Field Theory and Experiment in Social Psychology: Concepts and Methods

Author(s): Kurt Lewin

Source: American Journal of Sociology, Vol. 44, No. 6 (May, 1939), pp. 868-896
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/2769418 .
Accessed: 06/02/2014 03:42

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
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
American Journal of Sociology.

http://www.jstor.org

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY AND EXPERIMENT IN SOCIAL
PSYCHOLOGY: CONCEPTS AND METHODS
KURT LEWIN

ABSTRACT
The fieldtheoreticalapproachis instrumental in integratingdivergentphysiological,
psychological,and sociologicalfactson the basis of theirinterdependence.To explain
social behaviorit is necessaryto representthe structureof the total situationand the
distributionof the forcesin it. Certaingeometriesare offeredas adequate forrepre-
sentingsocial fields.The problemof adolescence and the concept "social group" are
discussedin detailas examples.

The sociologists, I suppose,have reasonto be satisfiedwiththe

recenttrendsinpsychology. Traditionally,
mostpsychologistsseem
tohavefeltmoreorlessobligedto emphasizethebiologicalcharacter
oftheindividual, to believein therealityofphysicalandphysiologi-
but to be rathersuspiciousofsocialcategories
cal processes, and to
regardas mysticthosewhoclaimedthatsocialfactswereas realas
physicalones.
Recently,however,a growingnumberof psychologists seemto
have abandonedthisview. They seemto be persuadedthatsocial
factsare equallyor evenmoreimportant forpsychology thanthe
so-called"physiological facts."These psychologistsrecognizethat
the childfromhis firstday of lifeis objectivelya partof a social
settingand woulddie withina fewdaysifhe wereto be withdrawn
fromit. Also,theso-called"subjective"psychological worldof the
individual,his life-space,is influencedin a muchearlierstageby
socialfactsand socialrelationsthananyonewouldhave expecteda
fewdecadesago. Already,at a fewmonths, thechildseemsto react
to anotherperson'ssmileand voice in a ratherspecificway. It is
probablysafeto say thatthe childis able to perceiveand to dis-
tinguish thefriendliness and unfriendlinessofanotherpersonat an
earlierage thanhe is ableto distinguish thepatternofphysicallines
in a countenance whichexpresses thesesocialattitudes.
Beginning withthisearlyage, the child'sbehavioris moldedin
everyrespectby hissocialsituation.Of course,his morale,his re-
ligion,and hispoliticalvaluesaredetermined byhisbeinga partof,
and reactingto, thesocietyin whichhe lives. If one considersthe
868

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 869

findingsof culturalanthropology and of experimental psychology,

onecan,I think,establishevidencethatsocialinfluences enterevery
actionof the individual,evenactionswhichseemto have nothing
todowithsociety.
Humanbehavioris eithera directedactionor an emotionalex-
pression.Experimental psychology has shownthattheformation of
goals dependsdirectlyupon the laws whichgovernthe level of
aspiration,particularly uponthe effect whichsuccessor failurehas
in raisingand loweringthe level of aspiration(7, IO, I3).' These
experiments makeit evidentthatthe levelof aspirationis greatly
influenced by suchsocial factsas thepresenceor absenceof other
personsor by the competitive or noncompetitive characterof the
situation.It has been shown,too, that the goal-setting depends
upon certainideal goals,upon whatthe sociologists call the "ide-
ology" of the person.Culturalanthropology proves that these
ideologiesvaryextremely amongdifferent cultures.As to theemo-
tionalexpression, experiments have shownthat,forinstance,the
emotionalreactionto failurecan be changedto a greatextentby
appropriatepraise or changein social atmosphere(7, I5). This
substantiates thegeneralthesisthatthemanagement oftension(g)
by the individualdependsupon his particularsocial and cultural
setting.
Fromthisit shouldbe apparentthatexperimental psychology is
instrumental in helpingthesociologists to realizetheirmostardent
dream:the demonstration of the fundamental, direct,and wide-
spreadeffect ofsocialfactsonbehavior.
Thereis a growingnumberof psychologists who emphasizethe
"historical,"social side of psychological facts;and eventhe hard-
boiledbelieversin a stimulus-reaction psychology showa peculiar
interestin gettingas muchof,and as closeto, socialfactsas they
can. I believethereis no longeranyneedforthetraditional oppo-
sitionbetweenpsychologists andsociologists inthisbasicissue.
I
intothesocialdependency
thisinsight
Unfortunately, ofbehavior
doesnotendtheproblemforthepsychologist. rather
His problems
I Numbersin parenthesesreferto workscited in the bibliography
at theend of the
article.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 870 THE AMERICAN JOURNAL OF SOCIOLOGY

beginhere.For the sociologist, too, theyshouldbeginhere.Psy-

chology,includingsocial psychology, cannotpossiblybe satisfied
withany "generalities"(howevercorrecttheymay be). It has to
judgescientific conceptsand theorieslargelyby theirabilityor in-
abilityto handle problemsof dynamicinterdependence and to
handlethemin a mannersufficiently to attacktheconcrete
specific
tasksofthelaboratory ortheclinic.
Of course,forhundredsof yearsthe beliefwas prevalentthat
personality, will,and emotionwerenot subjectto strictlaws and
that theycould not be studiedexperimentally. A similarview is
traditionally strongin sociology.In the long run,however,dira
necessitasis boundto be stronger in bothsociologyand psychology
thanthosemetaphysical prejudices,and sociologyseemsto be ready
now forimportant stepsaway fromtheseprejudices.Psychology
as a sciencemightbe said to be somewhatmoreadvancedtechni-
callyand conceptually, at leastinsomeofitsareas. However,on the
whole,and particularly in regardto social psychology, it, too, is
facingthetaskofdeveloping a generalapproachwhichoffers specific
conceptualtools forsolvingthe concreteproblems of a vast and
diversified area.
Socialpsychology indicates, probablybetterthananyotherpart
ofpsychology andofsociology, whatis needed.Its progress depends
upon an overcoming of certainmajordifficulties, whichincludeat
leastthefollowing:
a) The integrating of vast areas of very divergentfactsand aspects: The de-
velopmentof a scientificlanguage (concepts) which is able to treat cul-
and physicalfactson a common
tural,historical,sociological,psychological,
ground
b) The treatingofthesefactson thebasis oftheirinterdependence
c) The handlingofbothhistoricaland systematical problems
d) The handlingofproblemsrelatedto groupsas wellas to individuals
e) The handlingof all "sizes" of objects or patterns(social psychologyhas to
includeproblemsof a nationand its situation,as well as of a play groupof
threechildrenand theirmomentary struggle)
f) Problemsof"atmosphere"(suchas friendliness, pressure,etc.)
g) Experimentalsocial psychologywill have to finda way to bringthe large-
size patternsintoa framework small enoughforthetechnicalpossibilities
of
experimentation

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 87I

The varietyof factswhichsocialpsychology has to treatmight

reallyseemfrightening to evena boldscientificmind.Theyinclude
"values" (suchas religious and moralvalues),"ideologies"(such as
conservatism or communism), "the styleof livingand thinking,"
and other so-called "cultural" facts.They include sociological
problems, i.e., problemsofgroupand groupstructure, theirdegree
ofhierarchy and typeoforganization; or suchproblemsas thedif-
ference betweena ruraland an urbancommunity, theirrigidityor
fluidity, degreeof differentiation,
etc. They also includeso-called
"psychological" problems,such as the intelligenceof a person,his
goals and fears,and his personality. They includesuch "physio-
logical"factsas theperson'sbeinghealthyor sick,strongor weak,
the colorof his hairand of his complexion. They include,finally,
such "physical"factsas the size of the physicalarea in whichthe
personora groupis located.
It is utterlyfruitless
and merelya negativescientific treatmentto
put thesefactsinto classificatory pigeonholes, howeveraccurately
builtand fittedtheymay be. It is widelyacceptedtodaythatwe
needpositivemeansofbringing thesevarioustypesoffactstogether
in sucha waythatone can treatthemon onelevelwithoutsacrific-
ingtherecognition The problemof
oftheirspecificcharacteristics.
adolescencewhichwe will discussas an exampleshows,I think,
particularlyclearlythat a way must be foundto treat bodily
changes,shiftofideology, andgroup-belongingness withinonerealm
of scientific language,in a singlerealmof discourseof concepts.
The questionis "How canthatbe done?"
Behaviorismhas triedto answerthis questionby interpreting
everything as a conditioned reflex.One ofthemainreasonsforthe
appeal of suchapproachis the same as thatwhichlies behindthe
popularappealofthe"unityofscience"idea: namely, it appearedto
put everyproblemon a "physiological" basis (althoughin factit
did not), and in thisway it seemedto promiseintegration of the
divergent factsononelevel.
Today mostresearchworkers in sociologyand socialpsychology
will agree that the programof describingand explainingsocio-
psychological processesby conceptsand laws of physicsor physi-

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
872 THE AMERICAN JOURNALOF SOCIOLOGY

ologymightat best be something to talkaboutas a distantpossi-

bilityfora speculativephilosopher. But sucha waywoulddefinitely
notbe a realisticresearch programforattacking thesociopsychologi-
cal problems oftoday. On theotherhand,to elaborateon the"fun-
damentaldifferences" betweenphysics,sociology,and psychology
and to restsatisfied
withsuchdistinctions is no helpeither.
To discussthese problemsadequatelywould involvea more
thoroughtreatment of certainquestionsof comparativetheoryof
sciencethanis possiblehere.As faras I can see thesolutionliesin
the direction(a) that a scienceshouldbe considereda realmof
problemsratherthan a realmof material;(b) that the different
realmsofproblems mightnecessitate different
universesofdiscourse
of constructs and laws (such as thoseof physics,aesthetics,psy-
chology,and sociology);and (c) that any one of them refers
moreorlessto thesameuniverse ofmaterial.
For anypracticalpurposeofresearch-andthat,afterall,is what
counts-scienceslikesociologyor psychology shouldfeelfullyfree
to use thosetypesofconstructs whichtheythinkmostadequatefor
handlingtheirproblems;and theyshouldattemptto findtheinte-
grationwe have discussedon theirownlevel. Theyshouldnotfeel
obligedto use constructs ofanothersciencemerely outofphilosophi-
cal reasons(e.g.,becausesomephilosophies or popularmetaphysics
apply"truereality"to physicalentitiesonly). On theotherhand,
feelingconfident in theirown right,thosesciencesdo not need to
be afraidofusingmethodsorconcepts(e.g.,mathematical concepts)
whichmightor mightnot have similarities with thoseof other
sciences.
approachis intendedto be a practicalvehicle
The field-theoretical
ofresearch.As is truewithanytool,itscharacteristics canbe under-
stoodfullyonlyby the use of it in actual research.Therefore, in-
stead of statinggeneralmethodological principles in abstractum,I
preferto discusstheproblemof adolescenceand thedefinition ofa
Our purposein discussingthemis
social groupas an illustration.
nottheproving ofcertainfactsortheories(whichmightormightnot
be fullycorrect)but to surveycertainmajoraspectsof the field-
theoreticalapproach.In discussing theseexamplesI willtherefore,
fromtimeto time,pointto similaraspectsin otherproblems.A

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORYIN SOCIALPSYCHOLOGY 873

forthcoming articleby Mr. Lippitt2offersa moredetailedexample

ofactualresearch.
II
We have chosentheproblemof adolescencebecausethechanges
in behaviorwhichare supposedto be characteristic forthisperiod
seem,at firstsight,to giveexcellent backingto a biologicalviewin
sociology.Obviously,adolescencehas something to do withsexual
hormonesand with certainperiodsof bodilygrowth.The more
recenttreatments of theproblemof adolescence,however,seemto
emphasizeitssocialaspect. Theypointparticularly to thefactthat
the behaviortypicalforthisage is ratherdifferent in different so-
cieties(4, 23). Considerableargumentation has been advancedfor
andagainstbothviews.
However,in regardto theproblemof adolescence, as in relation
to othersocialand psychological problems, it doesnothelpmuchto
arguewhetheradolescenceis a biologicalor psychological effect.It
does not help verymucheitherto tryto describe,on a statistical
basis,to whatdegreethisproblemis biologicalor psychological in
nature.Even if an answercouldbe found,it wouldbe of as little
value as, for instance,the determining of the degreeto which
heredityand environment affectintelligence.We stillwould not
have gainedany insightinto the way in whichbodilyand social
factorsareworking togetherand againsteachother,integrating the
concretebehaviorof the adolescent.It would seem to be more
fruitful to startwithan analysisof the settingin a concretecase.
This case shouldbe chosennotso muchaccordingto thefrequency
of occurrence as accordingto theamountofinsightit offers intoa
constellation whichis typicalat least fora part of the settingin
question.
In regardto the problemof adolescence,it mightbe helpfulto
referfirstto caseswhichshowtheso-called"typical"difficulties of
adolescentbehavior.A field-theoretical analysisofsucha situation
shouldgivesomehintsas to whatconditions wouldincreaseor de-
creasethesesymptoms.
The periodofadolescencecan be said to be a periodoftransition.
2
Scheduledforpublication Journalof
in the July,1939, issueof the American
Sociology.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
874 OF SOCIOLOGY
THE AMERICANJOURNAL

It seemsto imply,at least undercertaincircumstances, a more

rapidor deepershiftthan the periodbefore.Afterthe ratherim-
portantchangesaroundthe age ofthreeyears,oftena morestable
situationhas arisen.Maybe minorcriseshave comeup; but par-
ticularlyin caseswherethe adolescenceis characterized by special
disturbances, quietor stabletimemighthave preceded
a relatively
it. If one triesto characterizethenatureof thetransition, one can
pointtoseveralaspects.
a) One can viewadolescenceas a changein group-belongingness.
The individualhas been considered by himselfand by othersas a
child.Now he does notwishto be treatedas such. He is readyto
separatehimselffromthingschildishand to tryseriouslyto enter
adult lifein mannersand in outlookon occupation,as on lifein
general.Anychangein belongingness fromone groupto anotheris
ofgreatimportance forthebehavioroftheperson;themorecentral
forthepersonthisbelonging is, the moreimportant is the change.
A shiftingroup-belongingness is a "sociallocomotion."Thatmeans
it changesthepositionofthepersonconcerned.
It is a simplefact,but stillnot sufficiently recognizedin psy-
chologyand sociology, thatthebehaviorofa persondependsabove
all upon his momentary position.Often,the world looks very
different before and after an eventwhichchangesthe regionin
whicha personis located.That is the reasonwhy,forinstance,a
fait accompliis so fearedin politics. A changein position,forin-
stance,thelocomotion fromonegroupto another,changesnotonly
themomentary surroundings of a personbut moreor less thetotal
setting:whathas beena neighboring region,easilyaccessiblefrom
thepreviousposition,mightnowbe farther away or no longerac-
cessibleat all. On theotherhand,different regionsare nowneigh-
bors,and newonesmaybe accessible.The shiftintothegroupof
theadults,forinstance,makespossiblecertainactivities whichpre-
viouslywereforbidden but whichare nowsociallypermitted. The
individualmightattendcertainparties,have accessto certainac-
tivities.On theotherhand,certaintaboosexistfortheadultsthat
do notexistforthechild(Fig. i, a and b).
b) The changefromthegroupofchildren to thatoftheadultsis a
shiftto a moreor less unknownposition.Psychologically, it is

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 875

an unknownregion,comparableto coming
equivalentto entering
intoa newtown.Experiments forexample,
in thefieldoflearning,
givesomekindofpictureofthefundamental differences
betweena

FIG. i.-Comparison of the space offreemovement of child and adult. The actual
activityregionsare represented. The accessible regionsare blank; the inaccessible
shaded. (a) The space offreemovementofthechildincludestheregionsI-6 represent-
ing activitiessuch as gettinginto the movies at children'srates,belongingto a boy's
club,etc. The regions7-35 are not accessible,representing activitiessuch as drivinga
car, writingchecks forpurchases,political activities,performanceof adults' occupa-
tions,etc. (b) The adultspace of freemovementis considerablywider,althoughit too
is boundedby regionsofactivitiesinaccessibleto the adult,such as shootinghis enemy
or enteringactivitiesbeyondhis social or intellectualcapacity (representedby regions
including29-35). Some of the regionsaccessible to the child are not accessible to the
adult, forinstance,gettinginto the moviesat children'srates,or doing thingssocially
taboo foran adult whichare permittedto the child (representedby regionsI and 5).

situationwhichis familiarto an individualand thatwhichis un-

familiar.The unfamiliarcan be represented as a
psychologically
unstructured
cognitively region(i6). This meansthatthatregion

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
876 THE AMERICAN JOURNALOF SOCIOLOGY

is not differentiated into clearlydistinguishable parts. It is not

cleartherefore wherea certainactionwilllead and in what direc-
tionone has to moveto approacha certaingoal. Thislackofclear-
nessofthedirection in thefieldis one ofthemajorreasonsforthe
typical"uncertainty ofbehavior"to be foundinunknown surround-
ings. Studieson social pressureand on ascendantand submissive
behavior(I 2, 27) clearlyindicatethatan individualin an unfamiliar
surrounding is lessreadyto putup a fightor to showascendantbe-
havior.An unfamiliar surrounding is dynamically equivalentto a
softground.Or,to be morespecific, thelack of a cognitively clear
structure is likelyto makeeveryactiona conflicting one. The in-
dividual,not knowingwhetherthe actionwill lead him closeror
farther away fromhis goal,is necessarily uncertainas to whether
ornothe shouldcarryit out (i6).
The child'sdevelopment naturally leads to an openingup ofnew
unknownregions.Periodsof transition are characterized by more
thanthe usual impactof suchnew regions.Enteringa newsocial
groupcan meansomething verysimilarto beingthrownintoa cog-
nitivelyunstructured field,beingforcedto standon unfirm ground
and notknowing whether the"rightthing"is beingdone. The un-
certaincharacterof the adolescent'sbehaviorand his conflicts can
partlybe explainedby thelack of cognitiveclarityconcerning the
adult'sworldwhichhe is goingto enter(Fig. 2). It clearlyfollows
that thisuncertainty is greaterthe morethe individualhas pre-
viouslybeenkeptout of the adultworldand has beenkeptin the
darkaboutit.
c) One regionparticularly closeand important to theindividual
is his ownbody. Psychologically one's ownbodycan be treatedin
somerespectsin thesamewayas one's environment. Generally the
individual"knows"his body sufficiently. That means he knows
whathe can expectfromit and how it willreactundergivencir-
cumstances.The timeof sexual maturitybringswithit changes
whichmakethe individualsometimes disturbedby his own body.
Moreor less strangeand newbodyexperiences ariseandmakethis
partof the life-space whichis so close and vital to the individual
strangeandunknown.In thiscase thechangedoesnotmeanmerely
theusual uncertainties of a newand strangeenvironment; but,in

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 877

addition,a regionwhichpreviously appearedto be wellknownand

reliablebecomesnow unknownand unreliable.This changeneces-
sarilyshakesthebeliefoftheindividualinthestability
oftheground
on whichhe standsandperhapsevenin thestability oftheworldat
large. Sincethe regionof the bodyhappensto be veryimportant
and centralforanyone,thisdoubtingmightbe ratherfundamental.
It mightlead,on theonehand,to increaseduncertainty ofbehavior
and to conflicts;
on theother,to the aggressiveness
of someof the
adolescentreactions.

I , 9

FIG. ofthe adolescent

2.-The space offreemovement as it appears to him. The space
offreemovement is greatly
increased, manyregionswhichpreviously
including have
notbeenaccessibleto thechild,forinstance,
freedom
to smoke,returning
homelate,
drivinga car (regions7-9, II-I3, .. ). Certain regionsaccessible to the adult are
clearlynot accessible to the adolescent,such as voting(representedby regionsiO and
i6). Certainregionsaccessible to the child have already become inaccessible,such as
gettingintothe moviesat children'srates,orbehavingon too childisha level (regioni).
The boundariesofthesenewlyacquiredportionsofthespace offreemovementare only
vaguelydeterminedand in themselvesgenerallyless clearlyand sharplydifferentiated
than foran adult. In such cases the life-spaceof the adolescent seems to be full of
possibilitiesand at the same time of uncertainties.

Suchexplanation wouldbe in line,e.g.,withthefindingsofL. B.

Murphy(22) thatinsecuresituations lead bothto highlyaggressive
andhighlysensitivebehavior.The disastrouseffectwhichthebreak-
downofa previously firmgroundmighthave,is dramatically illus-
tratedby foster-children,
who discoverat a late age thetruefacts
concerningtheirparentage.The traumaofsucha collapseofa social
groundsometimes permanently destroys theirbeliefintheworld.
d) The "radicalism"whlchmakes some adolescentsflockto
extreme"left"or "right"politicalpartiesand be extremein many

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
878 THE AMERICAN JOURNALOF SOCIOLOGY

judgments has to deal also witha secondfactor.A periodofradical

changeis naturally a periodofgreater plasticity.The veryfactthat
a personis in thestateofmovingfromone regionA to a newregion
B, and is therefore cut loose fromthe regionA but notyetfirmly
established in theregionB, putshimin a less stableposition(Figs.
2 and 4) and makeshim,as anyobjectinstatu nascendi, moreforma-
tive.
The psychological environment has to be regarded functionally as
a partofone interdependent field,thelife-space, theotherpartof
whichis the person.This fundamental factis the keynoteof the
field-theoreticalapproach.In psychology it has become,in various
forms, moreand morerecognized and can be expressedsimplyby
the formula:Behavior= Functionof personand environment =
Function oflife-space (Be = F [P, E] = F [L Sp]) (3, i8). Thein-
stabilityof the psychologic environment leads, in some respects,
therefore, to greaterinstability of the person."Beingestablished"
meanshavinga well-defined positionand definite relationsto the
manyregionsof a highlydifferentiated life-space:undersuchcir-
cumstances anymajorchangemeansa greatnumberofstepsand a
shiftofinterrelation. In an unestablished, newsituationthefieldis
not verymuchdifferentiated, and whateverdifferentiation has oc-
curredis notveryfirm.The shiftofpositionoftheindividualfrom
one regionto another, whichin thelessdifferentiated fieldmightbe
merelyone step (Fig. 2), wouldhave to be considereda major
change(equivalentto manysteps) in a moredifferentiated field
(Fig. i, b). Similarly,whatin realityis a notverygreatand easily
made shiftin cognitivestructureof the ideologicalfieldof the
adolescent,whichcontainsrelativelyfewregions,appearsto be a
radicalshiftto the adult,withhis highlydifferentiated cognitive
field.The difference in cognitivedifferentiation is probablyone of
thereasonswhyadolescents easilygo to extremes.
e) The widening of thelife-space intounknownregionsconcerns
not onlygeographical surroundings (interestin traveling, hiking,
etc.) and social surroundings (more inclusivesocial groupslike
politicalor occupationalones) but also the timedimensionof the
life-space.Personsofall agesareinfluenced by themannerin which
theysee thefuture, thatis,by theirexpectations, fears,and hopes.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 879

The scopeof timeahead whichinfluences presentbehavior,and is

therefore to be regardedas a partofthepresentlife-space, increases
duringdevelopment. This changein timeperspective is one of the
mostfundamental factsof development. Adolescenceseemsto be
a periodofparticularly deepchangeinrespectto timeperspective.
The changecan be partlydescribedas a shiftin scope. Instead
ofdays,weeks,ormonths, nowyearsaheadareconsidered in certain
goals. Even moreimportant is thewayin whichthesefutureevents
influence presentbehavior.The ideas of a childof six or eightin
regardto his occupationas an adult are not likelyto be based on
sufficientknowledgeof the factorswhichmighthelp or interfere
withthe realizationof theseideas. They mightbe based on rela-
tivelynarrowbut definite expectations or mighthave a dreamor
playlikecharacter.In otherwords,"ideal goals" and "real goals"
forthedistantfuture arenotmuchdistinguished, andthisfuture has
morethefluidcharacter ofthelevelofirreality.
In adolescencea definite in regardto thetimeper-
differentiation
spectiveis likelyto occur.Withinthosepartsofthelife-space which
represent the future,levels of realityand irrealityare gradually
beingdifferentiated. That whichis dreamedofor wishedfor(level
ofirreality in thefuture)becomesseparatedfromwhatis expected
(levelofrealityin thefuture).Vagueideas have to be replacedby
moreorlessdefinite decisionsin regardto preparation forfuture oc-
cupation.In otherwords,one has to "plan": to structure thetime
perspective (io) in a waywhichis in linebothwithone'sownideal
goals or values and withthoserealitieswhichmustbe takeninto
accountfora realisticstructuring of theplaneof expectation (Fig.
3, a andb).
This taskis characteristic forall kindsofplanning.The situation
oftheadolescentin thisrespectis particular onlyin thathe has to
formthe timeperspective in regardto a fieldwhichis especially
greatand unknown.Whathe learnsfrombooksand adultcouncil
aboutwhatan individualmightaccomplish is fullofcontradiction:
the adultspraisethe herowhohas realizedwhatseemedto be im-
possible,and at the sametimepreachthemoralof "standingwith
bothfeeton theground."
In anotherrespectthe adolescentfindsthe adults (thegrouphe

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 a

ps pr Psfu'

3le:

b
FIG. 3. Time perspective ofchildandadult.The diagramrepresents thelife-space P as
ofan individual
it existsat a certaintime. At thattimethepersonsees himselfin a certainsituiationat presentand expects
certainsituations in hisimmediate or laterfuture(represented as "psychological time"ps pr= psycho-
logicalpresent, psful= immediate psychologicalfuture, ps fu2= later psychological future).Probablyeach
of thesesituationsincludesa realitylevel (R) equivalentto what thepersonreallyexpectsto cometrue,and
an irrealitylevel (Ir), correspondingto what he mightdream,hope for,or fear(forinstance,on theirreality
level the person,P, sees himselfcloserto the goal thanhe expectsto be at that time). Mathematicallythe
reality-irrealitydimensionand the past-present-future dimensionmakes the life-spaceexistingat a given
momenta manifoldwhichhas at least fourdimensions(to representchangeoflife-spacewould mean a fifth
one). (It is, of course,impossiblemathematicallyto representin a diagrama continuousfour-dimensional
manifoldadequately, but we trust that this discontinuousrepresentationwill suffice.)
(a) Life-spaceof a child. (b) Life-spaceof an adult. !'(I) The adult life-spaceshowsa greaterdegreeof
differentiation in thedimensionreality--irrealityfora givenpsychologicaltime,forinstance,forthepsycho-
logicalpresent.(2) The timeperspective oftheadultinfluencing hispresentbehaviorgenerallycoversa
largertimespan and is moredifferentiated in regardto timesequence. (3) The adultis generallymoreable
to distinguishbetweenwishesand realisticexpectation.That is trueparticularly forthefuture;thestructure
of the realitylevel of his life-spacein the psychologicalfutureis less directlydependentupon theirreality
level (althoughforthe distantfuturethisinfluencemightbe not much less than fora child). For instance,
the positionin whichtheperson,P, sees himseffon the futurerealitylevel is forthe child probablymore
similarto thaton theirrealitylevel thanforthe adult.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 88i

is to enter)fullof contradiction.A varietyofconflicting religious,

political,and occupationalvaluesis obviouslypowerful withinthat
group.A childmay failto bringto adolescencea well-established
framework ofvalues,orhe mayhavethrown thevaluesofhischild-
hoodaway. In eithercase thestructure ofhis adolescenttimeper-
spectivewillbe unstableandundetermined, owingtotheuncertainty
ofnotonlywhatcan be done(whichwe have discussedpreviously)
but also whatshouldbe done. The uncertain character oftheideals
and values keepsthe adolescentin a stateof conflict and tension
whichis thegreaterthemorecentraltheseproblemsare. The wish
to structure thesefieldsin a definiteway (and in thismannerto
solvetheconflict) seemsto be one of thereasonsbehindthereadi-
nessoftheadolescentto followanyonewhooffers a definite
pattern
ofvalues.
f) The transition fromchildhoodto adulthoodmay be a rather
suddenshift(forinstance,in someof theprimitive societies),or it
mayoccurgraduallyin a settingwherechildrenand adultsare not
sharplyseparatedgroups.In caseoftheso-called"adolescencediffi-
culties,"however, a thirdstateofaffairsis oftenprevalent:children
and adultsconstitute clearlydefined
groups;theadolescentdoesnot
wishto belonganylongerto thechildren's groupand, at thesame
time,knowsthathe is not reallyacceptedin the adultgroup.In
thiscase he has a positionsimilarto whatis calledin sociologythe
"marginal man."
The marginalmanis a personwhostandson theboundary(Fig.
4, b) betweentwogroups,A and B. He doesnotbelongto eitherof
them,or at leasthe is notcertainabouthisbelongingness. Not in-
frequently thissituationoccursformembersof an underprivileged
minority group,particularly forthemoreprivileged members within
thisgroup.Thereis a strongtendency forthemembers oftheunder-
privileged minority groupto cut loose and to tryto enterthema-
joritygroup(I 7). In case thepersonis partlysuccessful in estab-
lishingrelationships withthe privilegedgroupwithoutbeingfully
accepted,he becomesa marginal man,belonging to bothgroupsbut
not fullyto eitherof them.The factof beinglocatedin a social
''no man'sland" can be observedin verydifferent typesofminority
groups-forinstance, racialgroupsorthehard-of-hearing, whichis a
marginal groupbetweenthedeafand thenormalgroup.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
882 THE AMERICAN JOURNAL OF SOCIOLOGY

Characteristicsymptomsof behaviorof the marginalman are

They tend to unbalanced
emotionalinstabilityand sensitivity.
too much
or shyness,exhibiting
behavior,eitherto boisterousness
shiftbetweenextremes
tension,and a frequent be-
ofcontradictory

b
FIG. 4.-The adolescent as a marginal man.(a) Duringchildhood and adulthood
the"adults"(A) and"children" (C) areviewedas relatively separatedgroups,thein-
dividualchild(c., C2) and theindividualadult(a', a2) beingsureoftheirbelonging to
theirrespective groups.(b) The adolescent belonging to a group(Ad) whichcan be
viewedas an overlapping regionofthechildren's(C) andtheadults'(A) groupbelong-
ingto bothofthem,oras standing between them,notbelonging toeitherone.

havior.The marginal manshowsa typicalaversionto thelessprivi-

legedmembersof his owngroup.This can be notedin the hostile
attitudeof somesubgroupsof the Negroesor otherraces against
members oftheirownrace,andthehard-of-hearing againstthedeaf.
To someextentbehaviorsymptomatic forthemarginalman can
easilyshifted
be foundin the adolescent.He too is oversensitive,
fromone extremeto the other,and particularly sensitiveto the

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 883

shortcomings ofhisyounger fellows.Indeed,hispositionis sociologi-

callythe same as thatof the marginalman: he does not wishto
belonganylongerto a groupwhichis, afterall,lessprivileged than
thegroupof adults,but at the same timehe knowsthathe is not
fullyacceptedby the adults.The similarities betweentheposition
ofthemembers oftheunderprivileged minority and theadolescent
and betweentheirbehaviorseemsto me so greatthatone might
characterize thebehaviorofthemarginalmembers oftheminority
groupas thatofpermanent adolescence.
We mightsumup our discussion of theadolescentin thefollow-
ingmanner:
a) The basic factconcerning thegeneralsituationof theadoles-
centcanbe represented as thepositionofa personduringlocomotion
fromone regionto another.This includes(i) the wideningof the
life-space(geographically, socially,and in timeperspective[Figs.
I, 2, and 31) and (ii) the cognitivelyunstructured characterof the
newsituation(Fig. 2).
b) Somewhatmorespecifically, the adolescenthas a socialposi-
tion"between"theadultand thechild,similarto a marginalmem-
berofan underprivileged minority group(Fig. 4, b).
c) Thereare stillmorespecificfactorsinvolvedin adolescence,
such as the new experiences withone's own body,whichcan be
represented as the baffling changeof a centralregionof the estab-
lishedlife-space.
Fromthisrepresentation onecan deriveconceptually:
I. The adolescent'sshyness,sensitivity, and aggressiveness,ow-
ingto unclearness and instabilityofground(followsfroma, b, and
c).
II. A moreor less permanent conflictbetweenthe variousatti-
andstylesofliving(follows
tudes,values,ideologies, fromb).
III. Emotionaltensionresultingfromthese conflicts(follows
froma, b,andc).
IV. Readinessto takeextreme attitudesand actionsand to shift
froma, b,andc).
hispositionradically(follows
V. The "adolescentbehavior"shouldappearonlyifthestructure
and dynamicofthefieldis suchas represented by a, b, and c. The
degreeand particulartypeofbehaviorshoulddependuponthede-

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
884 THE AMERICAN JOURNALOF SOCIOLOGY

greeofrealizationofthisstructureanduponthestrength ofthecon-
flictingforces.Aboveall, thedegreeofdifferenceand ofseparation
betweenadultsand children whichis characteristic
fora particular
cultureis important;also,theextentto whichtheparticular adoles-
centfindshimself in thepositionof a marginalman. According to
fieldtheory,actual behaviordependsupon everypartof the field.
It followsthatthedegreeofinstability ofthe adolescentshouldbe
greatlyinfluenced also by such factorsas generalstabilityor in-
stabilityoftheparticularindividual.

III
BeforeI discussthemethodological aspectofourexample,I wish
by an additionalexampleoneparticular
to illustrate point,namely,
thecharacterizations ofeventsand objectsbytheirinterdependence
ratherthan by theirsimilarity or dissimilarity
of appearance.Al-
readyin the exampleof adolescence,onlysucha proceduremade
possiblethe linkingof such divergentfactorsas group-belonging-
ness,bodilychanges,andattitudes.
To mymind,it is hopelessto linkthedifferent problems involved
in socialpsychology in a propermannerby usingclassificatory con-
ceptsof thetypeof theLinneansystemin botany.Instead,social
psychology will have to use a framework of "constructs."These
constructs do notexpress"phenotypical" but so-called
similarities,
"dynamical"properties-properties definedas "typesofreactions"
or "typesofinfluences." In otherwords,theseconstructs represent
certaintypesofinterdependence. The transition fromphenotypical
conceptsto dynamic(genetic, conditional-reactive)constructs
based
on interdependence is, to mymind,one ofthemostimportant pre-
requisitesfor any sciencewhichwishesto answerquestionsof
causation.Psychology is in the midstof a processof transition to
of and
thistype concept.Socialpsychology, sociology too,willhave
to turndefinitely in thisdirection.It is truethatsucha transition
can be madeonlyifand whenthereis a sufficient amountofpheno-
typical"facts"gathered andclassificatory workhasbeendone. This
state,however, seemsnowto havebeenreachedbothin psychology
andinsociology.
As an exampleof the typeand importance of thisshiftto con-

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 885

structsbasedon interdependence, I mightpointto thedefinition of

"social group." Since this concept will be used extensivelyin the
articlebyMr. Lippitt,I canlimitmyself to briefremarks.
The definition of the concept"group"has a somewhatchaotic
history.The termis interwoven withphilosophical and metaphysi-
cal considerations. Oneofthemainpointsofdiscussion waswhether
or not thegrouphas a group mind and is thereforean entityover
and above the individual.Besides this,the discussionwas domi-
nated frequently by the emphasisupon the difference between
Gemeinschaft and Gesellschaft,whetherone has to deal merelywith
mattersof formalorganization or whetherthereexistssomething
likea "naturalgroupunity,"basedon suchfactorsas empathy.
To thepsychologist whohas observedthehistorical development
of the conceptof "whole,"or Gestalt,in psychology, mostof the
argumentation about thegroupmindsoundsstrangely familiar.It
tookpsychology manystepsbeforeit was discovered thata dynamic
wholehas properties whichare different fromtheproperties oftheir
partsor fromthe sum of theirparts. Even relatively recently(in
the earlyGestaltpsychology) the statementwas frequently made
that"the wholeis morethanthesumof its parts." Today sucha
formulation can be considered hardlyadequate. The wholeis not
"more"thanthe sum of its parts,but it has different properties.
The statement shouldbe: "The wholeis different fromthesumofits
parts." In otherwords,theredoesnotexista superiority ofvalueof
the whole.Both wholeand parts are equallyreal. On the other
hand,the wholehas definite properties of its own. This statement
has lost all its magichalo and has becomea simplefactofscience,
sinceit was discoveredthatthisholdsalso forphysicalpartsand
wholes(Koehler,citedin 6). In addition,psychology todayrecog-
nizesthat there existwholes of all degrees of dynamic unity:there
exist,on the one extreme, aggregates of independent objects;then
wholesof smalldegreesof unity,of mediumdegreesof unity,of a
highdegreeofunity;finally, at theotherextreme, wholesofsucha
highdegreeofunitythatitis hardlyadequateto speakofparts.
Whatever hasbeenofscientific valueintheconceptofgroupmind
resolvesitselfintotheconcreteand familiar problems ofdynamical
wholesinsociology andsocialpsychology.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 886 THE AMERICAN JOURNALOF SOCIOLOGY

Conceiving of a groupas a dynamicwholeshouldincludea defi-

nitionofgroupwhichis based on interdependence of themembers
(orbetter,ofthesubpartsofthegroup).It seemsto me ratherim-
portantto stressthispointbecausemanydefinitions ofa groupuse
the similarity of groupmembersas the constituent factorrather
than theirdynamicinterdependence. Frequently,forinstance,a
groupis definedas composedofa numberofpersonswhoshowcer-
tain similarities,
particularly
a similarityof attitudes.I thinkone
shouldrealizethatsucha definition is principally
different froma
definitionofa groupbasedon interdependence ofitsmembers.It is
verypossiblethata numberofpersonshave a certainsimilarity-
forinstance, ofsex,ofrace,ofeconomic position,ofattitudes-with-
out beinga groupin thesenseofbeinginterdependent partsofone
socialwhole.For instance,womenall overthe world,or unskilled
workers, or the farmers,mayshowa certainamountof similarity.
It mightevenbe possibletopickouta groupofNegroesinLouisiana,
poor whitetrashin Kentucky,and peasantsin China withgreat
economicsimilarity. It mightbe properto distinguish, in thisre-
spect,"types" or "classes."3However,that does not implythat
thesenumbersof personsare interdependent to any greatextent.
One of the developments in moderntimesis forsome of these
economicalclassesto showan increasing degreeofinterdependence,
in
i.e., theyshow trends the directionof developmentto inter-
nationalgroups.
A group,on theotherhand,doesnotneedto consistofmembers
whichshowgreatsimilarity. As a matteroffact,it holdsforsocial
groups,as forwholesinanyfield,thata wholeofveryhighdegreeof
unitymaycontainverydissimilar parts.Doubtless,forinstance,a
man,wife,and babywithinone familymayshowmuchgreaterdis-
similarity than each of the membersof thisgroupshowsto other
individuals(babies,men,women)outsideofthisgroup.The degree
ofunityofa groupgoes,byno means,handin handwiththedegree
ofsimilarity ofitsmembers.It is typicalofwell-organized groupsof
highdegreeofunityto includea varietyofmembers whichare dif-
3 Theterm"socialclass"is usedgenerally bothaninterdependent
todesignate group
ofpersons
anda number whoshowsimilar properties.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 887

ferentandhavedifferent functions withinthewhole.Not similarity,

buta certaininterdependence ofmembers constitutes a group.
One shouldrealizethat even a definition of group-membership
by equalityof goal or equalityof an enemyis stilla definition by
similarity.The sameholdsforthedefinition ofa groupby thefeel-
ingofloyaltyor ofbelongingness oftheirmembers.However,such
an equality,as wellas theequalityofgoal or ofenemy,constitutes
sometimes, also,a certaininterdependence ofthepersonswhoshow
thesesimilarities. Therefore,ifone wishesto use the feelingofbe-
longingas thecriterion ofa group,onecan do so,ifonepointsto the
interdependence establishedby thisfeeling.However,one should
realizethatloyaltyorfeeling ofbelongingness is onlyoneofa variety
ofpossibletypesof interdependence whichmayconstitute a group
(othersare, e.g., economicdependence,love, livingtogetherin a
certainarea). The kind of interdependence of the members(i.e.,
whatholdsthegrouptogether) is equallyas important a character-
isticofa groupas thedegreeoftheirinterdependence and thegroup
structure.
Stressingsimilarity or dissimilarity,
ratherthaninterdependence,
is typicalforthe descriptive "classificatory" epoch,whichcan be
observedin a relativelyearlystageof development in practically
everyscience(I9). It governsalso,to a largedegree,theeveryday
thinking concerning groups.The discrepancy betweenwhatpeople
"shoulddo,iftheywouldbe guidedby theirrealinterest," andwhat
theyactually do is frequently causedby the factthata personfeels
himselfbelongingto thoseto whomhe is similaror to whomhe
wishesto be similar.On theotherhand,his "real interest"would
demandthathe shouldfeelbelonging to thoseuponwhomhis de-
pendenceis greatest.Thus, the behaviorof personsbelongingto
underprivileged groupscan hardlybe understood withoutrealizing
thatthemembership in sucha groupis determined by actualinter-
dependence but thatmanyunderprivileged personsfeelthemselves
(andoftenare)moresimilartopeopleoutsidethatgroup(I 7).
In relationto the problemof group-belongingness, as wellas to
anyothersocialproblem, onemustbecomesensitive to thedifference
betweenconceptsbasedon interdependence (including similarityof
interdependence) and thosebased on similarity withoutinterde-

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 888 OF SOCIOLOGY
THE AMERICANJOURNAL

pendence.I am persuadedthat in the furtherdevelopmentof

willmoreand moreper-
theformer
sociologyand socialpsychology
vadeandguide.
CONCLUSIONS
This cursoryexamination of theproblemof adolescenceand the
definition of "social group" is meant to illustratethe following
generalpointsconcerning thefield-theoreticalapproach:
a) It is possibleto linkin a definite mannera varietyoffactsof
individualand socialpsychology which,froma classificatory point
ofview,seemto have verylittlein common(suchas theprocessof
learningand orientation, timeperspective, planning,problemsof
individualmaturation, conflictsand tension,group-belongingness
and themarginal man,andbodilychanges).
b) This can be accomplishedby the use of constructswhich
characterize objectsand eventsin termsofinterdependence rather
thanofphenotypical similarity or dissimilarity. It may seemthat
emphasizing interdependence will make the problemof classifica-
tion even moredifficult because,generally, it is moredifficult to
in
describea fact terms of its effecton others and its being affected
byothers(itsconditional-genetic properties)thanin termsofitsap-
pearance(phenotypical properties).However,as soonas onegrasps
the idea, it becomesevidentthatif one characterizes an objector
the
eventby thewayit affects situation, everytype of factis placed
on thesameleveland becomesinterrelated to anyotherfactwhich
affectsthesituation.The problem ofwhether ornotoneis permitted
to combine,e.g., conceptsof values withthoseof bodilyweight,
vanisheswhenconfronted withthesimpletruththatbothfactsin-
fluence thesamesituation.
The transitionto constructswhich expressinterdependence
includes:
c) The systematization offactsby "classification" shouldgradu-
allybe replacedby an orderbasedon "construction," "derivation,"
and "axiomatization" oflaws(21).
d) It is possibleto takeintoaccount"general"trends,as wellas
more"specific"onesinvariousdegreesofspecificity (forinstance,to
linkthegeneralfactoroflocomotion fromone regionto anotherto
the morespecificone oflocomotion to an unknownregion,or to a

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD TIHEORY IN SOCIAL PSYCHOLOGY 889

locomotion fromone socialgroupto another, and finally to thestate

ofthemarginalman"between"twogroups).Insteadofpickingout
isolatedfacts,and lateron tryingto "synthesize"them,the total
situationis taken into accountand is represented fromthe be-
ginning.The field-theoretical approach,therefore, meansa method
of"gradualapproximation" by wayofa stepwiseincreasing specifi-
city.Pickingout isolatedfactswithina situationmaylead easily
to a picturewhichis entirely distorted.A field-theoretical represen-
tation,on theotherhand,can and shouldbe essentially correctat
anydegreeofperfection.
e) Whetheror nota certaintypeofbehavioroccursdependsnot
on the presenceor absenceof one factor of a numberof factsas
viewedin isolationbutupontheconstellation (structure and forces)
of the specificfieldas a whole.The "meaning"of the singlefact
dependsuponits positionin thefield;or,to say the samein more
dynamicalterms,the different partsof a fieldare mutuallyinter-
dependent.Thisis offundamental importance in socialpsychology.
It goesa goodwayin explaining, e.g.,theeffect ofruraland urban
surroundings (i, 13, 24), of nursery schoolsand orphanages(5, 25,
26),on thedevelopment ofintelligence,or,moregenerally, theeffect
of the stateof the environment (its degreeof differentiation, ten-
sion,etc.) on the stateof theperson,becausepersonand environ-
mentarebothpartsofonedynamical field.
f) The properties of a fieldas a whole,suchas its degreeof dif-
ferentiation, its fluidity,and its atmosphere, shouldbe emphasized
sufficiently.
g) The representation of sociopsychological factsby dynamic
constructs permitsderivationof the conditionswhich influence
behaviorin one direction or theotherand of the conditionsunder
which"exceptions"shouldbe expected.It coverstheusual case as
wellas theexceptional one.
h) If theviewsofthefield-theoretical approachare correct, there
is a goodprospectofapproaching experimentally a greatnumberof
problemswhichpreviously seemedout of reach: if the patternof
thetotalfieldis generally moreimportant than,forinstance,size,it
becomespossibleto studyfundamental socialconstellations experi-
mentallyby "transposing"theminto an appropriategroup-size.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
890 THE AMERICAN JOURNALOF SOCIOLOGY

(Gestaltpsychology understands by"transposition" a changewhich

leaves the essentialstructural characteristicsunaltered[6].) If the
experimenter is able to createsucha transposition, he doesnotneed
to be afraidof creating"artificial,""unlifelike"situations.(Ex-
periments becomeartificial ifmerelyone or anotherfactoris real-
ized,butnottheessentialpattern.)
Finally,I shouldlike to stresscertainmethodological points,
whichseemto meofparticular importance, one concerning thecon-
ceptualtoolsofsocialpsychology and sociology, anotherconcerning
theirresearchprogram, a thirdconcerning theirtechniqueof fact-
finding.
i) It is truethatall constructs inpsychology and sociology should
be operational(2); i.e., it shouldbe possibleto co-ordinate to each
of themobservablefactsor procedures.However,it is equallyim-
portantthatthe conceptualproperties of the constructs,i.e., their
logical-mathematical interrelations,be welldetermined. The latter
necessity, I think,has beenrelatively moreneglected inpsychology.
One of themostimportant amongtheseconceptualproblemsis
finding a geometry whichis able to represent the psychological or
socialfieldadequately.
Psychology has to dealwitha multitude ofco-existingfactswhich
areinterrelated andhavea relativepositionto eachother;inmathe-
maticalterms,it has to deal witha "space." Mathematics knowsa
varietyofdifferent typesofspaces. It is an empirical questionas to
what kindof geometry is best suitedto represent the dynamical
interdependence ofthatrealmoffactswhichis treatedina particular
science.SinceEinsteinit has beenknownthatEuclideangeometry,
whichpreviously was theonlygeometry appliedin physics,is not
best fitted for representing empiricalphysical space. For
the
psychology, a recently developednonquantitative geometry, called
"topology,"can be used satisfactorily in dealingwithproblemsof
structure and positionin a psychological field(i8). This spaceper-
mitsrepresentation of the positioninsideor outsideof a certain
region,therelationbetweenpartsand whole,and a greatnumberof
structural characteristics. All of this is done in a mathematically
exactway but does notpresupposethe quantitativedetermination
ofsize,whichis generally notpossiblein a psychological field.The

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 89I

topologicalspace is too "general"forrepresenting thosedynamical

psychological problemswhichincludethe conceptofdirection, dis-
tance,or force.Theycan be treatedwitha somewhat morespecific
geometry, whichI have called"hodologicalspace" (i6). Thisspace
permitsus to speakin a mathematically precisemannerofequality
and differences ofdirection, and ofchangesin distance,withoutpre-
supposingthe "measuring"of angles,directions, and distances,
whichis usuallynotpossibleina sociopsychological field.
It is, I suppose,beyondquestionthatsociology, too,dealswitha
"multitudeof co-existent interdependent facts"-in otherwords,
with the "empiricalspace."4The sociologistsand psychologists
shouldrecognize whathasbeenlongknown,thattheempirical space
is nothingotherthana multitude of factsexistingat a giventime
andshowing certaintypesofinterdependence. Indeed,sociology has
fora longtimeused a greatnumberof spatial concepts(such as
social approach,changein directionof action,etc.). The popular
prejudicethat the physicalspace is the only empiricalspace has
madesociologists regardtheirspatialconceptsas merelyan analogy.
Betterinsight intothemeaningofspaceinmathematics and physics
shouldreadilylead to the understanding that the social fieldis
actuallyan empirical space,whichis as "real" as a physicalone.
Euclideanspace generally is notsuitedforadequatelyrepresent-
ingthestructure ofa socialfield-forinstance,therelativeposition
ofgroups,or a sociallocomotion.For example,in a socialfieldwhat
is meantby a straight lineor an angleof 200 cannotbe determined
(at leastnotat present).However,thetopological and thehodologi-
cal spaceare,as faras I can see,applicablewithinsociology proper
as wellas in socialpsychology. For,in sociology, as in psychology,
oneis frequently able to determine relationsofpartsand wholeand
changesin distanceor directionwithoutbeingable to determine
quantitative relationsofsize,distance,or angle. In addition,these
geometries seemto be particularly suitableforrepresenting thepe-
culiar combination of and
"cognitive" "dynamic"factors, whichis
characteristic ofpsychological and socialfields,as wellas a number
4 This does not mean that everysociologicaltermwhichsoundsgeometricalis really
a geometricalconcept. The term "social distance," for instance,is probably not a
geometricalconcept(20).

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
892 THE AMERICAN JOURNAL OF SOCIOLOGY

of otherfundamental of the sociopsychological

properties dynamic
(i6).
The useofthesamekindofgeometry inpsychology and sociology
wouldnotimplythattheyareone and thesamescience.The ques-
tionof the "unity"of bothsciencescouldremainopen. However,
thetaskofsocialpsychology would,ofcourse,be greatlyfacilitated
bysucha similarity ofconceptual tools.
Independent ofthesolutionofthisproblem, sociology, as wellas
psychology, willhaveto decidewhatkindofgeometry it is goingto
applyin representing thespatialcharacteristics of its field.Before
thisquestionis answered, neithersociologynorpsychology can hope
to producescientific derivationsmoresolid than the "statistical
rules"basedon a co-ordination offactstreatedmoreorlesswithout
regardto theirparticular positioninspecificfields.
j) Both psychology and sociologycontain"historical"and ahis-
torical("systematical")problemscloselyinterwoven. Differently
frompsychology, sociology has beenfighting repeatedly againsttoo
greatan emphasison the historicalaspectof its problemsalmost
fromthebeginning. The transition to dynamicconstructs makesit
necessary to see thisproblemas clearlyas possible.It cannotbe the
task of sociologyor psychology to eliminatethe historicalside of
theirproblems.On the contrary, a field-theoretical
approachcan-
notavoidtakingintoaccountthehistorical character ofeveryfact
anditsspecific historical setting.
Nevertheless, it shouldbe recognized thatsystematic problems of
interdependence are different from historical of
problems origin.
The questionconcerning the "nature"and conditionsof a social
process-inotherwords,concerning "cause and effect"-isa syste-
maticonebothinpsychology andsociology.The first andmaintask
of a field-theoreticalapproachcan be characterized as thedetermi-
nation of "what situationsare empiricallypossible and which
situationsare not"; thisis identicalwiththe taskof finding laws.
For instance,does a dictatorship necessarily suppressdiscussion?
Does it need scapegoats? What formsof dictatorships or of de-
mocraciesare possible,and howdo theyaffect groupstructure, the
styleofliving,theideology,and individualbehavior?Questionsof
sucha systematic typeofcausationwillhaveto be answeredexperi-

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 893

mentallybeforethe dynamicaspect of "historical"problemsof

origincanbe treatedsatisfactorily.
k) Finally,a point concerning fact-finding shouldbe stressed
whichis technicalin naturebut nevertheless important fora field-
theoretical approach.It appliesto experimental as wellas to other
investigations.
It has alreadybeenemphasized thatthe validityof sociopsycho-
logicalexperiments shouldbe judgednot by theproperties of iso-
lated eventsor singleindividualswithinthe fieldbut mainlyby
whether or nottheproperties ofthesocialgroupor thesocialsitua-
tionas a wholeare adequatelyrepresented. This impliesthatone
of the foremosttasks of fact-finding and observationin social
psychology is to supplyreliabledata aboutthoseproperties of the
fieldas a whole.
How shouldthisbe done? Suppose,forinstance,thatthelifeof
a groupcontaining fivemembers wereto be observedduringa cer-
tain period.Let us assumethat fiveobserversare available.The
naturalprocedure mightseemto be to assignone observerto each
member ofthegroup,andinthiswaytogatherall thenecessary data
aboutthegrouplife. Generally, however, sucha procedure is hardly
thebestone. Whattheobservers willbringhomewillbe fiveminia-
ture"biographies" of fiveindividuals.It is truetheoretically that,
if thesebiographieswereto be perfectin securingall individual
data, and if,in addition,thetimeindexesforeveryactionwereac-
curateup to the second,the total grouplife mightbe "recon-
structed"on thebasis ofsuchmaterial.In reality,of course,these
biographies willbe neithercompletenorsufficiently accurateas to
time.As a rule,therefore, it willnotbe possibleto reconstruct even
suchsimpledata aboutgrouplifeas: a continuous recordofthesize
and character ofsubgroups, theirchange,and theirdegreeofunity.
Generally, thiswillbe as impossible as to constructmeaningfully the
behaviorandthepersonality ofan individual fromseparateaccounts
of the historyof his variousmuscles.Anyobservation necessarily
meansselection.The observer, confronted withthetaskofobserv-
ing an individual,naturallywill selectthosefactswhichare im-
portantfortheindividualeveniftheydo notmattermuchforthe
group.He simplywillnot"see" factsimportant forthegroupas a

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 894 THE AMERICAN JOURNALOF SOCIOLOGY

whole(e.g.,foritsorganization and atmosphere) iftheydo notim-

mediately strongly
reflect in theindividual's behavior.
At best,thedata aboutgroupproperties gatheredin thisfashion
on thebasisofindividual biographies are"indirectly reconstructed."
They cannotclaim to have the strengthof directobservations.
However,suchdirectobservations aboutproperties ofthegroupas a
whole arepossible.Frequently theycan be carriedthrough as easily
and as accuratelyas an observation on singleindividuals.In our
example,forinstance, it is possibleto assignoneofthefiveobservers
to directobservations of the subgrouping occurring in the group,
anotherone to recording thekindand characterof interactions. I
am persuadedthat,as a rule,forthe study of sociopsychological
problems sucha procedure is boundto be morefruitful and morere-
liablethantheassignment ofone observerto eachindividualofthe
group.
Ofcourse,giventhesocialdata,specific observation ofindividual
"biographies" mayproveveryvaluable. I do not doubt,however,
thatevenfortheunderstanding ofthecharacter and thebehaviorof
an individualthe firsttypeof observationwillgenerallybe more
significantthana recordoftheindividualwithoutthedata abouthis
socialbackground.Because the observation of thegroupwillpro-
videmoreandbettermaterialforthecharacterization oftheposition
and theroleofthisindividualwithinthegroup,theywilldetermine,
therefore,themeaning ofhisactionmoreaccurately thanwhatcould
be achievedby observing himmoreor less as a separatedentity.5
It wouldbe notat all surprising tomeifsucha sociological procedure
would becomea key techniqueeven forproblemsof individual
psychopathology.
The articleby RonaldLippitt,dealingwithexperiments on auto-
craticand democratic grouplife,willgivea moreconcreteillustra-
tion of how field-theoretical conceptsand techniquesmightbe
handledinactualresearch.
5 The stressing approachin regardto thetechnique
ofthefield-theoretical offact-
findingin socialpsychology thatundercer-
doesnot,ofcourse,excludethepossibility
tainconditions thebehavior can be treatedas a symptom
ofan individual forcertain
propertiesofthegroup.

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
 FIELD THEORY IN SOCIAL PSYCHOLOGY 895

BIBLIOGRAPHY
i. BALDWIN,BIRD T.; FILLMORE,EVA A.; and HADLEY, LoRA. Farm Children:
An Investigation
ofRural ChildLife in SelectedAreas ofIowa. New York:
D. Appleton-Century, I1930. Pp. xxii+337.
2. BRIDGEMAN,PERCY WILLIAM. The Logic of ModernPhysics. New York-
Macmillan, I932. Pp. 228.
3. BROWN,J. F. Psychology
and theSocial Order.New York: McGraw-Hill,
I936. Pp. V+529.
4. COLE,LUELLA.Psychology
ofAdolescence.New York: Farrar& Rinehart,
I936. Pp. ix+ 503.
5. CRISSEY,ORLOL. MentalDevelopmentas Relatedto Institutional
Residence
("Universityof Iowa Studiesin ChildWel-
and EducationalAchievement.
fare," Vol. XIII, No. I [II937). Pp. 8i.
6. ELLIS, WILLIS D. (ed.). A Source Book of GestaltPsychology.London:
Kegan Paul, Trench, Trubner & Co. Ltd., I938. Pp. ix+403.
7. FAJANS, SARA. "Erfolg, Ausdauer und Aktivitat beim Sauglings und
Kleinkind: Untersuchungzur Handlungs-und Affektpsychologie XIII"
["Success, Perseveranceand Activityin the Infant and Young Child:
Studiesin the PsychologyofActionand EmotionXIII"], ed. Kurt Lewin,
Psychologische XVII (I933), 268-305.
Forschung,
8. FRANK,JEROMED. "Some PsychologicalDeterminantsof the Level of
Aspiration," American Journal ofPsychology,XLVII (I935), 285-93.
9. FRANK,LAWRENCEK. "The Management of Tensions," American Journal
of Sociology, XXXIII (I928), 705-36.
Io. . "Time Perspective." (To be publishedin the Journalof Social
Philosophy,IV [I939].)
ii. HOPPE, FERDINAND. "Erfolg und Misserfolg" ["Success and Failure"],
Psychologische
Forschung,
XIV (I1930), I-62.
I2. JACK,LOIS M. "An ExperimentalStudy of AscendantBehavior in Pre-
school Children," in LOIS M. JACK,ELIZABETH MOORE MANWELL, IDA
GAARDERMENGERT, and OTHERS, Behavior of the Preschool Child. ("Uni-
versity of Iowa Studies in ChildWelfare," Vol. IX, No. 3 [I1934]). Pp. I7I.
I3. JONES, H. E.; CONRAD, H. S.; and BLANCHARD,M. B. "Environmental
Handicap in Mental Test Performance," Universityof California Publica-
tionsin Psychology,V, No. 3 (I932), 63-99.
I4. JUCKNAT, MARGARETE."Leistung, Anspruchsniveau und Selbstbewustsein"
["Achievement, Level of Aspiration and Self-Esteem"], Psychologische
Forschung,XXII (I937), 89-I 79.
I5. KEISTER, MARY ELIZABETH. The Behavior of Young Children in Failure:
An Experimental
Attempt
To Discoverand To ModifyUndesirable
Responses
of Preschool Children to Failure. ("University of Iowa Studies," Vol. XIV
[I937].)

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions
896 THE AMERICAN JOURNALOF SOCIOLOGY

i6. LEWIN,KURT. The ConceptualRepresentation and Measurement

ofPsycho-
logicalForces. ("Contributionsto PsychologicalTheory," Vol. I, No. 4
[I1938]). Durham,N.C.: Duke University Press.
I7. . "Bringingup JewishChildrenin the United States," Menorah
Journal,I939. (In press.)
i8. . Principlesof TopologicalPsychology.New York: McGraw-Hill,
I936. Pp. Vii+23I.
I9. . "The Conflict
betweenAristotelian
and GalileianModes ofThought
in ContemporaryPsychology,"Journalof GeneralPsychology, V (I193I),
I4I-77.
20. . "Psycho-sociological Problemsof a MinorityGroup," Character
III (I935), I75-87.
andPersonality,
2I. . Die Verwandtschaftsbegriffe in Biologieund Physik. Berlin: Ge-
brulder
Borntraeger, I1920. Pp. 34.
22. MURPHY, Lois BARCLAY. Social Behavior and Child Personality:An
ExploratoryStudy of Some Roots of Sympathy.New York: Columbia
University Press, I937. Pp. viii+344.
23. REUTER, E. B. "The Sociology of Adolescence," AmericanJournal of
Sociology,
XLIII (I937), 4I4-27.
24. SHIMBERG, MYRAE. "An Investigationinto the Validityof Normswith
Special Referenceto Urban and Rural Groups," Archivesof Psychology,
XVI, No. I04 (I928-29), 5-84.
25. WELLMAN,BETH L. "Mental Growth from Preschool to College," Journal
ofExperimental Education,VI (I937), I27-38.
26. . "The Effectof PreschoolAttendanceupon the I.Q.,"Journalof
ExperimentalEducation,I (I93 2-33), 48-69.
27. WIEHE, F. "Die Grenzendes Ichs," reportedfromthe manuscriptin
Kurt Lewin,A DynamicTheoryof Personality,pp. 26I-64. New York:
McGraw-Hill,I935.
IOWA CHILD WELFARE RESEARCH STATION

This content downloaded from 27.48.225.66 on Thu, 6 Feb 2014 03:42:53 AM

All use subject to JSTOR Terms and Conditions