetwork,italsooffersdifferentandenriching perspectives. In this collection, practitioners in language and genderresearch illustrate the potential and the limitations of the concept of CofP inextendingtheoreticalandmethodologicalboundariesforthoseinterestedinnewinsights into the relationship between language, gender, and community.
NOTES
1Wengerdistinguishes“peripheral”membersofaCofPfrom“marginal”members,dependingon
whetherthepositionistemporaryanddynamic.EventhoughcoremembersofaCofPmayperceivesomeone as a potential member, their participation may be peripheral as they gradually learn thepractices that will eventually make them a core member. However, individuals may elect to remainperipheral.Bycontrast,marginalmembersareindividualspreventedfromfullparticipation.Wengergives as an example the fact that “we often find it hard to be grown-up participants within our ownfamiliesofbirth”(1998:175);i.e.,ourownpracticesandthoseofothermembersoftheCofPinstan-tiateourmarginalposition.ThedistinctionbetweenmarginalandperipheralmembersofaCofPmaybe of importance in a theory of learning. However, we believe it remains to be shown that the dis-tinction is salient for the synchronic study of (linguistic) behavior.
2Asanyonewhohasworkedinsuchpolicyunitscanattest,theyalsodiffermarkedlyintheextent
towhichtheyarecharacterizedbythe“rapidflowofinformationandthepropagationofinnovations.”
3The concept of “evaluative behaviors” has sometimes been mistakenly interpreted in terms of
“attitudes”(e.g.Hudson1980:27).TheevaluativebehaviorstowhichLabovisreferringinclude,e.g.,the tendency of all speakers to shift toward variants with higher overt prestige as they pay greaterattentiontotheirspeech.Theymaybemarkedlydifferentfromclearlyheldorarticulable attitudes .
4There are obvious parallels, too, in the shift within social psychology to treat an individual’s
group(orpersonal)identitiesashighlylocalconstructions,orasthecumulativeresultofanumberofinteractions. Examples include research into the constructions of a disabled identity (Fox & Giles1996)andtheconstructionofanidentityasamemberofaparticularagecohort(Couplandetal.1991,Ryan et al. 1995).
REFERENCES
Abrams,Dominic(1996).Socialidentity,selfasstructureandselfasprocess.InW.PeterRobinson
(ed.),Social groups and identities: Developing the legacy of Henri Tajfel , 143–67. Oxford: But-
terworth Heinemann.
Bergvall,VictoriaL.,etal.(1996),eds. Rethinkinglanguageandgenderresearch:Theoryandprac-
tice. London: Longman.
Bortoni-Ricardo,StellaM.(1985). Theurbanisationofruraldialectspeakers:Asociolinguisticstudy
in Brazil. Cambridge & NewYork: Cambridge University Press.JANET HOLMES & MIRIAM MEYERHOFF
182 Language in Society 28:2 (1999)
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core

Bucholtz,Mary,etal.(1996),eds. Culturalperformances:ProceedingsoftheThirdBerkeleyWomen
andLanguageConference,April1994 .Berkeley:BerkeleyWomen&LanguageGroup,University
of California.
Cameron,Deborah(1992).“Notgenderdifferencesbutthedifferencegendermakes”:Explanationin
research on sex and language. International Journal of the Sociology of Language 94:13–26.
Coupland, Nikolas, et al. (1991). Language, society and the elderly . Oxford: Blackwell.
Crawford, Mary (1995). Talking difference: On gender and language . London: Sage.
Eckert,Penelope(1988).Adolescentsocialstructureandthespreadoflinguisticchange. Languagein
Society17:245–67._,&McConnell-Ginet,Sally(1992).Thinkpracticallyandlooklocally:Languageandgender
as community-based practice. Annual Review ofAnthropology 21:461–90.
Fox, SusanAnne, & Giles, Howard (1996). Interability communication: Evaluating patronizing en-
counters. Journal of Language and Social Psychology 15:265–90.
Giles,Howard,&Coupland,Nikolas(1991). Language:Contextsandconsequences .PacificGrove,
CA: Brooks /Cole.
Giles, Howard, et al. (1987). Speech accommodation theory:The first decade and beyond. Commu-
nication Yearbook 10:37–56.
Gumperz, John J. (1971). Language in social groups . Stanford: Stanford University Press.
Hall,Kira,&Bucholtz,Mary(1995),eds. Genderarticulated:Languageandthesociallyconstructed
self. London: Routledge.
Holmes, Janet (1997). Analysing power at work: An analytical framework. Paper presented at the
SixthInternationalConferenceonLanguageandSocialPsychology,UniversityofOttawa,16–20May1997.(EducationalClearinghouseonLanguagesandLinguistics,ERICdocumentED414733FL024823.)
_(1998a). Women’s role in language change: A place for quantification. In Natasha Warner
etal.(eds.), Genderandbeliefsystems:ProceedingsoftheFourthBerkeleyWomenandLanguage
Conference, 1996 , 313–30. Berkeley: BerkeleyWomen and Language Group._(1998b). Doing collegiality and keeping control at work: Small talk in government depart-
ments. In Justine Coupland (ed.), Small talk , to appear. London: Longman._;Stubbe,Maria;&Vine,Bernadette(1999).Constructingprofessionalidentity:“Doingpower”
in policy units. In Srikant Sarangi & Celia Roberts (eds.), Talk, work and institutional power:
Discourseinmedical,mediationandmanagementsettings ,toappear.Berlin:MoutondeGruyter.
Hudson,RichardA.(1980). Sociolinguistics .Cambridge&NewYork:CambridgeUniversityPress.
Kerswill,Paul(1994). Dialectsconverging:RuralspeechinurbanNorway .Oxford:Clarendon;New
York: Oxford University Press.
Labov,William (1972). Sociolinguistic patterns . Philadelphia: University of Pennsylvania Press.
Lave,Jean,&Wenger,Etienne(1991). Situatedlearning:Legitimateperipheralparticipation .Cam-
bridge & NewYork: Cambridge University Press.
Lippi-Green,RosinaL.(1989).Socialnetworkintegrationandlanguagechangeinprogressinarural
alpine village. Language in Society 18:213-34.
Meyerhoff, Miriam, & Niedzielski, Nancy (1994). Resistance to creolization:An interpersonal and
intergroup account. Language and Communication 14:313–30.
Milroy, Lesley (1987). Observing and analysing natural language . Oxford: Blackwell.
Preston,DennisR.(1989). Perceptualdialectology:Nonlinguists’viewsofareallinguistics .Dordrecht:
Foris.
Ryan, Ellen B., et al. (1995). Communication predicaments of aging: Patronizing behavior toward
older adults. Journal of Language and Social Psychology 14:144–66.
Santa Ana, Otto, & Parodi, Claudia (1998). Modeling the speech community: Configuration and
variable types in the Mexican Spanish setting. Language in Society 27:23–51.
Tajfel, Henri (1978). Interindividual behaviour and intergroup behaviour. In H. Tajfel (ed.), Differ-
entiation between social groups: Studies in the social psychology of intergroup relations , 27–60.
London & NewYork:Academic Press._,&Turner,J.C.(1986).Thesocialidentitytheoryofintergroupbehavior.InWilliamG.Austin
& StephenWorchel (eds.), Psychology of intergroup relations , 7–24. Chicago: Nelson Hall.
Wenger, Etienne (1998). Communities of practice . Cambridge & NewYork: Cambridge University
Press.THE COMMUNITY OF PRACTICE
Language in Society 28:2 (1999) 183
https://www.cambridge.org/core/product/2FDA5017C035E21EC5AB06845482B3D9terms of use, available at https://www.cambridge.org/core/terms . Downloaded from https://www.cambridge.org/core . IP address: 82.137.14.241 , on 09 Aug 2017 at 12:38:09 , subject to the Cambridge Core