Você está na página 1de 17

Social Science

Author(s): Peter Winch


Source: The British Journal of Sociology, Vol. 7, No. 1 (Mar., 1956), pp. 18-33
Published by: Wiley on behalf of The London School of Economics and Political Science
Stable URL: http://www.jstor.org/stable/588128 .
Accessed: 29/08/2014 16:21

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.

Wiley and The London School of Economics and Political Science are collaborating with JSTOR to digitize,
preserve and extend access to The British Journal of Sociology.

http://www.jstor.org

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
Social Science
PETER WINCH

I. UNDERSTANDING ANDEXPLAINING IN THESOCIAL SCIENCES


B ERTAIN KINDS of explanationof humanbehaviourmade in non-
{ scientific contexts involve the use by the explainer of concepts
which are also possessedby the personswhose behaviouris being
explained. For instance,I say of someone: " His reason for cancelling
next week's lecture is that he intends going to Londonfor a conference."
Theintelligibilityof this explanationis conditionalon the fact that the concepts
usedin it-canceUing,nextweek,givixglectures,goingto Londoare in general
understoodby the personwho is cancellinghis lectures; thoughhe need not
of course necessarilyhave consciouslyemployed them on this particular
occasion. Explanationsof the behaviourof naturalobjectsobviouslycannot
sharethis characteristic, sinceit is onlyof humanbeingsthat we canintelligibly
say that they are in possessionof concepts. FurtherJthere are many kinds
of explanationof human behaviourwhich do not share this characteristic:
an examplewould be a biologicalprinciplesuch as " human beings eat in
orderto maintaintheir metabolism". Peopleeat even if they do not under-
stand what " metabolism"means,and this in no way affectsthe validity of
the above explanation.
Now explanationsin the more developedsocial sciences are obviously
not on the same level as are common-senseexplanationssuch as my first
example. The social occurrencesexplainedwill not have been intendedby
the membersof the society whose actions led up to them, neitherwill the
conceptsused in the explanationsbe immediatelyintelligibleto those agents.
Nevertheless,they are not for that reasonclosely analogousto those natural
scientificexamplesto which I havvealreadyreferred. For they will inrrolve
a referenceto socialactivities,andit is partof whatis meantby the expression,
{ social activity", that the conceptsused to describesuch an activity must
also be possessedby those who performit since to say of someonethat he is
performingsuch an activity entails saying of him that he is in possessionof
such concepts. This is not to confusethe performanceof an activity with
talking and thinking about its performance,but to draw attention to two
importantfacts.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
PETER Wl N CH I9
(I) A personcan be said to be perfortninga specificsocial activity on
a given occasiononly if he belongsto a society in which the performance
of at least the generalkind of activity, e.g. playinggames,religiousworship,
the use of money, is an establishedinstitution,and if his behaviourshows
that he understandswhat is involvedin doingthat kind of thing. To show
that one understandswhat is involvedin game-playingis to show that one
possessesthe conceptgame. In one society a set of actions performedby
a group of people might constitute a game of hide-and-seek; in another
societya superficiallyidenticalset of actionsmightconstitutea religiousrite;
and in a society whereit did not fall underany conceptrecogiizedin that
society it would be completelyaimlessbehaviour. (2) The use of the word
" concept" in this connectionsuggests,what is true, that linguisticbehanour
entersinto the morerefinedformsof socialactinty. You cannotplay bridge
without understandingand being able to use werds such as " revoke" and
" trump"; you cannot pay bills urithoutunderstandingthe use of arith-
meticalsyrnbols; you cannotmakefuturearrangements andreminiscewithout
understandingthe use of the future and past tenses of verbs; you cannot
be a Christianwithoutunderstandirlg the use of suchwordsas " God", " sin ",
" salvation" and so one might go on. This is what I mean when I say
that bills are paid in terms of arithmeticalconcepts,that religiousworship
is carriedon in terms of the concept of God, and so on.
Liquiditypreference is a technicalconceptof economics; it is not a con-
cept generallyused,for exanqple,by businessmenin theirconductof economic
affairs. The economistuses it to explain how businessmen's behaviourin
relation to money brings about changesin the value of money which they
had not envisagedin behavingin that way. But the concept of liquidity
preferenceis logicallytied to conceptswhichare understoodby businessmen
in that its use by the economistpresupposeshis understandingof what it
is to conducta business,and in that this involvesunderstanding suchbusiness
concepts as money, profits,cost and risk.
Again, a psychoanalystmay explain a patient's neuroticbehaviourin
termsof factorsunknownto the patientand in terrnsof conceptswhichwould
not be intelligibleto him. The explanationmay referto eventsin the patient's
early childhood; the descriptionof those events will presupposean under-
standingof the conceptsin tetms of whichfamilylife, for example,is carried
on in our society and in terms of which the patient as an infant viewed,
howeverrudimentarily,the membersof his family. The conceptof a father
is differentin oursocietyfromthe corresponding conceptamongstthe Trobriand
Islanders,and this differencewould have to be understoodby anyone who
unshedto give an account of the differencesin the respectiveaetiologiesof
the neuroseshere and there.
The difficultiesinvolved in rejectingthis view of the social scientist's
relationto his subject-matterare well illustratedby Pareto'stheory in The
Mind and Socicty. Accordingto Pareto, humanbehaviourcan be classified
into (I) LogicalConduct,which is conductdirectedtowardsthe achievement

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
SCIENCE
SOCIAL
has
20
use of means the efficacyof which
making the methods
of a purpose,and
definite reasoning" (i.e.roughly, all behaviour
established by " logico-experimental Conduct,comprising
been and (2) Non-Logical specifications of
science);
natural
of not measure up to the
in one way or another
does
recommendations for the study of
svhich " and
Conduct. Pareto'smethodological
Logical between what he calls " residues
distinction to occur
life are based on a are forms of activity which can be seen times.
social
". Residues diverseplaces and
"derivations
in human societies at the most constitutea sort of theo-
constantly
fairly
on the other hand are unstable; they to explain why the
actions
Denvations for residues, purporting
invalidrationalization
window-dressing
retical
performed, but in fact being a mainly whichare not really
are
inquestion for activities
to provide a pseudo-justificationthe resultof certainfundamental
concocted but whichare
for any reasonat all are largely unconscious.
performed 1
"sentiments" of which men
are relatively unimportantand can
derivations residues.
Pareto maintainsthat scientistwho wishesto understandthe than not
be ignoredby the social
safely concern
likely
with derivationsis morebehaviour,since
that serious social
saysindeed
He understandingof as some-
stand
to in the way of any scientific is that they representthemselvesattention
features
of their distinguishingand will tend to divert the sociologist's
one He
which they are not, whichhe is studying. he
thing causes of the behaviour thing
the real natureand one importantexample of the kind of doctrines
from
religious practices as and theological
cites
here ritual acts constitutethe residues, for their per-
in mind;
has provide pseudo-justifications whichare
thecorrespondingderivations of exactly what concepts
are the residues to
formance. But in terms
and thought of ? The concepts applied to the
beingstudied to be describedthey are performedshould be barredPareto
them in the society in which derivations.
the validity of the In Sections I2C) and
since these presuppose atlew.
sociologist a sort of Hobbesian
oftentalks as if he holds of The Mind and SocieGy he tries to take ourer
expres-
those immediatelyfollowingfor the descriptionof socialevents,using ";
movement
theterminologyof mechanics equilibrium ", " tie 2'," realandvirtual carry
sionssuchas " force", " not necessarily
that these termsdo not give any precise
he warns occasionally he does
andalthough as they do in ulechanics, apparentthat
quitethe same sensein whichthey aresupposedto differ. It is descriptionof
accountof the ways he is hankeringafter a purely physical he does not even
throughoutthe book he is discussingactualcases,however,are, for example,
socialactions. When The actionsdescribed
beginto carry out this programme. for one's country". But it is
or " self-sacrifice in question
" sacrificesto the gods", these modes of clescriptionof the actions of whichthey
sufficientlyobviousthatfrom the systemsof conceptsin terms
get their sense entirely in Pareto's thought.
It must
at least represents one important strand
sentiments in social
life is more
1 The above account account of the functioning of to extract from The Alind and
Pareto's be possible sentiments
be said, however, thatand ambiguous, and it wouldSection I696, which suggests that
than usually obscure theory. Cf., in particular,
Societya quite different out of residues.
are logical constructions

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
P E T E R W I N CH
2I
are performed,i.e. fromwhat Paretocalls their " deristations". Clearly,too,
no accountof actions,as opposedto merephysicalmovements,couldbe given
exceptin termssuchas these; and no attemptto describeactionsin physical,
or even biological,terms could result in anything remotely resemblinga
descriptionof a human society.
Curiouslyenough,the defects in Pareto'sconceptualapparatusseem to
result pretty directlyfrom his failureto see all the implicationsof the fact
that scienceis itself a formof socialactivity. Despitethe heavy ironywhich
he is constantlydirectingagainstworshippersof " the GoddessScience", he
treats the criteriaof scientificthinking as if they were somethingabsolute
and in a completelyspecialpositionvis-d-visthe conceptsused in otherforms
of social activity. But the " logico-experimental"criteriaused in scientific
thinkingdo not stand on their own feet; they cannotbe understoodexcept
as involved in science as a form of human activity. In fact the relation
betweenthe conceptsand theoriesof scienceon the one handand the actierities
of scientisttheir experimentation,etc. on the otherhand is in this respect
exactly parallelto that betweenPareto'sexamplesof derivationsand residues.
Scientificconceptsareevolvedby scientistsin the courseof theirinvestigations
and can only be understoodin the context of those investigations. The
investigationsof scientistson the otherhandcan only be understoodin terms
of the conceptsof science. Thoseconceptsare, as it were, tailor-madeto fit
into the pattern of " logico-experimental reasoning". It is logicallyabsurd
to expect the conceptsinvolrredin activities of quite a differenttype to fit
into that pattern. TheseconceptsiIl their turn can only be graspedin terms
cf the activitiesin the context of whichthey are applied,and those activities
similarlycan only be understoodin termsof the conceptsappropriateto them.
There is a related confusionin the whole idea of " logical action" as
used by Pareto. It is an idea which really makes sense only within the
frameworkof logico-experimental ideas. That is to say, within this frame-
work we can speak of one action being logical,anotherillogical it is symp-
tomatic that Pareto observes no clear and consistent distinction between
" non-logical" and " illogical", althoughthis distinctionis vital for his whole
thesis. For him any action wllichis not logicalis non-logical,and this term
is appliedto a very heterogeneouscollectionof activities. It is a confusion
to say, as Pareto does, that scientificactivity as suchis a speciesof logical
conduct. It is just as non-logicalas is the practiceof religion,thoughneither
is illogical; and illogical actions can occur within the frameworkof both
scienceand religion. To graspthis point is to go a long way towardsunder-
standingthe relationbetween conceptsand behaviourin social life.

II. WAYSOFDOINGTHINGS, HABITSANDRULES


A series of articles by ProfessorMichaelOakeshottin The Cambridge
Journal1doesmuchto bringout the connectionbetweenthe idea of reasonable
1 See in particular " The Tower of Babel ", Nov. I948, and " Rational Conduct ", OCt. I950.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
22 SOCIAL SCIENCE
conductand modesof sociallife. But Oakeshottobscuresthe nature
latter by drawingan oversharpdistinctionbetweenconductwhich is aof the
self-
consciousapplicationof verballyformulatedlules and conductwhichhe calls
tt habitual"* He argues,rightly, that there are fundamentalconfusionsin
what he calls the " rationalistic>'conceptionof rationality,according to which
standardsof reasonablebehaviourexist absolutelyand are broughtin, as it
were, from the outside to regulateour conduct. But in his reaction
this brandof rationalismhe stressestoo mllch the notion of habit, and from
both
misconceivesand underestimatesthe importancefor this subject of rxle-
following in human affairs. To say that somebody'sbehaviouris
by rulesis not necessanly,as Oakeshottoften seems to imply, to say governed
that it
exempliSesa self-consciousapplicationof explicitlyformulatedpciples. A
man'sobservanceof rulesmay show itself itl his behaviourwithout
for its recognitionby othersyany verbal acknowledgementor requinng,
by him formulation
By { habitual behaviour" Oakeshottseems to mean behaviour
has not been taught-by preceptbut which has been acqulredin the which
of practice. Exampleshe gives are cooking,speakinga language, course
morally,taking part in scientificenquiry. He is quite nght in saying behaving
sllch actinties as these are based very largelyon the unreflective that
of skills in practice; but he does not seem to have noticed that acquisition
whichis habitualin this sense is neverthelessoften correctlydescnbed conduct
as an
instanceof rule-followmg It is not on the same level as the
conditioned
reflex,nor even as what I will call " blindhabit ". An exampleof the
wouldbe the llabit someonemight acquireof alwaysputtingon his leftlatter
before his right. The distinction between blind habit and shoe
rule-governed
habitis not a sharpone. The habit of alwaysputtingon the left shoe
the right may start as a blind habit and developby imperceptible before
stagesinto
the observanceof a mle, or uice versa. Recognitionof such a change
in the
characterof a person'sbehaviourdoes not depend on any explicit avowal
by him that he is malsingit a rule to act in this way; it ls sufficient
if, for
instanceJhe becomesannoyedwith himselfor performsa propitiatoryrite if
onsome occasionhe shouldput on his shoesthe otherway round. In
hemust showthat he recopzes harringdonesomethlngwro?sg.In this short,
examplethe distinctionis no doubtnot very important; but it becomes trinal
very
muchso in connectionmth such fundamentalhumanactivitiesas
mentionsas examplesof what he calls " habitualbehanour". Oakeshott
A dog can be taught tricks (e.g. to balancea lump of sugaron its
andwe can use words like " correct" and (' incoITect" in our nose),
ofhis performanceof the trick. This, however,is an assessments
ofspeaking,unobjectionablein its context but misleadmg anthropomorphic way
if appealedto in
supportof any strong analogybetweenhumanand animalbehanour.
veryidea of performinga tnck, and the standardsof correctbehasriour The
goalong with that, cannotbe fully elucidatedsimplyby a which
descnption,how-
everdetailed)of what dogs can be observedto do, but requirea
referenceto

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
PETER WINCH 23
charactenstichuman forms of activity, to norms taken from a context of
humansocial life and appliedanalogicallyto animallife. The dog, we can
say, acquiresa habit of doing certainthings on certainoccasions balancing
the sugar when a word of cornmandis uttered; the behaviourist'scausal
terminologyof " stimulusand response" descnbes the situation fairly ade-
quately. If on someoccasionthe dog does not respondin the way we regard
as appropnate,we may look for an explanatorycause perhapsa sttmulus
operatingin the contrarydirection. We cannot say that the dog has done
anythingwrong,unless we look on his behaviourfrom the point of view of
our own purposesand critena in teaching him the trick.
Such expressionsas this, however,are indispensablefor the description
of a rery large area of human behaviour; that, namely, which exemplifies
rule-followingrather than blirldhabit. The dog's behaviouris stereotyped
in that his habit consistsin his always doing the same thing in responseto
the same stimulus. Now althoughto say of a man that he is followinga
rule in what he does also entails sapng of him that he always acts in the
same way in a particularsort of situation, neverthelessthe word " same"
cames here a sense which differsimportantlyfrom that which it calTiedin
its previoususe. In particular,it does not mean that the man's behaviour
is stereotypedas is the dog's. For instance,somebodysays that he makes
it a rule alwaysto be in good time for his appointments. This entails a lrery
definiteregularityin what he does on that kind of occasion. But the speci-
ficationof what constitutesthesamekitd of occasionis very muchmoreelastic
than it is in the case wherewe are dealingwith animalhabits. He will leave
himselfplenty of time no luatter whomhe is meetingand no matter whereJ
and there may be borderlinecases in whtch he has to decide whetherthe
situationhe is in constituteshavingan appoirltmentto keep or not. It must
be possibleto ask whetherhe is behavingconsistently, and if, on someoccasion)
he fails to try to be in goodtime for an appointmenthe cannotrebuta charge
of inconsistencysimply by saying that it is after all a differentpersonhe is
meeting on this occasionand in a differentplace}from what has been the
case on previous occasions. One cannot charge a dog with inconsistency
becausea dog's behaviourdoes not exhibit the featuresin connectionwith
which alone such a charge=makes sense. It must be possiblein principleto
makesomeonerealxze that he has behavedinconsistentlyif it is to makesense
to chargehim with inconsistency,and this is possibleonly with someonewho
is himselfcapableof applyingcriteriafor distinguishingsimilarfromdifferent
situationsin the regulationof his conduct. Animalscannot do this though
we may apply our own criteriaof samenessin describingwhat they do.
The applicationof such criteriain the regulationof behaviouroccurs
characteristicallrwherethat behaviourexhibitsregularityof a sterycomplex
sort, extendingolrersituationswhich in many respectsdiffer markedlythe
one fromthe other. A bliIldhabit on the otherhandis a tendencyto perform
a fairly stereotypedaction on a fairlystereotypedsort of occasiontwherethe
charactensticsof the actionor situationcan be definitelyspeciSedin concrete

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
4 SOCIAL SCIENCE
terms. A personwhose behaviouris rule-governedwill, when laced with a
situation markedlydifferentfrom any he has previouslyhad to deal with)
have to decidehow he is to interpret his rulein this new situation; but some-
body whois merelyactingIromblindhabitwill be facedwith no suchproblemn
He may be at a loss what to do, but his problernwill not take the same form
as that of the personwho is followinga rule; in particular,no questionof
consistencycan arisehere. It is noteworthythat Oakeshott'sapproachleads
him to say that dilemmasof the form " What ought I to do here? ", in
situationssuch as I am now envisaging,are likely to ariseonly for somebody
who is self-consciouslytrying to follow explicitly formulatedrulesJnot for
somebody whose behaviourexhibits the kind of implicit rule-followingof
which I have been speaking,that is, whose conductis what Oakeshottcalls
" habitual". Now it may be true that the necessityfor suchheart-searchings
is likely to be more frequent,and perhapsmore pressing,for a personwho
is trying to follow an explicit rule without a foundationof esTeryday experi-
ence in its application; but still questionsof interpretationand consistency
are bound to arise for anybodywhose conduct is rule-goverrsed in the way
I have described,i.e. for all of us, in dealingwith a situationforeignto his
previousexperience. In a rapidlychangirlgsocialenvironmentsuch problems
will arise frequently,not just as a result of the breakdownof traditional
customarymodesof behaviour,but as a resultof the novelty of the situations
in which that behaviourhas to be carriedon. To make this distinctionis
not to deny that a rapidlychangingenvironmentmay lead to a breakdown
in traditionsof behaviour. It is possibleto arguethat in some spheresthe
positionis the oppositeof what Oakeshott'srriewwould lead one to expect.
A rule is open and makes possiblethe reflectiveapplicationof past experi-
eilce to new kinds of situation; a blind habit is confinedto one particular
nar4ow type of situationand is likely to engendernothingbut bewildersent
wherethe conditionsin which a personhas to act are very novel. FurtherJ
it habbeen pointedout 1 that in the UnitedStates the existenceof a verbally
formulatedConstitutionsometimesmakesforgreaterflexibilityin the adminis-
tration cvfthe law than is possiblein a countrywhere there is no abstract
principleto be appealedto above that embodiedin the decisionsof precedent
cases. WThere a seriesof precedentsleads to an inabilityto cope adequately
with a novel social situation,the SuprerneCourt,by appealingback to the
originalterms of the Constitution,has sometimesbeen able to break the
deadlock.
Not all the behaviourwhichis studiedin the socialsciencescan be brought
directly under the concept of rul-following. It is said, for instance, tnat
Japanesemen are attractedby the backs of woInen'snecks ratherthan by
their faces. Such facts as these abound in works of sociolegy and social
anthropologSr, and they can be establishedby simple observationsubject to
controlsvery similarto those foundin the naturalsciences. Again,a political
1 By E. H. Levi in An Introductionto Legal Reasoning(publishedbU,>
the Law School of the
University o f Chicago).

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
PETER WINCH 25
scientistmight establishthat the voting behaviourof a certainsection of the
electoratewas completelyrandomand subjectto no rule at all. That would
be establishedin the same sort of way and might perhapsbe formulated
statistically. In itself it wouldnot requireany understandingby the investi-
gator of rulesfollowedby the objectsof his investigation. To such activities
of the social scientist as these the considerationsput forwardin this paper
have no direct relevance. My argumentis, however,that the conceptionof
a humansociety in generalcannot be graspedexcept in termsof the concept
of rule-followingand that rule-governedbehaviourconstitutes one of the
most interestingfields for investigationby the social sciences. Moreover,
even in cases such as those just mentionedone is never very far away from
rule-governed behaviour. Even if a man'sactualvotingbehaviouris random,
to graspwhat he is doingas an instanceof voting is impossibleunlessone has
someunderstanding of the organizationof politicalaffairsin his society. Even
if he casts his vote in a randomway, his behaviouris not completely random
and arbitraryif what he is doing is voting. He is participatingin a form of
social life and his behaviouris guided, y)erhapsunreflectively,by considera-
tions of what is and what is not appropriatein that form of activity. The
politicalscientistcan be assumedto be familiarwith those considerationsand
most of the time he has no need to bringthem into the open. But the results
of his scientiScand statisticalmethodsof investigationare relevantand im-
portant only in virtue of his and his readers'understandingof the way of
life in the context of which they are applied. Otherwiseone has merely a
rag-bagof facts without significanceor cohesion.

III. SCIENTIFIC
THEORIZING
ANDUNDERSTANDING
A WAY OF LIFE
In the light of the above considerationsI want now to ask in what sense
the behaviourof men in society may be said to exhibit regularitiessuch as
may be studiedby the socialscientistand made the basisof scientifictheories.
Mill'sview,in BookVI of the Systemof Logic,is that theseregularitiesareon
preciselythe samelogicalfootingas those whichmay be observedin the realm
of nature,the differencebeing one of complexityonly, and this view is still
tacitly assumedin many contemporarypronouncementson the nature and
methodsof social science. I wish to argue,on the contrary,that the whole
sense of the word " regularity" is differentin this context, and that con-
sequentlythe investigationof society is on quite a differentlogical footing
from the investigationof nature.
To say that there are regularitiesto be observedin a certain realm of
events is to say that in specifiablecircumstancesthe same specifiableevent
always,or nearlyalways,happens. Obviously,this presupposesthat we have
criteriaof sameness; we must have some methodof decidingwhen an occur-
rence in one situation counts as the same qualitativelyas an occurrencein
another situation. So the appropriatemethod of determiningthe type of
regularitywith which we have to deal in a given field will be to investigate

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
26 SOCIAL SCIENCE
the kind of criteriaaccordingto whichwe decidein that field
whenthe same
thing has happenedin morethan one situation. Whenwe are
a physical science these criteriamust be looked for in the dealingwith
context
methods and proceduresof investigatorsworkingin that science, of the
for
applicationof those criteriais unintelligiblein abstractionfrom those the
cedures. Consideran investigationlike Galileo'sinto the mechanicsof pro-
bodies,involvingexperimentswith balls on inclinedplanes,and falling
the investigatorperformstwo experimentswith two differently suppose that
He will say that the conditionsof these two experimentsare inclined planes.
the planesare differentlyinclined; he has of coursemethodsdifferent in that
for determining
when two planes are diflerentlyinclined, and what " differently
means cannot be understoodby someonewho does not understand snclined"
at
somethingof those methods. The fact that the two planes are made least
the same sort of wood is not taken into accounthere; neitheris from
of the wood,nor the nameof the carpenterwho constructedthe the colour
these factorsare not taken into considerationcan be understood planes. Why
basis of knowingsomethingabout the natureof the investigation only on the
physicistis corlducting. For somebodyelse, conductinga different which the
sort of
investigation,e.g. into the durabilityof differenttypes of wood,the
of a given carpenter'sxvork,the aestheticeffectsof differently adequacy
tures,these factors wollld of coursebe relevant. That is to coloured struc-
say, he would
decidewhethertwo situationswerethe " same" or not accordingto
criteriafromthose used by the physicistconductingthe kind of different
nowunder discussion. investigation
One's criteriafor (lecidingwhen two situationsare to be
beingof the same kind are relativetc the kind of activity in theregardedas
whichthe questionarises. I wish to considerthis relationmorecontext of
closely, in
connection with the formof socialactilritywe call stientificinvestigation. A
scientist'smethodsof investigationexemplifyconductcarrsedout
torules. It must always make sense to ask whetherhe has according
carriedout his
procedures correctlyor incorrectly,and this is only possiblewherethereexist
standards of correctnesswhichcan be appliedto them. Standardsof
nesspresupposethe possibilityof one person'smistake being correct-
otherpersons,a possibilityinvolvinginteractionbetweenpeople in corrected by
context. It is in fact a constitutivepart of the notion of scientific a social
tionXas of any other socialactivity, that its practiceby any one investiga-
bein principlesubjectto check and correctionby other personshall
persons;
thispossibilitydoesnot exist, thereis no possibilityof doing for where
somethingwrong,
thereforeno possibilityof doing anythingnght either.l
Tn charactenzingthe relation between an individualscientist
and his
{ellow-participantsin the traditzonof activity which makeshis actionson a
givenoccasionan exampleof scientificinquiry,one cannotbe
aset of preceptsor rules ol be}waviour content with
the followingof which constitutes
t Cf. Wittgenstein, Philosophical
a Private Language? ", Proceedings IetUestt;ations,
be I, 2f6 ff.; also Rush Rhees: " Can there
of the Aristotelian Society, SupplementaryVolume
XXVIII.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
number,
orhecopies
theserieso, I, 2, ^, 4, 5} ................................... hkethis:I, O, 3, 2, 5, 4,

PETER WINCH 27
engagingin scientificinquiry. For what is in questionis: what is it precisely
about the behanour of scientistswhich makes us say of them that they are
following these precepts, assuming that the latter could be satisfactorily
formulated? Oakeshott,in his article on " RationalConduct", notes that
such an activity goes beyond anythingthat could be formulatedin a set of
precepts,at least in respectof the fact that those pr2ceptshave to be applied
in practiceby scientists. One could add that any attempt to formulatethe
way the rules are appliedin terms of anotherhigher-orderset of rules wollld
start one off on the slopeof an infiniteregresspreciselyparallelto that pointed
out by Lewis Carrollin " What the Tortoisesaid to Achilles". The point
here is that the notion of followinga rule is presupposedby the specification
of any particularset of rules; such a specificationsays nothingat all about
the kind of behaviourin the context of which alone the idea of following
a rule makes sense. This notion could not be understoodby someonewho
was not alreadyfamiliarwith that sort of behaviour.
In Wittgenstein'sPhilosophicalInvestigationswe have a characterization
of the sort required.l I will here quote from Section I43 in extenso.
Let us . . . examine the followingkind of language-game: When A gives an
order B has to write down series of signs accordingto a certain formation-rule.
The first of these seriesis meant to be that of the naturalnumbersin decimal
notation.-How does he get to understandthis notation? First of all series of
numberswill be tten dolvn for him and he will be requiredto copy them....
And here already there is a normal,and an abnollllal hearer'sreaction. At first
perhapswe guide his hand in writingout the serieso to g; but then the possibility
Ofgettinghim to understandwill dependon his going on to te it down independ-
ently.-And here we can imagiIle,e.g. that he does copy the figuresindependently,
but not in the right order: he writessometimesone sometimesanotherat random.
And then communicationstops at thatpoint.-OI agaln he makes " mistakes " in
the order.-The differencebetween this and the first case will of coursebe one of
frequency.- Or he makes a systematic mistake; for examplehe copies every other
Here we shall almost be tempted to say he has understoodwrong.
Notice, however,that thereis no sharpdistinction between a randomriistake
and a systematic one. That is, betweenwhat you are inclined to call " random"
and what " systematic".
I said that Wittgensteincharacterizedsome generalfeaturesof conduct
which is subjectto rules. In fact, however,as he himselfpoints out in the
next section,what he has done here is to describewhat mighthappenin the
situation envlsagedbut in fact does not commonlyhappen, the purposeof
this procedurebeing as follows:
I wanted to put that picturebeforehim (sc. the personto whom Wittgenstein
has imaginedhimselfto be speakiIlg),and his acceptanceof the pictureconsistsin
his now beinginclinedto regarda given case differently: that is, to compareit with
thisrather than thatset of pictures. I have changedhis way of lookingat things.
One mightexpressthe way of lookingat thingsinducedby Wittgenstein's
remarksby saying: one is now inclinedto noticeand even to feel slightly
1 Cf. especially I, I43 ., and I, I85 ff.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
28 SOCIAL SCIENCE
surprisedabout a fact that is so obviousthat one would normally
it altogether; the fact,namely,that whensomebodyis taught overlook
to do something
he will normallygo on independentlyin the activity whichhe has
in the same way as the teacherhimselfwouldgo on. One sees been taught
unlessthe behaviourof most peopleagreedin this sort of way the furtherthat
rule-guidedbehaviourwould be completely inapplicable. The " notion of
reaction,in the sense of the reactionwhich most peoplewill in factnormal"
is elevatedinto a standardof correctbehaviourfromthe point exhibit,
of view of the
particularrule in question,and any behaviourwhich deviates from it
be said to constitute a mistake. will
I am not saying that peopleobservethat most of them act in
way in certainsituationsand then agreeto take that commonway of the same
as theirnormfor decidingwhat actionsare rightand what behaving
wrong. This may
of coursehappenin individualcases: for instance,supposethat
in
days of motortransportno particularrulewas observedconcerning the early
side of the road to drive along; supposefurtherthat with the the cosTect
trafficit becamenecessaryto institutesomesuch rule; in deciding increasein
to institute it might be observedthat peoplealreadyhad a what rule
natural
to drive on the left-handside ratherthan the right; and then tendency
it might be
decidedthat that mode of behaviourshould in future count as the
But the " agreement" to which I am refetTinghere is more norme
that, and a very large measureof it would alreadybe primitive than
presupposedin the
situationjust described. Forno sensecan be givento suchnotionsas
thatso and so is the case unless men alreadyhave a mode of noticing
comrnunication
a languageof some sort in the use of which rules are already
served. For one cannotnoticeanythingwithoutbeingableto identify being ob-
definite
characteristics;which means that one must have some conceptof definite
characteristics;and that involves the rule-governeduse of some symbol to
referto those characteristics.l The developmentof rulesrests on a
primitive,
urlselfconsciousconsensusof reactions. One cannot say that this consensus
ofreactionscomes first and is followedby the observanceof
rules; for the
relationbetweenthe notionof a ruleand the notionof a consensusof reactions
ofthis tvpe is a logicalone, not a relationof cause and effect.
The kind of
conDensus whichis herein questionis describableonly in termsof the notion
ofa rule; and the notion of a rule is indescrlbableapart from
Ofthat sert of consensus. the concept
Further)I am not saying that it must always be right to do what
majority of peopledo, and that deviationmust,froma sortof logical the
bewrong. One mllst reInemberthat a deviationfrom necessity,
commonlyaccepted
standards, howeverradical it may be, is never a completerejectionof all
cornmonly acceptedstandards. Howeverxriolenta disagreementmay be, it
cannot but occuragainsta vast bacligroundof commonpresumptions.
outsuch a background With-
the very notionof a disagreementwould be impossible.
Aparticularscierltist,for example, may disagreeprofoundly
with all his
1 Cf. Rush Rhees, " Can there be a Private l,anguage 2 ';
passim.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
PETER WINCH 29

colleaguesconcerningsomeimportantpoint of doctrineor method,as Einstein


disagreedwith his fellowphysicistsover quantummechanics. But if Einstein
had not shareda vast area of commonground,in the shape of a scientific
tradition comprisingboth theory and techniquesof investigation,he could
not even have begun to formulate not even to himself- those points over
which he disagreedwith his colleagues. That is tantamountto saying that
it would make no sense in such circumstancesto speak of a disagreement
between him and them. A disagreementmust be abo?wt something, and
where it is in principleimpossibleto formulatethe points of disagreement,
the idea of there beinga disagreementis logically absurd.
In the language-game describedin SectionI43 of the Philosophical
Investi-
gationsno questionarisesof what followingthe rule involvesin circumstances
differentfromthose in connectionwith whichit was taught. One might say
that the behaviourof somebodywho has satisfactonly learnt this rule is
stereotyped like the dog's performanceof a trick. Of courseit differsfrom
the latter in that a personfollowinga rule must himself understandwhen
the rule has been correctly,and when incorrectly,followed,if he is to be
said to have graspedthe rule rather than merely to have acquireda blind
habit; whereasthe dog can hardly be intelligiblysaid to understandthis,
althoughwe may say of the dog that he has acted correctlyor incorrectly.l
A parrotmight, nevertheless,be trainedto repeat the numeralsfrom o to 9
in theircorrectorderas wellas a man. But the ruleswhicharemostimportant
in humanlife are used to guide people'sactions in circumstancesrnarkedly
differentfromthosein connectionwith whichthey werefirstlearntor acquired;
and therein precisely lies much of their importance. To illustrate what
" differentcircumstances"means here I will again quote from Wittgenstein
this time Section I85 of the PhilosophicalInvestigations.
Now -judgedby the usualc:riteria-thepupilhas masteredthe seriesof natural
numbers. Next we teach him to write down other series of cardinalnumbersand
get him to the point of writing down series of the form,
o, n, 2n, 3n, etc.
at an orderof the forrn" + n "; so at the order" + T " he writesdown the series
of natural numbers. Let us supposewe have done exercisesand given him tests
up to TOOO.
Now we get the pupilto continuea senes (say + 2) beyondIOOO and he writes
IOOO, I004, I008, TOI2. We say to him: " Look, what you've done." He doesn't
understand. We say: " You were meant to add two: look how you began the
series." He answers: " Yes, isn't it right ? I thought that was how I was meant
to do it." Or supposehe pointed to the senes and said: " But I went on in the
sare.eway."-It would now be no use to say: " But can't you see . . . ? " and
repeat the old examplesand explanations. In such a case we might say, perhaps;
" It comes natural to this person to understandour order with our explanations
asweshouldunderstandthe order: ' Add 2 Up to IOOO, 4 up to 2000, 6 up to 3000
and so on.'"
In this example followinga rule involves performingactions different
from those onginally demonstratedin the teaching of the rule; and the
1 See Section II above.
C

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
3o SOCIAL SCIENCE
criterionfor decidingwhethersomebodyhas understoodthe rule is precisely
that he shouldbe able to go on in this way. What he has to do is in one
sensedifferentfromwhat he originallydid undersupervision,but nevertheless
counts,in relationto thatrgle, as going on in the sameway. Again,as in the
firstexample,the possibilityof the existenceof the ruledependson agreement
amongstthe majorityof peopleconcerningwhat particularactionsshall cognt
as the same. Neither criteriaof samenessnor, what is fundamentallythe
same thing, rules could exist if the behaviourof everybodywere eccentric
like the man's in Wittgenstein'sexample,and accordingto no fairly simple
pattern.
The aboveconsiderations relateto the notionof a ruleas such,not simply
to mathematicalrules or to any particularkind of rule. Of course,many
importanthumanactivitiesare characterizedby rules whichdiffermarkedly
from those governingthe pursuitof mathematics. For instance,in morality
politics, the law and aestheticsthere is much disagreementabout what, in
concreteterms, gottg or tn the same way demands}and very many of the
most importantdisputeswithin those fields can be regardedas attempts to
resolve such disagreementsand to find commonlyacceptableinterpretations
of whateverrule is in question. I will returnto this aspect of the matter
beforeconcluding; but at presentI wish to concentrateon the fact of agree-
ment over large areas rather than on the fact of disagreementover other
large areas, and to insist once again that the latter could not exist without
the former. Peoplecouldnot disagreeabout anythingat all unlessthey also
agreedabouta greatdeal: if aboutnothingelse, at least aboutthe application
of the rulesgoverningthe use of the languagein whichthey talk about their
disagreements.
This discussionof the natureof rule-governedbehaviouraroseout of an
attemptto understandthe natureof the criteriaaccordingto mrhich we judge,
in a natural science,when there are regularitiesin the phenomenawhich are
the subject-matterof that science. This was necessarybecausethe notion
of applyingcriteriamakessenseonly in the contextof rule-governed behaviour
and cannot be understoodwithout some understandingof what is involved
in the latter. I have tried to show that the detectionand study of physical
regularitiesrests on the existenceof regularitiesof a differentsort in human
behaviour. Now accordingto the Mill traditionthe idea of a social science
is the idea of a body of theory based on the detectionand formulational
regularitiesin humanbehaviourin society: that is, accordingto the present
analysis,regularitiesof the sort involved in rule-governedbeharriour.The
questionnow to be consideredthen is this: takinginto aceouIlt the peculiari-
ties of this kind of regularity,can its detectionand formulationrest on the
same kind of procedureas does the detection and formulationof physical
regularities?
The criteriaby which a phvsicist judges that a regularityexists in a
certain field of phenomenahave been evolved in the practiceof scientific
investigation. The individualscientist stands in relationboth to the phe-

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
PETER WINCH 3I

nomena which he is studying and also to his fellow scientists. Both these
relationsare essentialfor it to make sense to say of him that he is detecting
regularities. But they are relationsof a very differentsort. The phenomena
presentthemselvesto the scientistas an objectof study; his fellowscientists,
on the other hand, are related to him as fellowparticipantsin the activity
of study; without this relationthe word " study" could not intelligiblybe
appliedto the actionshe is performing. The conceptsused by the scientist
are formulatedandmodifiedaccordingto the way the phenomenaareobserved
to behave. But the role of the scientist'sfellow workersis differentfrom
this; he does not evolve his conceptson the basis of his observation of the
way they behave, but ratherhe is able to evolve them in virtue of the fact
that he is doing the same kind of thing as they are, namely investigating
in a certain way. The very concept of observationin the sense in which
it is relevantas a descriptionof part of the scientist'sactivities, is possible
only in virtue of this relationbetweenhim and his fellow workers,so that
this relation clearly cannot be that between observer and observed. To
maintainthat it is the same is to tie oneself up in a logical circle.
This relationis necessaryif the scientist is to understandthe concepts
employedin scientiScresearch. One of the chief tasks of.the socialscientist
who wishes to understanda particularform of human activity in a given
society will be, as I arguedearlier,to understandthe conceptsinvolved in
that activity. It would appearthen that he cannot acquiresuch an under-
standingby standingin the same relationto the participantsin that form of
activity as does the naturalscientistto the phenomenawhichhe is studying.
His relationto them must ratherbe that betweenthe naturalscientist and
his fellownaturalscientists; i.e. that of a fellowparticipant. It is of course
possibleto observeregularitiesin humanbehaviourand to expressthem in
the formof generalizations. But my point is that it is not possibleto under-
stand the principlesaccordingto which such behaviouris carriedon purely
by this method. For understandingthose principlesinvolves understandlng
the concepts involved in that behaviour,and these are masterednot by
observingand theorizing,but by actuallytakingpart in that activity together
uriththe other people involved in it.
This point is reinforcedby such common-senseconsiderationsas the
following. A historianof religionmust himself have some religiousfeeling
if he is to " get inside" the religiousmovementwhich he is studying; a
historianof art must have some aestheticsense and some understandingof
the kind of problemwhichconfrontsartists; a historyof philosophyis best
written by a philosopher; and so on.
The argumentof this paperalso providessome justificationfor the pro-
foundhistoricalscepticismand relativismof an Idealistphilosopherof history
such as Collingwood. Its practicalimplicationsbecome pressingwhen the
social scientificinvestigatoris dealingwith behaviourin a society which is
very remote from the ways of doing things with which membershipof his
own society has made him familiar. He cannot participatein that kind of

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
32 SOCIAL SCIENCE
behavioursince it is no longercarriedon; all he can do is to look at those
activitiesfor whichhe has historicalevidencefromthe outside. Thisaccounts
for the weightattachedto conceptslike " empathy" and " historicalimagina-
tion" by the Idealists. Howevermuch metaphysicalconfusiontheir use of
thoseconceptscarriedwith it, they did at least emphasizein a salutaryfashion
how differentare the criteriaby which we judge the value of an historical
account from those by which we judge the value of a scientifictheory.
In conclusionI wish to make some remarksabout the logicalbasis for
predictionsof social events and about the applicationof deterministicideas
to the realmof humansociety. To understandwhatis goingon in a particular
area of a given society one must understandthe rulesaccordingto whichthe
various activities being performedthere are carriedon. If one knows the
rule whichsomebodyis followingone can, in a largenumberof cases, predict
what he will do on a giverioccasion. For instance,if one knowsthat a man
is followingthe rule: " Start with o and add 2 till you reach IOOO", one
can predictthat, havingwrittendown" I04 ", he willnext writedown" I06 ".
Thusfar, the positionmightappearto be closelyanalogousto the use of one's
knowledgeof a naturallaw in orderto predicta futurephysicalevent-say
an eclipseof the sun. But sometimes and this is the sort of case whichis
importantin the developmentof a humansociety even if one knowswith
certainty the rule which a man is following,one canno!predict with any
confidencewhathe will do on a givenoccasion. Thisoccurswhenthe occasion
in questioninvolvescircumstancesmarkedlydifferentfromany to whichthe
rule has hitherto been applied,for here the questionmay very well arise:
what is involvedin followingthe rule in circumstanceslike these ? There
may then be deliberation,argument,the canvassingof rival interpretations,
followedperhapsby the adoptionof some agreedcompromiseinterpretation
or by the cominginto existenceof nval schools. Processeslike this can be
seen in the developmentof any historicaltradition: considerfor example
the developmentof the notion of empiricismin philosophy,of the notion of
realism in literature; considerthe relation between the music of Haydn,
Mozart,Beethovenand Brahms; or in politicsconsiderthe rival schoolsof
politicalthoughtwhich all claim, with some show of reason,to be based on
the Marxisttradition. The processis seen in a convenientlysimplifiedform
in legal procedures,the cantrassingof rival interpretationsof a rule of law
by opposingcounsel,followedby a judicialdecisionon what the rule is to
be taken to mean in circumstanceslike this.
The importantpoint to noticeis that in all cases wherea rzle is applied
in caseslike this somethingin the natureof a decisionis boundto take place,
even wherethe clear-cutdecisionprocedureof the law-courtsdoes not exist:
bound as a matter of logical necessity,since the conditionsof the situation
in questionincludethe propositionthat it has not yet been settled what the
ruleis to implyin these circumstances. A shrewdobserxrer may have a hunch
which way the decisionwill go, but the rule as thus far developeddoes not
entailone decisionratherthan another. Considerthe differencebetweenthe

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions
PETER WINCH 33
predictionsof a shrewdlawyeron the outcomeof a trial, or of a shrewdchan-
cellor of the exchequeron the outcomeof a particulareconomicpolicy, on
the one hand, and on the other hand the predictionsof an engineeron what
will happento the behaxiourof a particularpiece of machineryif a certain
modificationis made to it. Both the engineerand the chancelloror lawyer
may be eitherright or wrongof course; but the latter may be right or wrong
for reasonswhichit wouldbe nonsensicalto apply to the successor failureof
the engineer'spredictions.
A humansocietycannotbe adequatelydescribedwithoutuse beingmade
of the notion of a developingtradxtion.This notion is inapplicableto the
behaviourof objects such as are studied by the naturalsciences-either of
inanimateobjectsor of non-humananimals; herethe behaviourbeingstudied
does not exemplifythe applicationof rules. The behaviourof the natural
scientificinvestigatordoes of course exemplifythis, and the scientificlaws
which he formulatesdo dexelop in accordancewith the scientifictradition
of whichthey forma part. But the investigationof the scientist'sbehaviour
and of the developmentof a scientifictradition,belongsto the social, not
the natural, sciences.

Current Notes
Owing to productiondifficultiesthis numberof the British Jogrnal of
Sociologyis 80 pages insteadof the usual 96. We hope to makeup for these
missing I6 pages in our next issue, where an article by the late Professor
Hobhouse," Some of the SimplestPeople", will be published.
A series of six lectures are being held on " Psycho-Analysisand Con-
temporaryThought" at Friends House, Euston Road, London,N.W.I, in
connectionwith the centenaryof the birth of SigmundFreud, on May 6th,
I956. These lectures are being arrangedby the British Psych>Analytical
Society,63 New CavendishStreet, London,W.I,and will be taking place on
AprilI3th, I7th, 24th and 27th and MayISt and 8th. Detailsof the lecturea
may be obtainedfrom the above Society.
A new Journalox BehaviouralResearchRelevantto the PeacefulSolution
of InternationalConficts is in preparation. Sociologists,political scientists
and psychologistson the EditorialBoard include Robert Angell, ArthurI.
Gladstone,Daniel Katz, Herbert Kelman and Frank Pinner. Sociologists
and psychologistswho wish to offer articles and related material should
eommunicatewith the ManagingEditor, Robert Hefner, Departmentof
Psychology,Universityof Michigan,Ann Arbor,Mich., U.S.A.

This content downloaded from 146.83.195.38 on Fri, 29 Aug 2014 16:21:52 PM


All use subject to JSTOR Terms and Conditions

Você também pode gostar