Documente Academic
Documente Profesional
Documente Cultură
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=ucpress.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The
American Journal of Sociology.
http://www.jstor.org
WhatIs Agency?'
MustafaEmirbayerand Ann Mische
New Schoolfor Social Research
1 Thisis a fully
coauthoredarticle.Earlierdraftswerepresentedat thePaul F. Lazars-
feldCenterfortheSocialSciencesat ColumbiaUniversity, theWorkshopon Politics,
Power,and Protestat New YorkUniversity, theColloquiumon Cultureand Politics
at theNew SchoolforSocialResearch,themeeting oftheAmericanSociologicalAsso-
ciationat Los Angeles,and variousseminarsat theNew SchoolforSocial Research
and Princeton We wouldliketothanktheparticipants
University. in thoseforums for
theirmanyusefulcomments. We wouldalso liketothankJeffrey Alexander, Bernard
Barber,RichardBernstein,Donald Black, Mary Blair-Loy,David Gibson,Chad
Goldberg,JeffGoodwin,Michael Hanagan, Hans Joas,MicheleLamont,Edward
Lehman,CalvinMorrill,MichaelMuhlhaus,ShepleyOrr,MargaritaPalacios,Mimi
Sheller,CharlesTilly,Diane Vaughan,Loic Wacquant,and HarrisonWhitefortheir
manyilluminating insights,
criticisms,andsuggestions.Directcorrespondence toMus-
tafaEmirbayer, Department ofSociology,New SchoolforSocial Research,65 Fifth
Avenue,New York,New York 10003.
? 1998by The University
of Chicago.All rightsreserved.
0002-9602/98/10304-0004$02.50
963
AmericanJournalof Sociology
THEORIZING AGENCY
Many of the tensionsin present-dayconceptionsof human agencycan
be tracedback to the Enlightenment debate over whetherinstrumental
rationalityor moraland norm-basedactionis thetruestexpressionofhu-
man freedom.Teleologicaland instrumentalist conceptionsof actionfu-
eled the philosophicalindividualismof the earlyEnlightenment, which,
whilestillgroundedin thereligiousmoralityofthetimes,allowed forthe
subsequentinventionoftheindividualas a "freeagent"able tomakeratio-
nal choices for (him)selfand society(Lukes 1973). With JohnLocke's
(1978) rejectionof the bindingpower of tradition,his locationof beliefs
in individualexperience,and hisgroundingofsocietyin thesocialcontract
964
Agency
965
AmericanJournalof Sociology
966
Agency
subjectiveconstruction
of choicesfromthetemporalvantagepointsof contextually
embeddedactors.
967
AmericanJournalof Sociology
3 We are not concerned here with Mead's engagementin this work with functionalist
evolutionarytheorynor with his debate with metaphysical theoristsof temporality.
AlthoughMead develops his theoriesthrougha comparison with more general physi-
cal and biological (i.e., nonhuman) processes and has been criticizedforveeringaway
fromaction theorytoward metaphysics(Joas 1985), he also provides the philosophical
core of a temporal and relational understandingof the intersubjectivedevelopment
of agentic capacities, which is of critical importance for a theoryof action. For a
related discussion, see also Mead's (1938) work, The Philosophy of the Act.
968
Agency
969
AmericanJournalof Sociology
970
Agency
971
AmericanJournalof Sociology
972
Agency
973
AmericanJournalof Sociology
I Sucha position
doespresentus witha certaindifficulty:
namely,
thatcorporate
actors
suchas firms, states,or otherorganizational
entitiescannoteasilybe accommodated
withinthetermsof such a framework unlesstheyare themselves giventheoretical
statusequivalentto thatof naturalpersonsor selves(forexamplesof thismodeof
reasoning, see Coleman[1990],Luhmann[1990],and White[1992]).Whilenotaverse
tosucha movein principle, we do notpursueall ofitsmanyimplications in thepages
to come,or grapplesystematically withthe specialchallengesin translation thatit
wouldnecessarily entail.
10 It is worthnotingthatWiley'sperspective is itselfself-consciously
groundedin the
pragmatist tradition(see also Wiley1994,pp. 10, 29, 47); fora similarperspective,
see Taylor(1991),Colapietro(1990),and Gergen(1994).Moreworkneedsto be done,
ofcourse,in theorizing thesystematicblockagesto such"openand free"intrapsychic
communication or dialogue.
974
Agency
975
AmericanJournalof Sociology
" See,e.g.,Camic'sdiscussions
ofW. I. Thomas,FlorianZnaniecki,RobertPark,and
TalcottParsons,amongothers,in Camic (1986,pp. 1072-75).Camic adds thatthe
historical
reasonsforthistendency
aretwofold:ontheonehand,theemergence during
thelate 19thcentury ofDarwinianevolutionarytheoryand ofexperimental physiol-
ogyand,on theotherhand,theriseduringthatsameperiodofa "militantly scientistic"
new fieldof psychology.Betweenthem,thesedevelopments led to an identification
ofhabitualactionwiththemostelementary behavioralprocessesofthehumanorgan-
ism,akinto thoseofthelowerspecies(Camic 1986,pp. 1048-49).
976
Agency
12
"For Aquinas,. . . a habitusputsone's activitymoreunderone's controlthanit
mightotherwise be. In thissense,to have a habitusis to be disposedto someactivity
or other-notbecauseone tendsto thatactivityon everypossibleoccasion,butbe-
cause one findsit natural,readilycopedwith,an obviousactivityto engagein, and
so on" (Davies 1992,pp. 225-26; emphasisin theoriginal).
977
AmericanJournalof Sociology
quant 1992) and Giddens (1979, 1984) build upon the insightsof both
pragmatismand phenomenology, as well as upon earliertraditionsof
thought.Bourdieuuses theAristotelian/Thomistic idea ofhabitusto illu-
minatetheformativeinfluences ofthepast upon thecognitive,corporeal,
and intentionalstructures ofempiricalaction.Throughtheincorporation
of past experiencesin the body,he maintainsthat social actorsdevelop
a setofpreconsciousexpectationsabout thefuturethatare typicallyinar-
ticulate,naturalized,and takenforgrantedbut nevertheless strategically
mobilizedin accordancewith the contingenciesof particularempirical
situations.Bourdieurecognizesthecompatibility ofsuchnotionswiththe
insightsofbothDewey and thephenomenologists: "The theoryofpractical
sensepresentsmanysimilarities withtheories,suchas Dewey's,thatgrant
a centralroleto thenotionof habit,understoodas an activeand creative
relationto the world"(Bourdieuand Wacquant 1992,p. 122).
In similarfashion,Giddensconceptualizestheagenticdimensionofrou-
tinebehaviorin termsof what he calls thestratification modelofaction
(Giddens1979,p. 56). By distinguishing betweenthreelevelsofconscious-
ness-the unconscious,practicalconsciousness,and discursiveconscious-
ness-he in effectconstructsa continuumbetweenthe unreflective and
reflectivedimensionsof action.But despitethisnod towarddiscursivity,
Giddensgivesroutinizedpracticalconsciousnessa privilegedplace in the
explanationofsocial reproduction, themasterkeyof
callingroutinization
his theoryof structuration. Such consciousnessemergesout of a back-
groundof"tacitlyemployedmutualknowledge"(Giddens1979,p. 58), by
means of which social interactionsare reflexively monitored.In under-
scoringtheagenticmomentin thereproduction he also de-
of structures,
velops the importantidea of recursivity:structures(which Giddens
definesas "rulesand resources")are reallyonly"virtual"structures (para-
digmaticpatterns)thatmustbe recursivelyactivatedwithinsocial prac-
tices.The agenticdimensionof routinizedaction lies preciselyin the re-
cursiveimplementation of structuresby humanactors."3
13 Giddens(1991)is particularly
interestedin theconceptofroutinization becauseof
hisontological he emphasizestheneedfor"basictrust"and"ontologi-
presuppositions:
thatdriveshumansto routinizetheirpracticesand to give orderand
cal security"
to theirrelationships,
stability especiallyin the faceof the growingcomplexityand
diversityofmodernsociety(fora similarperspective, see White[1992]).
978
Agency
979
AmericanJournalof Sociology
980
Agency
Iterationin EmpiricalResearch
The iterationalorientationof agencyhas alreadyproveda richsourceof
researchquestionsin a varietyof social sciencedisciplines.Here we ex-
plorehow such researchopens up a numberofintriguing linesofinquiry
intothereciprocalrelationship-theongoingdialogueor conversation-
between the agency in its iterationalmodalityand a wide range of
temporal-relational contextsof action.
Culturalcompetences.-Researchbuildingupon Bourdieu's notionof
habitusproveshighlyusefulin showinghow different formativeexperi-
ences,such as thoseinfluencedby gender,race, ethnicity, or class back-
grounds,deeplyshape theweb ofcognitive,affective, and bodilyschemas
throughwhichactorscometo knowhow to act in particularsocial worlds.
Ann Swidler(1986) evokesBourdieuin speakingofthe"culturaltoolbox"
of practicalcompetencesthatpredisposeactorsto feela fitwithinsome
actionsand notothers.AlthoughLoic Wacquant (1992b)criticizestheim-
plicitinstrumentalism of Swidler's account,his work on boxingin Chi-
cago ghettoneighborhoodssoundssimilarthemesby exploringhow em-
bodied competencesand classificatory schemas firstlearned withinthe
streetenvironment underlieboxers'subsequentengagementofthe"pugi-
listicfield."Likewise,MicheleLamont's (1992) researchintomoney,mor-
als, and mannersin France and the UnitedStatesexamineshow classifi-
catoryschemas developed withinparticularclass, race, and national
settingsinfluencetheboundaryworkofsocial actorsin articulating tastes
and aspirations,as well as in distinguishing
themselvesfromothersocial
groups.In suchways,thesewritersclaim,theagenticreactivationofsche-
mas inculcatedthroughpast experiencetendsto correspondto (and thus
to reproduce)societalpatterns:"Social structuresand cognitivestructures
981
AmericanJournalof Sociology
982
Agency
983
AmericanJournalof Sociology
984
Agency
985
AmericanJournalof Sociology
986
Agency
16 Cornelius
Castoriadis(1987,p. 87) drawsheavilyuponHeidegger-as wellas Aris-
totleand Marx-in his own theoryof"theimaginative of society":"To
constitution
do something, to do a book,to make a child,a revolution, or just doingas such,
is projectingoneselfintoa futuresituationwhichis openedup on all sides to the
unknown."
17In
contrastto mostrationalchoicetheorists, Schutz(1967,p. 69) maintainsthat
choicesare highlyunstableand onlygainrelativeclarityaftertheact has beencom-
pleted,throughex postfactoreflection: "The erroris to supposethatthe conscious
state,whichonlyexistsafterthedeed is done,lies back at some 'pointof duration'
beforetheactual choice."
987
AmericanJournalof Sociology
less helpfulin showingwhat projectsare good for-that is, how our pro-
jectivecapacityis essentialto problemsolvingwithina community. Here,
once again,we can turnto the pragmatists, who in additionto theircon-
cern with routine,are deeply attunedto the imaginativeflexibility of
actors' deliberationsabout the future.Dewey (1981, p. 61), forexample,
characterizestheexperimental withthefutureas an essential
relationship
dimensionofhumanaction:"Experiencein itsvitalformis experimental,
an effortto changethegiven;itis characterized byprojection,byreaching
forwardintotheunknown;connectionwiththefutureis itssalienttrait."
Human intelligenceis based upon the capacityto "read futureresultsin
presenton-goings"(Dewey 1981,p. 69); thisprojectivecapacitypermits
thekindof responsivechoiceand inventivemanipulationofthephysical
and socialworldsthatis so essentialtodemocraticparticipation. Likewise,
Mead (1934) stressesthe essentiallyintersubjective dimensionof projec-
tivity,arguingthatour basic self-concept is developedfromthecapacity
to projectourselvesintotheexperiencesofothers.The imaginativecapac-
ityof the "I" to move betweenmultiplesituationallyvariable "me's" is
what constitutesfreedomand maneuverability in relationto established
roles,as well as makingpossiblesocial coordination, joint problemsolv-
ing,and collectiveprojectsofsocial reform. In thepragmatistview,proj-
ects are not constitutedmerelyby "thrownness" intoan uncertainworld
that condemnsus to freedom,but also by the practicalexerciseof that
freedomalong withothersin pursuitof a commongood.
988
Agency
989
AmericanJournalof Sociology
990
Agency
maturgicanthropology. VictorTurner(1974),forexample,describesthe
"social dramas"thatare enactedduring"liminalperiods"in whichsocie-
tiesritualistically
reversesocial roles.AlthoughTurnerstresseshow such
dramas reinforce the social order,we would argue thattheseliminalex-
perimentalperiodsmay also have a transformative and renovationalef-
fectupon the largerculture,as new possibilitiesforhuman interactions
are imagined,tested,and (perhaps)definedon a collectivescale.
Projectivityin EmpiricalResearch
In consideringhow past patternsof interactionare imaginatively recom-
posed to generatenew futurepossibilities, we open up a richlysuggestive
fieldforsociologicalresearch.This is in contrastwithmuchof empirical
sociology,where,despiteits extensivephilosophicallegacy,the notionof
projectshas largelybeen ignored,due in part to its perceivedsubjective
nature and the apparent incompatibility of "imaginative"phenomena
with behavioral observation,survey techniques,and macrostructural
analysis.We arguethatprojectivity needsto be rescuedfromthesubjec-
tivistghettoand put to use in empiricalresearchas an essentialelement
in understanding processesof social reproductionand change.Many of
theelementsoutlinedabove have, in fact,been addressedby a wide body
of literaturein various social sciencedisciplines,albeit in an undertheo-
rizedand residualway. Here we discussseveraloftheseapproaches(and
theirlimitations),in orderto pointtowardfutureresearchon theinterplay
betweenthe projectivedimensionof agencyand the different temporal-
relationalcontextsof action.
Time perspectives.-While most life course approaches in sociology
have tendedto focusupon theinfluenceofpast experienceson subsequent
life paths, a well-developedsubfieldin social psychologyhas explored
questionsmoredirectlylinkedto projectivity. Since the 1940s,research
has been carriedout on "timeperspective"and itsinfluenceon such mat-
tersas academicsuccessand civilianmorale(Lewin 1948);morerecently,
researchers in thisarea have investigatedchangesin timeperspectivedur-
ingdifferent developmentalperiods,such as childhood,adolescence,mid-
dle adulthood,and old age. Of particularrelevanceto projectivityare
studiesoftheconstruction offutureexpectations, examiningsuch factors
as variabilityin the densityand extensionof imaginedfutureevents,
linkedto cognitivedevelopmentand/orparticularsocial contextssuch as
familyor class background(Cottle and Klineberg1974; Devolder and
Lens 1982;Wessmanand Gorman1977;Greene1986,1990).Whilemuch
ofthisresearchis limitedby overlybehavioraland correlationalassump-
tions,recenttheorists ofnarrativity have added an interpretive dimension
991
AmericanJournalof Sociology
to lifecoursestudies(Gergenand Gergen1983,1984,1988;Bruner1986;
Sarbin 1986),exploringhow personalconceptionsabout past and future
are transformed at key momentsof transitionand/orcrisis(Riegel 1975,
1977; Cohler 1982).
Propheticmovements. -A second line of work that directlyengages
projectivity is the extensiveliteratureon prophetic,utopian,and revolu-
tionarymovements.Whilesuchliterature can be criticizedforitsoverem-
phasis on cultural(as opposedto social-structural or social-psychological)
factors,we argue (along withDesroche [1979] and Ricoeur[1991]) that
theprojectiveimaginationas expressedin collectiveideals and aspirations
plays a constitutive, not just an epiphenomenal,role in a wide variety
of historicalphenomena,rangingfrommillenarianmovements,religious
cults,alternativecommunities, and revolutionary organizations, to more
generalizedformsof culturalrevival.For example,NormanCohn (1977,
pp. 16-17) arguesthatmillenarianprojectionsappearingin Europe dur-
ing the llth-16th centuriesresultedin very different kinds of move-
mentsthan the more limitedlocalized peasant or artisanrevoltsof the
period:"The usual desireof the poor to improvethe materialconditions
of theirlives became transfusedwithphantasiesof a world reborninto
innocence througha final, apocalyptic massacre." Likewise, William
McLoughlin(1978,p. 2) claimsthatmajor"greatawakenings"duringpe-
riodsofuncertainty and changein Americanhistoryled to catharticreviv-
als that"eventuatedin basic restructurings of our institutions and redefi-
nitions of our social goals." Finally, both Marxist and non-Marxist
historiansof revolutions(e.g.,Thompson1966, 1993; Walzer 1965, 1980)
explore the projectivedimensionsof revolutionarymovements,which
Walzerdefinesas "consciousattemptsto establisha new moraland mate-
rial worldand to impose,or evoke,radicallynew patternsof day-to-day
conduct."Revolutionitself,he concludes,"is a project"(Walzer 1980,
pp. 202-3).
Framingprocesses.-The projectiveimaginationis also a factorin less
apocalypticformsofsocial movementsand efforts at institutional
reform.
Most work in this well-researched area fails to adequatelytheorizethe
projectivedimension,due in part to the paradigmaticsplit duringthe
1970s and 1980s between"strategy"and "identity"(Cohen 1985). This
split,whichgoes back to theKantian divisionbetweeninterestsand ide-
als, has had the effectof severingtwo intrinsically linkeddimensionsof
projectivity:strategiesare strippedofmeaningand reflexivity, whileiden-
titiesare temporally flattenedout and shornoftheirorienting power(Mis-
che 1994).18Recentattemptsto bringthetwoparadigmstogether (see Mor-
developedthiscritiquein an empiricalstudyof
18 Mische(1997,pp. 7-8) has further
992
Agency
993
AmericanJournalof Sociology
994
Agency
concernswithexplicitdecisionproceduresand a widespread"flightfrom
ambiguity"(Levine 1985) and judgmenthave become evidentin a host
of analyticalperspectives-notonlyrationalchoice theory,but also less
explicityet equally instrumentalist conceptionsof social action,dating
back at least to Max Weber's discussionsofZweck-and Wertrationalitat
(Weber 1978).Even Durkheim(1961,pp. 31, 26) sees morality, by defini-
tion,as a "systemof commandments," of special rules [that
"an infinity
are] fixedand specific.""To the extent,"he writes,"thatthe rule leaves
us free[and] does notprescribein detailwhat we oughtto do, the action
beingleftto ourownjudgment,to thatextentthereis no moralvaluation"
(Durkheim1961,pp. 23-24).
Aristotelianperspectiveson practicalwisdom.-We mustreturnonce
again to Aristotle'swritingson ethicsforone of the earliest(and most
fullydeveloped)theoriesofprudenceor practicalwisdom.In markedcon-
trastto later rule-basedtheories,Aristotleholds that "threefeaturesof
'the matterof the practical' . . . show why practicalchoices cannotbe
adequatelyand completely capturedin a systemofuniversalrules"(Nuss-
baum 1986,pp. 303-4): themutability oftheparticular,itsindeterminacy
(complexityand contextualvariety),and its inherentnonrepeatability.19
Also, the values, rules,and principlesthatare constitutive of a good hu-
man lifeare themselvesplural and incommensurable; hence,a concern
forsituatedjudgmentssupplantsany simplebeliefin the unproblematic
applicationofuniversalnormsor imperatives(Nussbaum 1986,pp. 303-
4, 294-95). In Aristotle'sview, practicalwisdom can refervariouslyto
meansor to ends;itcan be eitherstrategicand calculative-in whichcase,
he says,we can speak of personsas beingclever,crafty,or cunning-or
it can be concernedwithbroaderquestionsofthegood lifeitself(Aristotle
1985,p. 153).Aristotlesees practicalwisdomas intrinsically communica-
tive in nature;thatis, it entailsa deep involvementand participationin
an ongoingcommunity of discourse.Far frombeingpurelyindividualor
monological,itremainsopento dialogueand persuasionand is profoundly
implicatedin commonvalues, interests, and purposes.
Theoriesofjudgmentand critical deliberation. -A significant break
withthislegacycomesaboutwithKantianethics,whichregardsprudence
not as a virtue,as did so many earliermoral theories,but ratheras a
vehicleforcold and selfishcalculation,expediency,and pragmatism. And
yet,especiallyin his laterwork,even Kant indirectlyprovidesa theory
ofpracticalevaluation.20 Moreover,he adds thatpracticaljudgments(spe-
19
"Perhapsthemostobviousand astonishing absencefromAristotle's
thoughtforany
modernreaderis thatthereis relativelylittlementionofrulesanywhere intheEthics"
(MacIntyre1981,p. 141).
viewsofprudence,see Kant (1956),pp. 16,37-38; (1964),
20 For Kant's earlycritical
995
AmericanJournalof Sociology
cifically,
judgmentsoftaste)fallwithinthepotentiality ofall personssince
they"depend . .. on our presupposing the existence of a commonsense
[sensuscommunis]"(Kant 1951,p. 83). Kant linkssuch "commonsense"
to what he calls thecapacityforan "enlargedmentality," in whichjudg-
mentis carriedout by abstractingfromone's own limitedexperiencein
orderto put oneselfin thepositionofeveryoneelse and thusto deliberate
over thecollectivegood. Such an idea recallsAristotle'snotionof a com-
munityof discourse,as well as the more distinctively modernthemeof
autonomy,sincejudgmentno longerdependsupon the subjectivityand
capriceof concreteindividuals.
More recentexamplesof theoriesthat fullyembracethe criticaland
dialogic aspects of practicalevaluationcan be foundin the writingsof
JohnDewey, Hannah Arendt,and JurgenHabermas. Dewey subsumes
Kant's insightson reflective judgmentintohisown pragmatist-and emi-
nentlyrelational-theoryof judgment.In "The Logic of Judgmentsof
Practice"(1985),he pointsout thatall suchjudgmentsbeginwitha prob-
lematicexperience,a forkin theroad,whichtheyattemptexperimentally
to resolve.Judgmentsgain intersubjectivevalidityfromassumingthe
standpointof a sensuscommunis,"a wholeofcommoninterestsand pur-
poses"(Dewey 1978,p. 286). Arendt(1984,p. 36) also expandsupon Kant
by maintainingthat reflective judgmentis not limitedto aestheticsbut
represents "themostpoliticalofman's mentalabilities."She buildsupon
Kant's notionof the enlargedmentality, whichshe terms"representative
thinking," describingit as the abilityto see thingsfromthe perspective
ofothers,"an anticipatedcommunication withotherswithwhomI know
I mustfinallycome to some agreement"(Arendt1977a,pp. 220; see also
Arendt1977b;Benhabib 1992b).And finally,Habermas (1990, 1993) en-
gages Kant's doctrineof judgmentwhile insistingthat he is correcting
Kant's ethicalrigorism;to a Kantian "discourseof justification" he adds
a moreAristotelian"discourseof application."In developinghis theory
ofcommunicative action,Habermasretainsa Kantianemphasisupon de-
liberationand intersubjective validity,even as he objectsto theemptiness
of Kantian ethicsitself.21
996
Agency
Feministtheories.-Meanwhile,manyfeminist thinkerscritically
draw
upon both Aristotelianand Kantian outlookson practicalevaluationin
analyzingthe particularcapacities,experiences,and historiesof women,
while also generalizingfromthese experiencesto develop broader(less
"essentializing")theoriesofmoraland practicalreasoning.One important
contribution is Carol Gilligan's(1982, p. 22) In a DifferentVoice,which
stressesgenderdifferences in theuse of situatedreasoningand a "contex-
tual mode ofjudgment"and therebyseeksto overcomethelimitationsof
Kant's abstractuniversalisticconceptionsof moraljudgmentand action
(e.g.,Kohlberg1981).Froma verydifferent perspective,Donna Haraway
(1988),too,criticizesestablishedunderstandings of"objectivity,"
and calls
insteadfor"situatedknowledges"groundedin theparticularities ofpartial
"limited"locations.Finally,Seyla Benhabibstressesprocessesofdialogue
and publicdeliberationin herown communicative conceptionofpractical
judgment:thereis "no incompatibility," she writes,"betweentheexercise
of moral intuitionguided by an egalitarianand universalistmodel of
moralconversation[Kant]and theexerciseofcontextualjudgment[Aris-
totle]"(Benhabib 1992a, p. 54; see also Benhabib 1987, 1992c).
997
AmericanJournalof Sociology
998
Agency
999
AmericanJournalof Sociology
PracticalEvaluation in EmpiricalResearch
Finally,we outlineresearchfindingsthatpertainto empiricalmanifesta-
tionsof practicalevaluation,in orderto conveya clearersense of what
is entailedby this analyticalaspect of agencyand to show how it can
be investigatedsociologically.These findingsunderscorepossiblewaysin
whichpracticalevaluationmightbe elicitedin particularcontextsand in
whichit affectsin turnthe abilityof actorsto engage with,respondto,
and potentiallytransform theirstructuralenvironments.
Temporalimprovisation.-Oneresearcharea thatprovidesinsightinto
the temporalcontextualization of both ritualand purposiveaction in-
cludesstudiesofsequencingprocessesin social interactions. For example,
Bourdieu's investigations of the manipulationof the temporalstructure
of giftexchangereveal that the same gift-giving act can have different
meaningsat different times,alteringtheeffectiveness oftheintendedact.
Temporal strategiesthat enable actorsto controlintervalsbetweenex-
pectedritualtransactions-forexample,by "holdingback or puttingoff,
maintainingsuspense or expectation,"or otherwisemanipulatingthe
"tempos"of action-allow themto gain significant materialand/orsym-
bolic advantages vis-a-vistheirpartnersin exchange (Bourdieu 1977,
p. 71). Additionalexamples of temporalimprovisationinclude "turn-
taking"patternsin everydayconversationalinteractions.Conversation
analystsin the traditionof Schutz and Garfinkel(e.g., Sacks, Schegloff,
and Jefferson 1974) investigatethe subtletiesof timingand delay in the
social organizationoftalk,showingat a microlevelhow agenticmanipu-
lationsoftimeallow actorsto engagein repairwork,to avoid or (alterna-
tively)initiateconflict,and in myriadotherways to advance theirown
interests.
Resistance,subversion, -Another openingforpractical
and contention.
1000
Agency
1001
AmericanJournalof Sociology
1002
Agency
Structure,
Action,and Agency
A varietyofrecentattemptsto rethinktherelationship betweenstructure
and agencyhave arguedthatthe Kantian dichotomybetweenideal and
materialrealms-togetherwithparalleldistinctions betweenfreewilland
necessity, voluntarismand determinism-mustbe replacedbyan outlook
thatregardstheseelementsas reciprocally constitutingmomentsofa uni-
fiedsocial process.Seminal workin thisarea includesBourdieu's (1977,
1990) attack on the divisionbetweensubjectivismand objectivism,as
well as Giddens's(1979,1984)theoryofstructuration, whichcharacterizes
structure and agencyas mutuallyconstitutive (and henceinseparable)ele-
ments.This notionhas been a salutaryand fruitful one forsociological
theory,making possible empiricalresearchthat underscoresboth the
causal significanceof structureas the constrainingand enablingcondi-
tionsof action,and of praxis as "an active constituting process,accom-
plishedby,and consistingin,thedoingsofactivesubjects"(Giddens1976,
p. 121). But it has also broughtin its trainseveraltheoreticaldisadvan-
tages.Foremostamongtheseis a tendencytowardwhatMargaretArcher
(1982, 1988) termsthe "fallacyof centralconflation": the tendencyto see
structureas so closelyintertwined witheveryaspectofpracticethat"the
constituentcomponents[of structureand agency]cannot be examined
separately.... In the absence of any degreeof autonomyit becomesim-
possible to examinetheirinterplay"(Archer1988, pp. 77, 80; emphasis
in the original).22
speaking,Archermeansby "centralconflation"
22 Strictly an elisionof thetwo key
elementsof "CulturalSystem"and "Sociocultural Interaction."
We generalizefrom
hercriticismsto makea broaderpointabouttherelationship betweenagencyand its
pluralityofstructural If,as Archer(1988,pp. 89-90) putsit,"thepowersof
contexts.
Mephistopheles [structure]
ultimately dependon Faustus[agency]continuing to in-
vokethem,"theconstraining and enablingpowersofspecificactualstructures cannot
be determined.Andcorrespondingly, ifactors"areassumedtoenjoya constant degree
oftransformative thenthecircumstances
freedom," underwhichoneencounters "more
voluntarism" or "moredeterminism" also cannotbe specified(on thispoint,see also
Alexander[1994]).
1003
AmericanJournalof Sociology
1004
Agency
EmpiricalPropositions
1005
AmericanJournalof Sociology
1006
Agency
1007
AmericanJournalof Sociology
1008
Agency
1009
AmericanJournalof Sociology
1010
Agency
24 Abbott,
however,is less interestedin the subjectivecomposition of such turning
pointsthanin thestructural characteristics
thatmakethemparticularly susceptible
to transformative
action.Trajectories,he claims,can be conceivedof as narratively
constructed"networksthrough time,"linkedbyoccasionaltransitions thatbringabout
ofthelogicgoverning
a reformulation theconnection betweenpastandfuture possibil-
ities.Turningpointsare the"peculiarly essentialjunctures. . . whereactionmight
makeparticularlyconsequential bridgesbymakingor breakinglinksbetweenmany
networks" (Abbott1997b,p. 99).
1011
AmericanJournalof Sociology
CONCLUSION
We have argued throughoutthis essay that human agencyneeds to be
radicallyreconceptualized.Neitherrationalchoice theory,norm-based
approaches,nor any of the othersociologicalperspectivesextanttoday
providea fullyadequate understanding ofitssignificanceand constituent
features.Nor do such perspectivessatisfactorily answerthe questionas
to how agency interpenetrates with and impacts upon the temporal-
relationalcontextsof action.
We have contendedthatone keyto understanding thevariableorienta-
tionsofagencytowarditsstructural contextsliesin a moreadequate theo-
rizationof the temporalnatureof human experience.Actorsare always
livingsimultaneously in the past, future,and present,and adjustingthe
varioustemporalities of theirempiricalexistenceto one another(and to
theirempiricalcircumstances)in more or less imaginativeor reflective
ways. They continuously engagepatternsand repertoires fromthe past,
projecthypothetical pathwaysforwardin time,and adjust theiractions
to the exigenciesof emergingsituations.Moreover,thereare timesand
places whenactorsare moreorientedtowardthepast,moredirectiveto-
ward the future,or more evaluative of the present;actors may switch
between (and reflexivelytransform)their orientationstoward action,
therebychangingtheirdegreesofflexible,inventive,and criticalresponse
towardstructuring contexts.Such a perspectivelays thebasis fora richer
and moredynamicunderstanding ofthecapacitythatactorshave tomedi-
ate thestructuringcontextswithinwhichactionunfolds.We have referred
to thisperspectiveas relationalpragmatics.
Finally,thispointof view also opens up the possibilityto conceiveof
moraland practicalissues regardinghumanfreedom,creativity, and de-
mocracyin a moresatisfactory and powerfulway. In thisessay,we have
notlaid out a normativetheorythatactuallydistinguishes between"bet-
ter"or "worse"agenticprocesses,"moreor less morallyworthy"projects.
The elaborationof such a theorywould requireeven longerand more
complexargumentsthan thosepresentedhere. Yet, we have delineated
theanalyticalspace withinwhichreflective and morallyresponsibleaction
mightbe said to unfold.Throughout,we have stressedthereconstructive,
(self-)transformative potentialitiesof human agency,when faced with
contradictory or otherwiseproblematicsituations.What are commonly
referredto as normsand values, we can now add, are themselvesby-
productsofactors'engagementwithone anotherin ambiguousand chal-
lengingcircumstances;theyemergewhen individualsexperiencea dis-
1012
Agency
REFERENCES
Abbott,Andrew.1988."Transcending GeneralLinearReality."SociologicalTheory
6:169-86.
. 1997a."Of Time and Space: The ContemporaryRelevanceof the Chicago
School." Social Forces 75:1149-82.
Social Research16:
. 1997b."On theConceptofTurningPoint."Comparative
85-105.
Abbott,Andrew,and AlexandraHrycak.1990."MeasuringResemblancein Sequence
1013
AmericanJournalof Sociology
1014
Agency
1015
AmericanJournalof Sociology
1016
Agency
1017
AmericanJournalof Sociology
1018
Agency
1019
AmericanJournalof Sociology
1020
Agency
1021
AmericanJournalof Sociology
1022
Agency
1023