SOCIAL

KNOWLEDGE

An Essoy on the
Notureond Limits
of SociolScience

PoulMottick,Jr.

Hutchinson
London Melbourne Sydney Aucklond Johonnesburg
HutchinsonEducation

An imprint of Century HutchinsonLtd

62-65 ChandosPlace,London WC2N 4NW

Tobleof Contents
Century HutchinsonPublishingGroup (Australia)pty Ltd
l6-22 Church Street,Hawthorn, Melbourne,Victoria 3122

Century HutchinsonGroup (NZ) Ltd

32-34 View Road, PO Box 40-086, Glenfield,Auckland l0

Century Hutchinson(SA) (Pty) Ltd

PO Box 337, Bergvlei 2012, SouthAfiica

First publishedin the United Statesof America

b y M . E . S har p e In
, c. 1986 Preface VII

First publishedin GrearBritain 1986 L Introductory 3

2. "Subjective"ond "Objective" 12
O M . E . S har p e In
, c. 1986
3. Understonding on Alien Society 30
Printed in the United Statesof America
4. Understonding ond Explonotion 50
5. Foundotionond Superstructure 71
British Library Cataloguing in Publication Data
6. Scienceond Society 100
Mattick, Paul
Nofes 120
social knowledge:an essayon the natureand limits of socialscience.
l. Socialsciences-Philosophy Works Cited 132
L Tirle
lndex 135
300' .I H6 l
About the Author 137
tsBN 0 09 1654602
 Prefoce

'fhis book is one of severalpublished in recent years dealing with

issuesin the philosophyof social scienceso basic as to call into
question the very possibility of such a form of knowledge. It
differs from most of thesebooks in that it examinessomeof these
issueswith an eye to their appearancein the contextof a particular
theory, Marx's critique of political economy.This reflectsboth
my interest in this theory for its own sake, and my feeling that it is
at least in part the absenceof discussionof the problems raisedby
actual social theory that gives much philosophy of social science
its peculiarly sterile flavor.
On the one hand, argumentsfor the impossibility of scientific
study ofone or another realm ofphenomena are in general dubi-
ous. The developmentof new methodsmay always open hitherto
closed areas to scientific analysis. On the other hand, proofs of
the mere possibility of a science (as exemplified by Nagel's de-
fense of the social sciencesin The Structure of Science)can be
nearly as uninteresting.Marx's Capital, meanwhile, containsthe
statementof a theory constructedon a sufficiently large scale to
be comparableto interestingly general theories in other domains.
It meets the main requirements for a scientific theory: on the
formal side, it proposesa set of principles for the explanation of
 vilr : ; o ( I n t KN( ) Wl I lx,l
observablercalitiesand the predictionof clcflnitetrcnds; it is
falsifiable,though it is in fact quite well confirmed. This exam-
ple, however, has only rarely been discussedin the literature of
the philosophyof social science.To begin with, Marx's work is
not recognizedby the mainstreamof academicsocial sciencewith
respect to his own estimate of its importance: as a theory of
capitalistaccumulationand crisis. Even among those social sci-
entistswho style themselvesradical or Marxist, the majority do
not acceptMarx's own theory as an accuraterepresentationof
social reality. From the point of view of the philosophyof social
sciencethis is doubly unfortunate:not only becauseMarx's work
representsthe outstandingexample of a social theory, but also
becauseit includes, as part of its basic framework of ideas.
explicit attentionto problems of interestto philosophers-while
raising a new one: the curious inability of social theorists to
accept the validity of Marx's analysisof capitalism despite its
remarkablescientific strength.
My own interestin the philosophicalquestionswith which this
book is concerneddevelopedwhile teachingcollege courseson
Capital. The greatestbarrier to students' understandingMarx,s
ideas,I discovered,was their difficulty with seeingtheseideasas
formulated from a point of view outside of the academicdisci-
plines of economics,sociology,or political science.The basic
task I faced was to explain what Marx meant when he described
his theory as a critique of political economy rather than as a
contribution to economics.Explaining this involved discussing
Marx's conceptionof economicsas notjust a theorybut a system-
atizationof ideasthat help define as well as explainthe forms of
behavior constitutingcapitalistsocial life.
In this way I was led to the classicanthropologicalproblem of
the relation between the cultural insider, the native for whom
reality is defined by culturally developedforms of experienceand
thought, and the outsider,the scientific investigatorwho wishes
both to graspand to explainthe native's way of life. The problem
is posedto the extentthat there is conflict betweenthe anthropolo-
gist's and the native's understandingof the latter's worlcl: what,
rl :rrr.yllrirrg.r'lrrrjrrstily tlrcscicrrt
ist'scluinrlo providcan cxplana-
Irorrol rlrtivc cust()llts srrpcriorto thc nativc'sown, while doing
;rrst icc (o tlrc rrrlc playcd by culturc in the vcry constructionof
('\l)r'l'icnccas wcll as in its comprehension?
I lirrrntlit usclul to intmduccMarx's anthropologicaltreatment
ol lris own culture with a short discussionof Evans-Pritchard's
;rrrirlysis ol'Azandewitchcraftbeliefsand practices.Justas Evans-
l'r itchardattemptedto explainAzande ideasand rituals in terms
ol thcir place in native social life, Marx wished to explain the
t'ontinuedfaith in economics-despiteits striking weaknessas a
scicnce-displayedby the nativesof his (and our) own culture in
lcnns of the centralrole played in capitalistsocietyby this system
ol'idcas. This comparisonproved helpful for my students;it led
rrrcto the questions,aboutthe natureof scientificthinking and its
rclation to our everyday knowledge of social reality, and about
thc nature of that reality itself, discussedin this book.
While I believe that Marx's theory provides a starting point
I'rom which thesequestionscan be answered,the presentessayis
ncitheran introductionto that theory nor a fullscalediscussionof
Marx's conceptionof scienceand its realizationin his work. The
basic question it confronts is an abstractone: how is scientific
knowledgeof sociallife possible?That an answerto this question
does not merely depend on general philosophical considerations
but requires reference to the particular historical circumstances
of the development of social theory will emerge from simulta-
neousconsiderationof a sister question: why does such knowl-
edge seem impossible, despite some two hundred years of intel-
lectual effort? My argument, in a few words, is that the
difficulties of the social scienceshave been due, not to the inher-
ent resistanceof social life to scientific explanation,but to the
culturally determined inability of would-be social scientiststo
subject their own categoriesfor social experience-those of cap-
italist society-to the cross-cultural comparison on which the
possibility of scientific understandingof social life depends. It
was by his ability to look at capitalismfrom the perspectiveof its
eventual abolition that Marx succeededboth in explaining the
 SOCIALKNOWLEDGE

limits of bourgeoissocial theory and in constructinga scicntific

alternative.

Even so short a book as this one is a product of the intellectual

work of many people besides the author. My thinking on
the topics here explored reflects years of discussion with other
members of the group Root & Branch, and especiallywith Eliza-
beth Jones Richardson. By inviting me to teach with him, Fred
One sees how subjectivism and objectivism, spiritualism and
Moseley made possible the sharedstudy of capital in the course
rnaterialism,activity and passivity,lose their antitheticalcharac-
of which the comparison of economics and witchcraft first oc-
tcr, and hencetheir existenceas such antitheses,only in a truly
curred to me. Hilary Putnam supervised the dissertation from
social stateof affairs; one seeshow theoreticalantithesescan be
which this book has developed, in the processkindly discourag-
rcsolvedonly in a practical way, only through people'spractical
ing me from chasing a number of philosophical red herrings'
cncrgy, and how their resolution is therefore by no meansonly a
Rochelle Feinstein and Ilse Mattick read the penultimate draft
task fbr knowledge,but a real problem of life, which philosophy
with care; their suggestionsled to many improvements.Finally,
lrls bccn unableto solvepreciselybecauseit hasconsideredit as a
anyonewho knows his writings will recognizehow much I owe to
prrrcly theoreticalproblem.
my fatheq Paul Mattick. From him I learned not only how to
understandMarx but why it is important to do so: that the point of
Marx, Economic and Philosophical Manuscripts (1844)
understandingthe world really is to change it.
 I
Introductory

In his preface to the secondedition of The Structure of Scientific

Theories (1977), Frederick Suppe expressesconfidence that the
"confusion and disarray within philosophy of science" of the
previousdecade"is becomingresolved" and has given way to a
relatively coherentstateof the field.t However true this may have
been, the same certainly neither could nor can be said for the
awkward subtopicof Suppe'sfield, the philosophyof social sci-
ence. In the words of the author of a recent survev.

The initial impressiononehasin readingthroughtheliterature

in andaboutthe socialdisciplinesduringthe pastdecadeor so
is that of sheer chaos. Everything appearsto be "up for
grabs." Thereis littleor no consensus-exceptby membersof
the sa-me schoolor subschool-aboutwhat are the well-estab-
lishedresults,the proper researchprocedures,the important
problems,or eventhe mostpromisingtheoretical approaches
to the studyof societyand politics.2
 5()(_ tAl KNO W TI t x , t
This statcrnentonly rcflectsa striking clcgrccol'agrecmenton
the unsatisfactorystate of social theory itself, not only among
philosophicaloutsidersbut among stock-takinganthropologists,
sociologists,economists,and political scientiststhemselves.Al-
though economics is often touted as the best developed of the
social sciences, it is freely admitted by respectedpractitioners
that its more theoretical areas have little connection with eco-
nomic reality, while relatively empirical efforts, such as econom-
ic forecasting,also bear little impressivefruit.3In Tjalling Koop-
man's words, "after more than a century of intensive
activity . . . [one] is led to concludethat economicsas a scientif-
ic disciplineis still somewhathangingin the air." He is forced to
plead the case for his discipline on the basis of its unknown
future, in a statementso striking as to deservequotation at length:
If overestimation of the rangeof validityof economicproposi_
tionsis the Scyllaof "informal" economicreasoning, a cor_
rect appraisalof the limited reachof existingeconomictheory
may causeus to swerveinto the Charybdisof disillusionment
with economictheory as a road to useful knowledge.The
temptationto identify the resultsof existingeconomictheory
with economictheoryas such[sicl]-and to disqualifyboth in
onebreath-is strongestfor the experienced economicadvisor
to government or business, to whomthelimitationsof existins
theoryare most painfully apparent.a
Tom Bottomore, in his UNESCO-sponsored.,guide to the
problems and literature" of Sociology, expressedthe general
opinion when he wrote that "there is not, at the presenttime, any
general body of sociological theory which has been validated or
generally accepted."s This remark would, I think, be widely
accepted as applicable to anthropology, demography, political
science,and history (for thosewho classifythis field as, or in part
as, a science).In general, as Bottomore says, ..one powerful
argumentagainstthe scientificcharacterof the socialscienceshas
been that they have not in fact produced anything resembling a
natural law."6
l N l l l o ( ) tj cl Ol ? Y
Sortte(lrkctlrisstlrlcol trlllirs as trn intlicationol'thc "youth"
ol tlrc socialstutlics.rcla(ivcto (say)physics,but posita body of
objcctivcluwsas thc goal to which thesestudiesare well on their
lvrry. Othcrs scc thc lack of succcssin reaching this goal as
rcprcscntingsonrcthingrnore serious.indeedas demonstrating
thc l'undamentally misguidednatureof the current ideaof a social
scicnce itself. The question as to the very possibility of social
scicnce has roots far deeper than the insufficiencies of actual
work in these areas: ultimately in the only slowly dissipating
dil'l'iculty felt by Europeanculture with treating human beings,
arrdtheir socialinteractions,as part of nature.This difficulty has
tlcclined in physical and social anthropology since the day of
Darwin's struggle againstthe idea of man's essentialseparate-
ness; it is weakening further with the growing acceptanceof
biological approachesin psychology;but in the realm of social
theory the accessabilityof human phenomenato scientific think-
ing remains a topic for debate.
As May Brodbeck observesin the introduction to her reader in
The Philosophy of the Social Sciences,the modern idea of such
sciencescan be tracedto the thinkersof the Enlightenment."See-
ing men as part of the natural order, they envisioneda scienceof
man and society modeledon Newton's explanationof heavenand
earth, by whose explanationthe potentialitiesof man could be
realizedto form a more just and humane social order."7 It has
remaineda guiding assumptionin this tradition of social theory
that it is possibleto discover general laws of social behavior,
explanatory of observed phenomena. And though it has been
hoped that such laws would serve as guides to policy, believers in
the possibility of such scienceshave also emphasizedthe necessi-
ty and possibility of "value-freedom" in researchand analysisin
theserealms. That is, while the subject-matterof the social stud-
ies is such as to inspire strong feelings and valuations-such, in
short, as to express a point of view or way of life-this tradi-
tion has held that these feelings and values must and can be
put aside in order to attain results as objectively valid as those
 SOCIAI KN O W II t ) G I
of the physicaland biologicalscicnces.
A line of descentof equal age, however,stems from Vico's
distinctionbetweenhuman and naturalhistory: that we may have
special knowledge of the former as it is we who make it. There
are many versions current of the idea that consideration of the
human realm demandsspecial modes of inquiry, radically differ-
ent from the methods appropriate to natural science. In general,
this line of thought involves some form or another of the claim
that
what is neededaboveall is a way of lookingat socialphenom-
enawhich takesinto primaryaccountthe intentionalstructure
of humanconsciousness, andwhichaccordingly placesmajor
emphasison the meaningsocialactshavefor the actorswho
performthemandwho live in a realitybuilt out of their subjec-
tive interpretation.8
People have not only made their worlds: they make and remake
them. This radically "subjective" aspectof human phenomena
poses (or is held to pose) all sorts of problems for scientific
research.It is, for instance,supposedlynot the scientist'stask to
advocatenormative positions. But since humans are self-inter-
preting beingsdifferences in interpretationof social experience-
as betweena socialgroup and a scientistinvestigatingit-involve
differences in the ways in which the potential data for a social
theory are constructed.On a relativelysimpletechnicallevel, this
problem appears as that of the interference of the researcher's
valueswith thoseof the objects of his study. A more fundamental
objection,to the very possibilityof social science,maintainsthat
social life can only be understoodfrom "within"-that is to say,
in terms of the reasonsand values that define what the actions
constituting a way of life are, to begin with. If the student of
societyis "outside" the culture investigated,he will not be able
to understandit; on the other hand, if he is "inside" it, it is
difficult to see what meaningcan be given to the "objectivity"
associatedwith the idea of science.
One reason the methodology of the social sciencesis so well
IN tI? OD U C IOR Y
is lhc llrct (lrlr(thcsctwo lincs ol'dcscenthavealways
poptrlir(ctl
lrccn inlinralclyintcrtwincd il'not incestuously, at any rate by
t'vcly varictyul'cross-cousin marriage.No one, for example,has
insistcd nrorc lbrccfully than Max Weber on the necessityfor
valuc-liccdornand scientific objectivity in the social sciences.
Yct it was hc who described Verstehen,"interpretive understan-
tling" of subjectivepoints of view, as a fundamentalmethod of
thcsc fields. Similarly, Ernest Nagel has argued that one can both
admit the subjectivenature of the stuff of social analysisand hold
that "the logical canons employed by responsible social scien-
tists . . do not appear to differ essentially from the canons
crnployed . . in other areasof inquiry."e
This attempted synthesis of objectivity and subjectivity, I
think, characterizesthe majority opinion within the social sci-
ences today: such sciencesare supposedto produce objective
knowledgeof the causes,nature, and consequences of the com-
plex structuresof subjective meaning that constitute social phe-
nomena. Max Weber's definition is classical:
Sociology. . . is a sciencewhich attemptsthe interpretative
understanding of socialactionin orderto arriveat causalexpla-
nationof its courseand effects.In "action" is includedall
humanbehaviorwhen and in so far as the actinsindividual
attaches
a subjective meaningto it.
The idea of "subjectivity" here puts emphasison the individ-
ual as the basic unit of analysis,whose "action is social in so
far as, by virtue of the subjective meaning attachedto it by the
acting individual. , it takesaccountof the behaviorof others
and is thereby oriented in its course."rO An example of such
"interpretive understanding" would be the explanation of eco-
nomic behavior in terms of the ' 'profit motive' ' of entrepreneurs.
In contrast, Marx's social methodology appearsto champion
"scientism" in its purest form. What greater contrast to Weber's
attemptedsynthesisof subjectiveand objective elementscould be
found than Marx's declaration in the Preface to Caoitctl. Volume
I, that
 SOCINI KNO W II I ) G I
indiv idualsare h e rcd c a l t w i th o n l y s o l a r a s thcy are pcrsoni fi -
cationsof econorniccategories,the bearersof particular class-
relations and interests. My standpoint, fiom which the devel-
opment of the economic formation of society is viewed as a
process of natural history, can less than any other make the
individual responsiblefor relations whose creaturehe remains,
socially speaking, however much he may subjectively raise
himself above them.rr
Marx's attitude is explicitly connected with his intention to con-
struct a science of society on the model of physics and chemistry.
Beside the passage just quoted we may set the statement of his
"standpoint" which is one of the best-known passages from
Marx's works:
In the social production of their existence,men inevitably enter
into definite relations, which are independent of their will,
namely relations of production appropriate to a given stage in
the development of their material forces of production. The
totality of theserelations of production constitutesthe econom-
ic structure of society, the real foundation, on which arises a
legal and political superstructureand to which corresponddefi-
nite forms of social consciousness.The mode of production of
material life conditions the generalprocessof social, political,
and intellectual life. It is not the consciousnessof men that
determines their existence, but their social existence deter-
mines their consciousness. [C]hanges in the economic
foundation lead sooner or later to the transformation of the
whole immense superstructure.In studying such transforma-
tions it is always necessaryto distinguish betweenthe material
transformation of the economic conditions of production,
which can be determined with the precision of natural science,
and the legal, political, religious, artistic or philosophical-in
short, ideological-forms in which men become conscious of
this conflict and fight it out. Just as one does not judge an
individual by what he thinks about himself, so one cannotjudge
such a period of transformation by its consciousness,but on the
contrary, this consciousnessmust be explainedfrom the contra-
dictions of material life, from the conflict existing betweenthe
social forces of production and the relations of production.r2
IN II? OI) U C IOR Y
'l'lrisprcscrilrtiorr
sccrrrsin corrlirrrtrity
to thc work, the critique
ol poliliclrlcconorrry. to whoscl'irstpublicationit servedas pref-
rrlory nurlter. I)ucs l.urtMarx displaya model of capitalistsociety
irr which cconolnic lirrces compel a developmentalhistory in
wlrich pcoplc arc caughtup and which thrustsideologicalviews
rntl nxrdesof social action upon them? Even an historian like
lr. P. Thompson,who claimsa Marxian heritage,protestsagainst
wlrat he considers the exclusion of subjectivity from Capital,
judging that work to be "a study of the logic of capital, not of
capitalism," for "the socialand political dimensionsof the histo-
ry, the wrath, and the understandingof the class struggle arise
I'nrm a region independent of the closed system of economic
Iogic." t:
At the same time, we should not forget, as Thompson does,
that Capital is presented not as an economic theory, but as a
critique of economics,itself describedby Marx as "superstruc-
tural" . If we look at Marx's actualanalysisof capitalistsociety,a
picture seeminglyin conflict with his methodologicalprescrip-
tion emerges. Capital begins with the analysisof commodities
and so of commodity exchange.The latter is not, as may at first
appear, to be identified with the transfer of objects from one
person'shand to another's.The owners of commodities, ratheq
m:rstplace themselves in relation to one anotheraJ persons
whosewill residesin thoseobjects,and mustbehavein sucha
way thateachdoesnot appropriatethe commodityof the other,
and alienatehis own, exceptthrough an act to which both
parties consent.The [exchangers]must thereforerecognize
eachotheraJ ownersof privateproperty.. . . The contentof
thisjuridical relation . . . is itselfdeterminedby theeconomic
relation. Here the personsexist for one anothermerely as
representatives, and henceowners,of commodities.. . . tl]t
is asthebearersof theseeconomicrelationsthattheycomeinto
contactwith eachother.ra
The last sentenceestablishesphraseologicalcontinuity with the
passagequoted from the Prefaceto Zur Kritik. And yet the over-
ro soctAr KNowtt D(;t t Nlt ?ol) uclolt Y II

all meaningseemsto bc quite diffbrent.Marx's idca is that the rol i;11[111

v1vlctlgcclrnnotbc olr.jcctivc bccauscthc criteriafbr the
modern conceptof property,as defining a certain social point of ol
tr rrtlr sociologic;.rl
prropositions arc alwaysdefinedasrelativeto
view and rule for behavior, is necessaryfor capitalist market Hcrc I will turn to anotherstandbyof the
tlrccrrl(rrrcin clr"rcstion.
exchangeto take place (since, for example, collectively owned plrilosophicallitcraturc on the methodologyof social science,
products cannot be exchangedwithin the collective). The eco- lrvrrnsPritchard'sstudyof Azandewitchcraft,to arguethat this is
nomic relation is fundamental in the sense that the exchange- rrrruntcnableposition. Chapter4 will sketchan alternativecon-
relation is involved in the definition of "property" (since a t r'ption of socialunderstanding,and in Chapter5 I will return to
good'sbeingone's propertyhere implies one's havingthe right to Mrrrx's attemptto formulate methodologicalpresuppositionsfor
sell it). But at the same time the exchange-relationitself is de- rr scicntific studyof social life in sucha way as to do justice to its
scribed with referenceto the kind of phenomenacommonly called rrrtcntionalaspect.Finally, I will examinesomeimportant conse-
"subjective" in the social scienceliterature. It is only because (luencesof thesepresuppositionsfor our understandingof social
they treat each other as private proprietors that the individuals scicnceand its relation to other forms of social knowledse.
concernedcan be "bearers" of this economic relation. In fact,
this formulation seemsnot incompatiblewith Weber'sdefinition
of "social action", cited above. The economicrelation depends
upon the existenceof the juridical relation, and the latter clearly
involvesan element,at least,of "consciousness":at any rate we
are here within the "superstructure" which the economic "foun-
dation" is supposedto determine!
With this we touch, of course,on one of the classicproblemsof
Marxism. By bringing this problem of the relation of basis to
superstructureinto contact with the bourgeois social theorists'
discussionof the relation betweensubjectiveand objective as-
pects in the construction of social reality, I wish to illuminate the
issuesraisedin both traditions. I will begin with an examination
of the concepts of "subjective" and "objective" as they have
been developedin modern social theory. Like many others, I will
use Peter Winch's critique of Weber as a starting-point. Despite
its serious weaknesses,Winch's work remains valuable for its
isolation of the central issuesin modern social theory. I will try to
show that the categoryof "subjectivity" has maintained its place
in this contextlargely thanksto a confusionof the "intentional"
characterof social phenomenawith the idea that such phenomena
are unobservable. Having shown that the latter idea rests on a
misunderstandingof the nature of scientific observation,I will
turn in Chapter 3 to an examinationof Winch's argument that

"Subjective"ond
"Objective"
According to F. A. Hayek, "there are no better terms available
to describe [the] difference between the approach of the natural
and the social sciencesthan to call the former 'objective' and the
latter 'subjective.' Yet these terms are ambiguous and might
prove misleading without further explanation."r On the one
hand, he says, the facts of social phenomenaare as objective as
natural facts, in the sensethat they are given independentlyofthe
observer,and are thus public, given equally to every observer.On
the other hand,the socialsciencesare orientedtowardssubjectiv-
ity in that they "deal in the first instancewith the phenomenaof
individual minds, or mental phenomena, and not directly with
physical phenomena. They deal with phenomenawhich can be
understoodonly becausethe object of our study has a mind of a
structuresimilar to our own. " In addition, "the term 'subjective'
stresses. . . that the knowledge and beliefs of different people,
while possessingthat common structure which makes communi-
cation possible,will yet be different and often conflicting in many
respects." Thus, finally, the social sciencesmay be describedas
t2
'':;( l l i l l( l tvt " n N t) "o BJtci l vt" l3
''('()n('cr.nc(lwitlr nliul's corrsciousor rcllcctcd action, actions
ulrr'r'cil l)crs()r)
can l'rcsaid to chooscbetweenvariouscourses
(rlx'nlo hinr. ."1
It'lrooscHryck us a startingpoint not becausehis analysisis
prrlticularlyclcar or pcrsuasive,but becauseit includes-without
t lt'rrrlyclistinguishingthem-the chief usesof the pair "subjec-
tivc/objcctive" to be found throughoutthe literatureof and about
tlrc social sciences.Three usesof this pair of terms can be distin-
guished in the passages just quoted. (1) First of all, social facts
rrrcheld to be objectiverelativelyto the observer'ssubjectivity,in
(hc sensethat they exist as objectsfor the observer'sconscious-
rrcss,rather than as its products.(2) However,saysHayek, if we
Iook at these objects of social study we see that they themselves
have a subjective character, in that their existence depends on
somepersons'minds-the minds of the personsstudied.For this
rcason the facts relevant to social theory are, "in the first in-
stance," those which exist for the membersof the social group
under investigation.One result is that social phenomenaexist
distinctively in a world defined by choice and valuation, rather
than by naturalcausality.(3) Finally, Hayek stressesthe individ-
ual as the basicunit of the socialworld. Society'sstuff is subjec-
tive becauseit is composed of individuals and becauseno two
individualsare alike.
This third use of "subjective" will not be dealt with here. It
has played an important role in various "individualist" formula-
tions of social theory, and there is a large literature devotedto the
topic.3In general,for obvious reasonsthe individualsdealt with
by social theorieshave not been real, particular people, but ideal
representativesthereof, defined either in terms of a putative
human nature or in terms of the institutions. norms. etc. of a
given social situation. This is clearly to be seenin the caseof the
classicalconcept here, Max Weber's "ideal type" or "the theo-
retically conceivedpure type of subjective meaning attributed to
the hypotheticalactor or actors in a given sort of action."4 (An
example is the ideal type of the entrepreneurin capitalist society,
supposedlymotivated purely by economic gain.) Models of social
14 SOCIAI K NO W LTDG T
action usingsuchideal typesarc intendedto yicld approxintations
to, and so explanationsof, real social behavior, in which the
actual individuals are much more complexly motivated. As is
already indicatedby the phrase, "ideal type", Weberianindivid-
ualism (like methodological individualism generally) amountsto
little more than the recognition that society is made up of individ-
uals; the actual burden of explanation is carried by the type
concepts whose definition requires reference to social relations
betweenpersons. 5 One could even say that the ' 'ideal individual' '
of social theory has been a conceptual representation of what
anthropologists call a "culture"-a systemof learnedpatternsof
life sharedby members of a society-as embodiedby the individ-
uals who have learned it. We will return to the conceptof culture
and its relation to the "subjective" below.
SinceHayek's first useof "subjective/objective" merely indi-
catesa minimum condition for use of the word "knowledge"-
that there be an object to be known in some senseindependentof
the knower-the rest of this chapterwill focus on Hayek's second
use of these terms. To begin with, he indicates, if we wish to
explain human behavior with respect to any natural phenomena,
the latter must "not be defined in terms of what we might find out
about them by the objective methods of science, but in terms of
what the personacting thinks about them. " Hayek elaboratesthis
in two ways. First of all, he claims, "any knowledgewhich we
might happen to possessabout the true nature of the material
thing, but which the people whose action we want to explain do
not possess,is as little relevantto the explanationof their actions
as our private disbelief in the efficacy of a magic charm will help
us to understandthe behavior of the savagewho believes in it. "6
We will spendmost of Chapters3 and4 dealing with a version of
this claim, for, unlikely as it may seem, it has been advancedas
basic to the nature of social knowledge. Interestingly, however,
this point has been developed by critics of the "mainstream"
social scienceof which Hayek is a representative,such as Peter
Winch.
We shall here deal onlv with the more conservative formula-
( , IIVI " NN D "OB]EC IIVE"
:i IJt}.II I5
Irorrol'tlrc crlrrlnrstbclwccn "sub.jcctivc"and "objective," as
rt'prcscrrlcrlby Wcbcr'sassuurption that in additionto the subjec-
Irvc sidc ol'a socialpracticethereare associated "processesand
plrcrrrlrrrcnawhich arc devoid of subjective meaning," but of
wlrichaccountmustbe takenin socialexplanation.Admitting that
tlrc rcalitiesof physicalnatureplay their role in sociallife, Weber
is concernedto stressthat this role is shapedby the impositionon
nrture of human-determined meanings.THayek presentsa similar
vicw, although he (typically) does not note its difference from its
rrxrreradically subjectivistcousin. In his words,
Neithera "commodity" or an "economicgood," nor "food"
or "money' ' canbe definedin physicaltermsbut only in terms
of viewspeoplehold aboutthings. . Nor couldwe distin-
guishin physicaltermswhethertwo menbarteror exchange or
whethertheyare playingsomegameor performingsomereli-
giousritual. Unlesswe can understand what the actingpeople
meanby their actions,any attemptto explainthem, i.e., to
subsumethem under rules which connectsimilar situations
with similaractions,are boundto fail.8
This passagemakestwo different points. First, it observesthat
socially significantphenomenaare culturally and not (ust) phys-
ically defined: what one group acceptsas food, for instance,may
be rejected as inedible filth by another. Second, Hayek extends
the point into a rejection of what is generallycalled "behavior-
ism" as a method for social science, arguing that to understand
humanaction requiresnotjust what he calls "physical" informa-
tion-e.g., observationsof body movements-but also informa-
tion about what the human subjectsmean by their actions.
It is important to seethat theseare two distinct points, however
closely relatedthey may be in the minds of the majority of writers
on social theory. The first only specifies something to which we
will give the vague name "culture" as the subject matter of the
socialsciences.The secondinvolvesa certainview of the relation
between "the mental" and "the physical," and so an inter-
pretation of those concepts themselves. The latter concept is
r6 socrAt KNowttt)Gt
identified, as the label indicates,with the dornainof'phenomena
open to senseperception,while thesein turn are takento be none
other than the objects describableby categoriesof physical the-
ory: categoriesof motion through space-time.Events in this
domain are to be explained by reference to causal laws ("rules
which connect similar situationswith similar actions." in the
caseof socialevents)."The mental", in contrast,is the world of
phenomenaunobservablewith the techniquesof natural science
but understood by way of the mind-to-mind communication of
meanings, and explained typically by reference to the reasons
given for actions.
Here again Max Weber makes carcfully explicit what Hayek
suggestswhen he statesthe difference betweenthe mental and the
physical in terms of the idea of "rationally purposive action."
The distinguishingfeatureof such action in Weber's conception
(as referred to in the definition of sociology quoted above) is that
it is behavior regulated by a consciousagent in accordancewith
somepurposeof the agent's-in Weber's words, "a subjectively
understandableorientation of behavior."e To understand it re-
quires interpretationof the action in terms of the grounds that the
agenthad, or can be seenashavinghad, for his action. This, what
Weber called explanatory "adequacy on the level of meaning,"
he contrastedwith "causal adequacy" of interpretation, which is
achievedby the subsumptionof a given action under an empirical
generalization.to
The aim of social science,for Weber the "correct causal inter-
pretation" of action, is achievedwhen a (typical) action-process
is "both adequatelygrasped on the level of meaning and at the
same time the interpretation is to some degree causally ade-
quate." r I The action, that is, must both be interpretedas rational-
ly motivated and be subsumed under a statistical law linking
occurrencesof a certain situation-say, increasing demand for a
good-with the action in question-Say, raising its price. Both
parts of the explanationare necessary,in Weber's view, since it is
only understanding the means-end structure of an action that
makes possible explanation of it as an action, as opposed to
( - IIVt '' N N I) ''OBJIC IIVE''
I;U I]JI 11
rnstirrctivc or rcl-lcxivclrchavior,whilc only cnrpiricalgeneraliza-
tiorrsyicld cllrsal laws.Socialcxplanationthusrequiresthe com-
lrirrrlionol'two nrcthodsin the humansciences.To the empirical
ruotlcol'study sharcdwith the naturalsciencesmust be addedthe
rrlrcnrtionof Verstehen,interpretiveunderstanding,in which the
irrvcstigatorattemptsto reconstructthe motivationalstructureof
thc actions he is studying. The first studiesactions as "objec-
tivc," the second as "subjective": in this way the claims of
scicnceand thoseof the specialnature of the social subjectmatter
:rrc both satisfied.
But can "subjective" and "objective," once distinguishedin
this way, be so smoothly combined in one explanatory frame-
work? Weber's claim is that Verstehenis as much a scientific
procedureas the making of empirical generalizations.Its essence
is not some form of intuitive "putting oneself in another's
place"-though this may have heuristic value as a means to
Verstehen-but modeling the means-endcalculation that makes a
given action rational. Ernest Nagel has pointed out an obvious
problem with Weber's idea:
In discussingthe adequacyof the methodof Verstehen it is
essentialto distinguishbetweenthat methodconceivedas a
wayof g,eneratrng suggestivehypotheses for explainingsocial
action,and that methodconceivedas a way of validatingpro-
posedexplanations.. . . Ult is generallyrecognizedthat the
methodof Verstehen doesnot, by itself, supplyanycriteria for
the validity of conjecturesand hypothesesconcerningthe
springsof humanaction.r2
Weber himself emphasizedthat "verification of subjective inter-
pretation by comparison with the concretecourse of eventsis, as
in the case of all hypothesis,indispensable."r3Verification is
here identified with what are assumedto be the basic procedures
for testing hypothesesin physical science-checking against the
sense-observationof eventsin space-time.Unfortunately, he add-
ed, such verification is really only possible in "the few very
specialcasessusceptibleof psychologicalexperimentation"-in

IB SOCIAL K NO W TI DG I
cases,that is, when the model of thc motivationof an individual's
actioncanbe checkedunderlaboratoryconditions.In mostcases,
"there remains only the possibility of comparing the largest pos-
sible number of historical or contemporary processeswhich,
while otherwise similar, differ in the one decisive point of their
relation to the particular motive or factor the role of which is
being investigated.' ' 14(An example is provided by Weber's most
famous study, in which the difference in outcome of two pre-
capitalist situationsis explained in part by the presencein one of
the "Protestant ethic. ")
Aside from the many difficulties inherent in such a method of
historical analysis, one can question the very compatibility of
Weber's idea of the empirical verification of subjective interpre-
tation with his conception of the subjective/objectivedistinction.
A large number of motivational models are compatible with a
given course of behavior, to begin with. So long as subjective
interpretation and causal generalizationare treated as two sepa-
rate, if parallel, modes of explanation, it is hard to see how one
can serve as a check on the other.
One way out of this difficulty opensif we reject the distinction
Weber makesbetweencausaland rational explanation. If, that is,
we allow that reasonsmay be causesof observablebehavior, the
two forms of explanation could be integrated, and motivational
models could be evaluated by their success in predicting and
retrodictingaction.That reasonscanbe construedascauseshas,I
think, been convincinglyarguedby a number of writers.t5I will
not go further into this question here, becausethis way out of his
dilemma is not open to Weber. Close though it is to his own
position, he is blocked from reaching it by the conceptualstruc-
ture of his subjective/objectivedualism.
This emergesclearly if we look at Weber's proposedmethod of
comparative sociology, which involves the comparisonof institu-
tional (and motivational) complexesof different social systems.
Such comparison requires the investigator to judge that institu-
tions from different systems are the same or different, or that
typical actions taken by members of different social groups are
..:;IJIIJI
( - IIVI " AN I) "OBJIC IIVt
srrnillrlol tlil'li'r'cnt.Ilrrt by wlutt critcriitarc thcscjudgmcntsof
srrrrillrrity Sinccthc socialcharacterof
irntld il'll'rcrtccto bc ntaclc'l
rrt'tionlirr Wcbcr consistsin thc subjectivemeaningsattachedto it
lry rrrcrnbcrso1'asociety,identificationof socialpracticesseems
to rcquirc knowledge of the correlation between motivational
strtrctureand physical behaviorprior to the observationof the
lrrttcr.But then suchobservationscannotserveto verify interpre-
tutionsof the former. As Peter Winch has put this objection, in
'I'ht ldea of a Social Science, "someone who interprets a tribe's
rnagicalrites as a form of misplacedscientificactivity will not be
corrected by statistics about what the members of that tribe are
likely to do on variouskinds of occasion. ."16
Winch is an interestingcritic of Weber,becausein essencehe is
ilttacking Weber on the basis of the latter's own concept of the
subjective,althoughit is rephrasedin terms drawn from linguis-
tic philosophy.Social relations, in Winch's view, are structured
by shared concepts, which are embodied in and so constitute a
language. Language is described, following Wittgenstein, as
both the expressionof and a condition for a "way of life."
Languageis a social phenomenonby nature, as meaning depends
on language-users' adherenceto rules,which of necessitymust be
sharedand not idiosyncratic.Languageis thus not only a medium
of social life but is paradigmatic of social activity, which is
generally defined by the rules that regulateforms of conduct. For
this reasonthe nature of knowledge about society "must be very
different from the nature of knowledge of physical realities."
The latter is based on the recognition of causal regularities, the
former on the understanding of meaning relations. Knowledge
about society is like knowledge of a language in that one cannot
know a languagewithout knowing the linguistic rules that define
membership in the social group that speaksit. If we leave mean-
ing relations out and just describe what we seewhen we observe
social life, we will leave out what representssociality par excel-
lence.tT
Despitethe difference in terminology, all this is (asWinch says
himself) in fundamental agreement with Weber's emphasis on
 s o( - t nt K N ()WiIl )(;l
subjectivemeaningas dclinitivc ol'sociaraction.as wcll as with
his analysisof subjectivemeaningin terms of norm-govcrnedor
purposive behavior. Thus Winch's ,,rules" correspond to
Weber's "typical motivation patterns,' and to the ..norms" of
post-Weberiansociology.tt On the other hand, on this basis
winch condemnsweber for his attempt to show "that the kind of
'law' which the sociologist may formulate to account for the
behavior of human beings is logically no different from a 'law' in
natural science" by showing that a motivationally comprehensi-
ble scenecan also be given a purely "objective" description.In
this attempt, as Winch describesit,
Insteadof speakingof the workersin [a] factorybeingpaidand
spendingmoney,he speaksof their being handedpiecesof
metal,handingthesepiecesof metalto otherpeopleandreceiv_
ing otherobjectsfor them. . . . In short,he adoptstheexternal
point of view and forgetsto take accountof the "subiectivelv
intendedsense"of the behaviorhe is talking about.
In fact, for winch such a description is not a description of social
behavior at all; weber "does not realize that the whole notion of
an 'event' carries a different sensehere, implying as it does a
context of humanly followed rules which cannot be combined
with a contextof causallaws in this way without creatinglogical
difficulties."re For example, one can break a rule, or,nuk" u
mistakein following one, but one cannotbreak a causalrelation.
Interpretations cannot be added to causal descriptions, for one
can't even describea social situationwithout knowing the inter-
pretation; and for the same reason observation of causal se-
quencesprovides no check on subjective interpretation.
winch makes essentiallythe same point in his discussionof
Pareto's attempt "to treat propositions and theories as 'experi-
mental facts' on a par with any other kind of such fact" or, more
generally,with Durkheim "to consider social facts as thinss."
Discussing the former, Winch comments that
In a senseParetohasnotcarriedhis empiricismfar enough.For
'l; ut llt( lr vl " nNl) "ol] . Jt cllvt " 2l
ls not
u'frlrl tlrc sot'iologicltlobsclvcr ltitsllrcsclttctlttt hi'r^r(n^t('.r
l rt l rl l p coplc lr olding ccr t aill t hcor ics. . . but pcople m aking
ccrtaln nlovctltctttsand s<lunds.Indccd, even describingthem
rrs "pcoplc" really goes too f-ar,which may explain the popu-
lurity ol'the sociologicaljargon word "organism": but organ-
isrns, as opposed to people, do not believe propositions or
crnbracetheories.20
( )nc can say that Winch himself is not carrying things far enough
lrcrc. After all, the senses are not presented with organisms either
but-if they are presented with anything-with sense data. If
thcse are perceived as attributes of physical objects, classifiable
as organisms, why shouldn't the latter be classifiable as people,
and indeed as people holding and discussing views?
There are really two different questions here, which Winch
like many writers on thesetopics is collapsinginto one: (1) how
do we go about describingand explaining a point of view, whether
an individual's or one shared by members of a culture? and (2)
how can we tell what's in people'sminds, when we can't observe
such mental contentswith our senses?The conflation of thesetwo
questionsinto one rests on a conception of observctrion-and this
on a whole system of epistemological categories-that, despite
their longevity, are severely flawed.
To begin with, observing is not done "with the senses" in
Winch's use of the words in any field of study, or indeed in daily
life (as Piaget's studiesof the child's construction of the categor-
ies of experiencehave shown). It is by now widely recognizedby
philosophers of science that such a view of observation is not
applicable even where social scientists and their philosophical
critics often imagine it to be preeminently the case: in physics,
where the categories of observation are clearly not given by
nature but have been historically constructedin the form of phys-
ical theory. This is obvious in the case of the observation of
quarksand leptons,but, as E.H. Hutten pointsout in his book on
The ldeas of Physics,"even the 'direct' observationof a familiar
object, say of a star, presupposes not only a long process
of learning acquired by the individual but also the knowledge
 22 1;O(_tnt KNOWTil)(,1
'SU BJTC IIVT"
N N D "OBJEC TIVE'' 2 3

accumulatedby mankindovcr thousandsr)t'ycars.":r

own pattcrnsol'thought and action, our own wishes,interests,
Given this, it would be perverseto limit s.ciological rbserva-
:rntl pluns, our own standards,criteria, ideals, categories,con-
tion to phenomenadescribable with the constructs of ethology,
ccpts,and fbrms of life. " This correspondsto the question,how
rather than thosedevelopedin the study of society itself. Since we
is social activity to be "objectively" understood?Just as both
know that for the members of any social group studied ..be-
Wcbcr and Winch conflatethis questionwith that of the correla-
havior" in the physicalistsensedoesnot exist, but only behavior
t ion of behaviorwith meaning,the radical translationproblem, as
within the conventionsof a given culture, there is no need to
Wallace sees it, connectsthese issuesby the way in which it
pretendthat the former can be an object of study for the scientific
visualizesthe anthropologist'sposition in the face of his data.
observer either. More generally, there is no reason to give the
Part of his "already acquiredconceptualschemeis immediately
concept of observation in social investigation a content-the
applicable," in the Quineanview to the alien sround on which he
"pure sense-observation of physicaloccurrences"-it has in no I'indshimself
other field of study.what social analystsobserve are precisely the
things to which Winch draws attention, under the rubric of the withoutfear of falsificationor distortion:the physicalpart.
"subjective": businessbehavior, thinking, kinship structures, Applicationof the otherpart, the mentalsystem,has to be
value systems,etc. The important questionis not, how does one workedout. In workingout theapplication of mentalconcepts
know aboutthings one can't observe?but, how doesone observe whatwe haveto go on aredescriptions in physicaltermsof the
things of this kind? environment and of the behaviorof the inhabitants. . . We
The situation here may be usefully comparedwith the problem havemarkedoff . . . an areaof description of humanbeings
of "radical translation" as posedby W.V.O. euine. This prob- which (i) is immuneto rhe dangerof projectionand (ii) pro-
lem indeed is raised in the form of a story about an imaginary videsan evidentialbasisfor applicationof descriptions from
our remaining,projection-prone stock.22
ethnological study. An anthropologist,encounteringa people
whose languagehe does not speak,attemptsto learn their lan-
This is pretty easilytransferableto Weber'saccountof the use of
guageand so their cultural concept-system.In a very interesting
statisticalgeneralizationsas a control on subjectiveinterpreta-
article John wallace has observedthat this problem can be seen
tions of social action;23Winch's position may be stated in its
as motivatedby two issuesimportantfor sometime in philosophi-
terms as the denial that the two systems,the mental and the
cal thinking. The first derives from the view of our conceptual
physical, are applicableto the same phenomena.
scheme as divided into two parts: the system of physical con-
Wallace comparesthis general model of radical translation
cepts,which is "basic and coverseverything"; and ,,our system
with an actualcase:the deciphermentof the Cretanscript, Linear
of semantic, psychological, moral, aesthetic, and social con-
B. According to his account,in the actualpracticeof interpreting
cepts," thementalsystem.The questionsarise:how is the bound-
and learningaboutotherpeople,physicaland mentalconceptsare
ary between the two systems to be drawn? and, how are they
mixed up in single reports in such a way that the two systems
related to each other'?The secondof thesequestionscorresponds
cannotbe treatedas independent.In the interpretationof certain
to the Weber-Winchproblem, how "observable behavior" is to
symbols in the Minoan script, the scholarsworking on the case
be correlated with motivational or rule-following subjectivity.
took as "immediately accessibleevidencewhat for the philos-
The secondissuefinding expressionin the problem of radical
ophers has to be theory [supportedby physicalistevidence],a
translationis that of "the danger of projecting onto others our
semanticalpieceof the mentalframework."2a Wallaceconcludes
24 SO(_ t nI KNO W Tt t ) c t
that thc correctanswerto the question,how the applicationof
mental concepts is constrainedby the application of physical
concepts,is: "The questionpresupposesa process-application
of mental conceptson the basisof physical concepts-which does
not exist." 25
To return from this digressionto our topic, Wallace's point
appliesboth to Weberiansocialscienceand to Winch's argument
againstit, asthey sharethe samedualisticstarting-point.(Indeed,
Winch criticizes Weber from a viewpoint defined by a more
consistentapplication of the latter's concept of the "subjec-
tive. ") The stayingpower of the word "subjective," despitethe
inappropriateness of its applicationto what all agree are inter-
subjective,social conceptsand norms, derives I think from the
implicit referenceto the supposed"mental" portion of the con-
ceptual schemewe have inherited from Descartes,as opposedto
the "objective" or "physical" portion. Even more than the
emphasison the individual as the basic unit of social life, the
continued use of "subjective" and "objective" indicates the
maintenanceof a radical splitting of the phenomenaof human life
from the rest of nature. This is visible, for instance, in the very
concept of "behavior, " which is taken to pick out the physical
side ofaction as distinguishedfrom and thus as observablesepa-
rately from the mental activity that gives it meaning.
This is, however, not the only way human action can be ap-
proached. For example, from the point of view of Piaget's theory
of cognitive development the psychological behaviorist's
"stimulus" does not really exist except as developmentallycon-
structed by the organism involved: so that description of a se-
quence of actions in terms of "stimulus" and "responsc" de-
scribes "behavior" only via implicit reference to what the
behaviorist considers "mental" phenomena.Piaget's conceptual
framework, by making this explicit-by considering intelligence
as the self-regulation of the "behaving" organism-deprives
"mind" and "body" alike of their statusas separateanalytical
objects.26Similarly, some anthropologistshave defined the con-
ceptof "culture'' in sucha way asto allow us to posethe problem
" AN I) "OBJEC IIVE"
SU I].JTC IIVI
ol'cstablishing
a scicnccol'sociallil'c in othertcrms thanthoseof
dcuronstratinga concordancebetween"adequacy of meaning"
ancl"causal adequacy."For example,Kluckhohn offers the fol-
Iowing definitionsof "culture":
thatpart[of humanlife] whichis learnedby peopleastheresult
of belongingto someparticulargroup, and is that part of
learnedbehaviorwhich is sharedwith others.. . . By "cul-
ture" we mean those historicallycreateddefinitionsof the
situation
whichindividuals acquireby virtueof participation
in
or contactwith groupsthattendsto sharewaysof life thatarein
particularrespectsand in their total configurationdistinc-
tive.21
We may then by extensionspeak of the society or group defined
by its possessionof such shared learning as "a culture."
Whateverthe limitationsof this conceptionof culture, suchan
approach to the phenomena of social life does not suggest the
problem of the relation between objective and subjective that
Weber'sdefinition of sociologybrings with it. Kluckhohn, defin-
ing "explicit culture" as "all thosefeaturesof group designsfor
living that might be describedto an outsiderby participantsin a
culture" or that might be observedby the outsider,is careful to
say that
To avoidconfusion,it shouldbe notedthat the basicdatafrom
whichthe anthropologists abstractexplicitcultureencompass
manifestations of "feeling" and"thought"andarein no sense
restrictedto objectsandactsin the narrowbehavioristicsense.
In otherwords,"explicit" doesrzoldraw the line which'.ob-
jective" is supposed to draw from the"subjective."zs
Aside from the matter of behaviorism(which, as noted, does not
characterize Weber's position) what is important here is that
elementsof culturesdescribableas "feeling" and "thought" are
to be thought of neither as privately individual nor as inwarcl, but
as public and in principle observable.An examplethat appearsto
support the Weber-Winch view of the matter will clarify the
 5 (X,INI KNO W II t X, t
point . I n a c ri ti q u c o 1 ' Wi n c h , A l a s d a ir Macl ntyre w rotc:
Supposethat a team of Martian socialscientistsis observing
human behavior. What they are watching we should describeas
chess-playing, but unhappily they lack the concept of a
game. . . They therefore do not discern the rule-governed
characterofthe players' behavior,although they arrive at many
statistical generalizationsabout the movement of small pieces
of wood by human beings. What is it that they do not under-
stand when they fail to understandthese movementsas a game
of chess?They fail of course first of all to grasp the players'
actions as distinct from their physical movements.It is not that
they wrongly explain what is done; rather they fail to identifu
the actions which are to be explained.2e
Maclntyre seems to be saying the same as Winch in the latter's
criticism of Weber. But there is an important difference. As
Maclntyre states it, the failure of the Martians stems not from
their limiting themselves to the "external point of view," from
which they attend only to what is "physically observable," but
from their lack of categories that would make the observation of
chess-playingbehavior possible. Because their culture lacks
games they cannot identify what they are seeing, and will be
forced to describeit in terms availablefrom their own culture.
They are thus in the toils of the problem of projection, but this
derives from the presenceof cultural blinkers on theory and not
from the "subjective" nature of social phenomena.30
To summarize the argument so far, we have seenthat free and
easy use of the opposition "subjective/objective" has allowed
confusion of recognition of the intentional character of social
phenomenawith the idea that such phenomenaare unobservable.
This confusion depends(at least in part) on a reductionist analysis
ofobservation as the passiveregistrationand subsequentorgani-
zation of sensory givens. As this analysis is not tenable for any
science,leastof all the physicalones,thereis no call to apply it to
the social studies.
On the other hand, abandoningthe reductionistred herring
AN D ''OBJEC TIVE'' 2 7
SU BJTC IIVT''
tkrcsnot clirtrinatcthc problem o1'projecticln.Indeed,the correc-
tion of Winch's mistakenview of naturalscienceallows his argu-
nrcnt about the impossibility of social scienceto return with a
vcngeance.The demiseof positivistic philosophyof sciencehas
led to challengesto the notion of objective truth, or the idea that
there is a gradualconvergence,over time, of theory to truth, in
natural science.Somephilosopherssuggestthat no sensecan be
given to the conceptof scientific objectivity. Rather, it is suggest-
ed that the conditions of verification of any theory are set by the
particular "community" of scientistswho work within the "dis-
ciplinary matrix," "paradigm," or "Weltanschaultng" em-
bodied in it. Following this line of thought, the scientist-turned-
sociologistBarry Barnes comparesthe scientistto specialistsin
other types of knowledge, "the prophet, the astrologer, and the
witch-doctor"'. " 'true' like 'good' is an institutionalizedlabel
used in sifting belief or action according to socially established
criteria."3r Winch, too, beginsfrom the fact that scienceis itself
a social activity. As a result, "to understandthe activities of an
individual scientificinvestigatorwe must takeaccountof two sets
of relations:first, his relationto the phenomenawhich he investi-
gates;second,his relation to his fellow-scientists."32The latter
is as determiningas the former; it is only becausethey are taking
part in the same kind of activity, or cultural institution, and
therefore accept the same criteria of identity with respect to
natural phenomena,that scientistscommunicate with each other,
cooperatein doing science,or, simply, do scienceat all. This is
important because,as Hayek noted in his attempt to distinguish
natural from social science, a given natural state of affairs will
come under different descriptions if different criteria are applied
to it.
Now, Winch argues,if sociology (or anthropology)is to be a
sciencelike the natural sciences.
Theconcepts andcriteriaaccording
to whichthesociologist
judgesthat,in two situations,
thesamethinghashappened, or
the sameactionperformed,mustbe understoodin relationto
 1; ( ) ( . t ntKN ()Wt
I t)(,t
C IIVT'' AN D ''OBJEC TIVE'' 2 9
SU B.JI

tlu, rulcs g,vcrtrirtg s..r'i,l,giutl in,t,.tti74ttli,tr.But herc we run

agalnst a difficulty: f<rr whcreas, in the case of the natural
scientistwe have to deal only with one set of rules, namely
those governing the scientist'sinvestigationitself, here whar
the sociologist is studying, as well as his study of it, is a human
activity and is therefore carried on according to rules. And it is
these rules, rather than those which govern the sociologist's
investigation,which specify what is to count as ,.doing the
same kind of thing" in relation to that kind of activity.33
For example, the question whether two actions performed by
members of a culture are both acts of prayer is a religious ques-
tion: when a sociologist or anthropologist studies religion the
criteria of identity of the institution studied must be taken not
from sociology but from the religion studied itself. The anthro-
pologist Marvin Harris has described the situation of the social
s c ient is t us ing Pi k e ' s d i s ti n c ti o n b e tw e en ,,emi c" and ..eti c' ,
observation and analysis. The former
have as their hallmark the elevationof the native informant to
the status of ultimate judge of the adequacyof the observer's
descriptionsand analyses.The test of the adequacyof emic
analysesis their ability to generate statementsthe native ac_
cepts as real, meaningful, or appropriate.In carrying out re_
search in the emic mode, the observer attemptsto acquire a
knowledge of the categoriesand rules one must know in orcler
to think and act as a native. . . . Etic operationshave as their
hallmark the elevation of observers to the status of ultimate
judges of the categoriesand conceptsused in descriptionsand
analyses.The test of the adequacyof etic accountsis simply
their ability to generatescientifically productive theories about
the causesof socioculturaldifferencesand similarities. . . .
Frequently,etic operationsinvolve the measurementand juxta_
positionof activitiesand eventsthat nativeinformantsmay find
inappropriateor meaningless.3a
cal" phcnonrcna. or with that bctwccn thc rcalm of the meaning-
lirl and that of thc observable. Indeed, according to Harris,
The importanceof Pike's distinction is that it leadsto a clarifi-
cation of the meaning of subjectivity and objectivity in the
human sciences.To be objective is not to adopt an etic view;
nor is it subjective to adopt an emic view. To be objective is to
adopt the epistemologicalcriteria . . . by which scienceis de-
marcatedfrom other wavs of knowins.35
But-Winch would reply-there's the rub. A culture is constitut-
edby a setof categories;naturehasno point of vieq but a way of
life is definedby its practitioners'.The abandonmentof Weber's
dual method, of Verstehenplus "objective" observation,neces-
sary on Weber's own premises, apparently leaves no criteria
availableforjudging the correctnessofan observer'sdescription
or analysisof a culture, to the extent that this involves interpreta-
tion of the culture in categoriesforeign to it. The sin of projection
seemsan Original one for social science:even aside from the
issuesof "subjectivity/objectivity" or of the diff'erencebetween
causesand reasons,the meaningfulcharacterof social phenom-
ena seemsto rule out the objectivity necessaryif we are to be able
to speak of a scientific study of society.

Harris is careful to point out that the emic/etic distinction is not

to be identified with a distinction between ,,mental" and .,physi_
 N N A[IEN SOC IETY
U N I) II? SIAN I) IN G 3I

ll'arncwork liont urotivational clcmcnts with which they are fa-

nriliar. As Allied Schutz describes "the paradox that dominates
Wcber's whole philosophy of social science," Weber

Understonding
on
postulatesas the task of social sciencethe discovery of intended
meaning-indeed, the intended meaning of the actor. But

AfienSociety this"intended meaning" turns out to be a meaning which is

given to the observer and not to the actor.2

But, objects Winch, this is to disregard the essence of culture-

that it consists of a set of shared meaninss.

When one is dealing with any kind of social "things," .

their being . . social, as opposedto physical, in character
dependsentirely on belongingin a certain way to a systemof
ideas or mode of living. lt is only by reference to the criteria
governing that system of ideas or mode of life that they have
any existenceas intellectual or social events. lt follows that if
the sociological investigator wants to regard them as social
Max Weber's exampleof explanatory,.adequacy
on the level of events. , he has to take seriously the criteria which are
meaning"-explaining .,what is, accorclingto our
current norms applied for distinguishing "different" kinds of actions and
of calculationor thinking, the correct solution to identifying the "same" kinds of actionswithin the way of life
an arithmetical
problem"t-sheds no right on the problemof he is studying. It is not open to him arbitrarily to impose his
the interpretationof
a culture different from the observer'sown. The own standardsfrom without. In so far as he does so. the events
action of arith-
metical calculationcan be interpretedwith adequacy he is studyinglose altogethertheir characteras soc:ialevents.3
on the level
of meaningby referenceto ..our" current norms
because,it is
assumed,the calculatorobservedis obeyingthe same So far, Winch's position seems uncontroversial. Surely, how-
norms. But
how is the behaviorof membersof a socialgrorp who ever we intend to analyze a set of cultural phenomena, we should
believe(for
example) in witchcraft or magic to be interpreted strive to identify its elements on the basis of the criteria of the
by would_be
scientificobserverswho do not sharethesebeliefs? culture studied. Examples Winch offers of his recommended
The latter can
statethe means-endrelation in the former's action (,,Thev approach certainly seem to embody ordinary anthropological
do that
to stop witches from attacking them") in terms common sense: e.9., that a psychoanalytical investigation of Tro-
of the group,s
own norms of thought. But to the question, ..What briand neuroses "could not just apply without further reflection
is that man
doing when he performs his anti-witchcraftritual?" the concepts developeC by Freud for situations arising in our own
the answer,
"warding off witchcraft," is not very iiluminating. society, " but would have to investigate differences between (say)
observers
who are not of the group believing in witchcraft European and Trobriand concepts of family structure.4 No psy-
c;n place this
action in an explanatory framework only if they choanalyst will take a prohibition of continuing investiga-
consiruct that
tion as a methodological guideline, and no anthropologist or

30
32 SOCIAI KNO W II t ) G t
sociologistwill adrnit to inrposinghis own cultural catcgorics
"arbitrarily" on the material of his field research.
In fact, if put in these terms the purported contrast of social
with natural scienceis at least questionable.While nature (as I
just said) can't be said to have a point of view, it has been said to
have or constitutean order of its own, to which the community of
scientistsmust accomodateits conceptseven while it assimilates
natural phenomena to them by defining "criteria of identity."
Physics is at any rate no more arbitrary than social sciencecould
be. So in what way is Winch's argument an argument againstthe
possibility of social science,rather than an argument againstbad
social science?To put it differently, at what point, as Winch sees
it, does the problem of projection reveal itself to involve an
irreducible conflict betweenthe cultural point of view investigat-
ed and the categoriesof the observing scientific community? It
might seem as though Winch means to rule out analysesof a
culture that go beyond what membersof it would give in explana-
tion of their own behavior.And this would certainly rule out a
great deal of what passesfor social science.In general,as Fritz
Machlup has pointed out, participantsin a culture, even while
they may (and indeedmust) know the rules and criteria regulating
social behavior in that culture, may have only a very vaguenotion
of how the parts of social life in which they participate fit togeth-
er. Machlup offers the example of a Martian anthropologist-
perhapsa colleagueof Maclntyre's Martian-who observesthe
stock market and interviews its denizens.
Sinceprobably999 out of 1000personsworking on the stock
marketdo not really know what it doesandhow it doesit, the
mostdiligent observer-plus-interviewerwould remainlargely
ignorant.Alas, economicscannotbe learnedeitherby watch-
ing or by interviewingthe peopleengagedin economicactivi-
ties. It takesa gooddealoftheorizingbeforeonecangraspthe
complexinterrelations in an economicsystem.s
This expressesin other words a point made in the previous
chapter,that observation is meaninglessexcept within categories
AN
U N [) II? SIN N I) IN C ; AIITN SOC IETY 33
proviclcclby sontcthcorcticaloricntation.In Winch's terms, sci-
cncc rcquircs categories,and criteria for applyingthem, that go
hcyond observation, in the sensethat they are createdby groups
of scientists.The workers in the stock market, lacking economic
theory, are in a bad position to observe their own activity; the
economist can do so just because he makes use of categories
different from those of the insiders-categories of economics,
rather than those of the mode of life that includes working in the
stock market. But this meansthat he will employ criteria unused
by the subjectsof his study; he will impose his own standardsin
the analysisof their activity, though certainly (he will claim) not
arbitrarily.
We can compare this case to that of a linguist's analysis of a
languagein terms of grammatical rules unelicitable from native
speakers.Despite the view of some linguists that their theories
representnot just theories of languagebut theories of language-
users, explaining what the latter know as opposedto what they
know how to do, it seemsthat at least we can say that linguists'
knowledge of a language is different from a native speaker's,
employing as it does theoretical conceptsunknown to the native.
Winch's views need not be taken to be in fundamental conflict
with such a point of view. Indeed, in his eyes,
thetestof whethera man'sactionsaretheapplicationof a rule
is not whetherhecanformulntett but whetherit makessenseto
distinguishbetweena right anda wrongway of doingthingstn
connectionwith what he does.Whenthat makessense'thenit
must also makesenseto say that he is applyinga criterionin
what he doeseventhoughhe doesnot, and perhapscannot,
formulatethat criterion.6
A closer look discovers difficulties in this formulation. It
seems odd to say that someone is applying a criterion that he
cannot formulate. As with the linguistic example, it seemsto me
more natural to speakof such a criterion as a feature of the theory
or description of the man's behavior,rather than of his knowledge
of or obedienceto a rule. And in any caseWinch's formulation of
 34 :;(X-IAI KNO W II I X; I
thc criterionfur thc uscol'"critcrion" thrcatcnshis own clistinc-
tion betweenthe researcher'scriteria of identity and thoseof the
object of study. To say that what is important is "whether it
makes sense" to formulate a criterion leavesopen the question,
to whom it must make sense,especiallyin the case where the
actor himself "cannot formulate the criterion."
winch's discussionof social rules, in fact, slidesin a superfi-
cial way over the fact, acommonplacefor ethnographers,that (in
Malinowski's words) informants produce "at best that lifeless
body of laws, regulations,morals, and conventionalitieswhich
ought to be obeyed,but in reality are often evaded."7The rules
discerned by the investigatormay therefore be quite different
from thosewhich may be elicited from an informant. As a result,
winch's test of when someoneis following a rule permits ethno-
graphersto formulate rules that contradictrules the nativesstud-
ied would formulate. It is possible, however,to state Winch's
criteria for non-projective social description in a way which
avoidsthis problem. what is essentialis not that all the categories
employed by the outside researcherbe given in the culture he
studies,but that they do not contradictthe classifications(not the
rules) of social activity basic to the culture. In Winch's words.
althoughthe reflectivestudentof society,or of a particular
modeof sociallife, mayfind it necessary to useconcepts which
arenot takenfrom theformsof activitywhichhe is investigat-
ing, but which are takenrather from the contextof his own
investigation,
still thesetechnicalconceprsof his will implya
previousunderstanding of thoseother conceptswhich belong
to the activitiesunderconsideration.s
what kinds of social studiesare unacceptableunder winch's
rule? Most clearly, it rules out certain kinds of cross-cultural
comparison of institutions and practices. winch develops his
argument to this point in the context of a critique (already noted)
of Pareto's theory of "residues" (psychologicallydetermined
constantsof human behavior)and "derivations" (ideologiesand
institutionsin which the residuesare fleshedout). The example
UNDTI?SINNI)ING
AN ALIEN SOCIETY
Winch cliscusscs is Parcto'sanalysisof Christianbaptismas a
spccialcaseof the allegedlytrans-culturalphenomenonof cleans-
ing rituals. Pareto'sidea is that this phenomenonis found in some
lirrm in all cultures;the variouscharacteristicsand explanations
of the ritual in different cultures being derivations from it, ab-
straction from which revealsthe essential,sharedcore (hence
"residue"). Christians share with members of other cultures a
practice of ritual cleansing; what differ between them are the
specific featuresof the practice and the reasoningsexplaining the
efficacy of these various features.
Now, says Winch, "a Christian would strenuouslydeny that
the baptismal rites of his faith were really the samein characteras
the acts of a pagan sprinkiing lustral water or letting sacrificial
blood. Pareto, in maintaining the contrary, is inadvertently re-
moving from his subject-matterpreciselythat which gives them
sociologicalinterest:namelytheir internalconnectionwith a way
of living. "e It is what differentiatesthe two ritualsthat givesthem
sociologicalinterest,that gives them, that is, the meaningsthey
have for the members of the cultures concerned: for the connec-
tion of this element with the remainderof their ways of life is
articulated exactly through the differer:ces. To identify elements
from different cultures is to commit the root error exemplified by
Weber's behavioristic description of monetary exchange (see
Chapter 2, above). There the descriptionof behavior,divorced
from the intended meaningsof the actors, lost all senseas social
description; here the sensegiven is a concoction by the social
theorist rather than a representationof the actual social meaning
involved. In Winch's words.
ideascannotbetorn outof theircontextin thatway: therelation
betweenideaand contextis an internal one. The ideagetsits
sensefrom the role it playsin the system.It is nonsensical
to
takeseveralsystemsof ideas,find anelementin eachwhichcan
be expressedin the sameverbalform, andthen claim to have
discoveredan ideawhich is commonto all the svstems.ro
Paretois no straw man here. It is worth rememberinsthat E.E.
 :;(X INI KNO W II I X, I
Evans-Pritchard, in his attack rcductivcthc.rics ol'rclisi'n.
citesParetoapprovinglyas one 'n who indicatedan importantd'irec-
tion for analysisand research:"in spiteof the superficialityand
vulgarity and the confusion of his thoughts, pareto saw the prob-
lem correctly."rt 1O take a modern example,Jack Goody, who
has expendedmuch effort in demonstrating errors of projection
committed by anthropologists imposing their categories on the
people they study,nonethelessdescribesthe "experimental
meth-
od" of anthropology as "comparing, not social systemsor soci-
eties or cultures as such, but specified variables under different
social conditions . ."12 Such comparison,of course, requires
abstractionof the "variables" from their social contexts, and
identification of them between systems. And Marx certainly
seemsto fall, as winch sayshe does,into the sameclassas pareto
with respectto winch's critique, given the seemingsimilarity
of
the conceptof "derivation" to Marx's conceptof .:ideology.-,,r:
Winch's argumentis in fact, as he wishesit to be, an attackon
the sociological and anthropological tradition as a whole. The
claim of the social sciences has always been that one can go
beyond ethnologicaldescription to the formulation of generaliza-
tions applying acrosscultures, yielding knowledge fundamental-
ly different from that possessedby cultural insiders. It is a com-
monplacethat anthropologyis "a comparativediscipline,', both
because"by conventionand establishedtradition of doing field-
work in a culture other than one's own, the anthropologisiuses
a
comparative framework in his study" and becauseanthropologi-
cal theorizing has always, "with very few exceptions," invorved
"intercultural or cross-culturalexplicit comparison."ra
Even in
questioningthe results of the anthropologists'searchfor
cross-
cultural regularities, Evans-pritchard identified this searchas the
only path to a possible social sciencecomparable to the physical
sciences,concluding, "my scepticismdoesnot mean that I think
that we should cease to look for such regularities as can be
establishedby various forms of the comparativemethod.',rs
It is obvious that some form of transculturalconceptformation
is essentialto the project of a general science of society. and in
I,IN I) II? SIN N I) INN
GN AIITN SOC IEIY 37
thi s proccss social scicncc inr poscs it s own cr it cr ia on cult ur al
phcnonrcna. As Harris puts it, in his scientific tough-guy way,
I cannot agree . . . that our etic conceptual resourcesfor the
study of the behavior stream are dependentupon emic studies.
The etic concepts appropriate for the study of the behavior
stream are dependenton their statusas predictive elementsin a
corpus of scientific theories.r6
The very voice of the Anti-Winch! For Winch, to adopt such a
perspectiveis to give up all hope of understandingsocial institu-
tions or practices,to dignify the error of projectionwith the name
of science. Gift exchangeand monetary exchangemay have in
common the mutual transfer of objects between persons. But to
treat them as in essenceidentical, as Mauss did in his celebrated
essayon The Gift, ignoring the sensegiven to money by its place
in the whole system of generalizedcommodity production, is to
commit in theory an error of the magnitude of that committed in
practiceby the Indians who, without knowing it, sold the island
of Manhattan.
Cultural elements can only be understood as rule-governed
behavior,and the rules exist only as statedin terms of the concepts
ofthe particularculture concerned.To abstractfrom the specific-
ity of theseconceptsis therefore to render cultural interpretation
impossible. It follows that interpretation is possible only from
within a culture, for rule-governed action can only be described
and explainedin terms of the relevantrules. "'IJnderstanding,"'
in social study, Winch says explicitly, "is grasping the point or
meaning of what is being done or said."17 So that
If we aregoingto comparethesocialstudentto an engineer.we
shalldo betterto comparehim to anapprenticeengineerwho is
studyingwhat engineering-thatis, the activity of engineer-
ing-is all about.His understanding of socialphenomena is
more like the engineer'sunderstanding of his colleagues'ac-
tivitiesthanit is like the engineer's
understanding of the me-
chanicalsystemswhich he studies.r8
 1X)(.tn I K NO W {|t ) ( . t
Socialunderstanding, cultural intcrprctation,is, thcn, lar I'rorrt
science,it is, for Winch, a species
of Wittgensteinian philosophy:
exploring the "grammar" of a culture, it aims not to discover
new "facts" but to "leave everythingas it is."
It will be instructiveto compareWinch's position, as he him-
self hasdone. with that of Evans-Pritchard. In his book on Witch-
craft, Oracles, and Magic Among the Azande the anthropologist
comes to ask,
In writingaboutthebeliefsof primitivepeoplesdoesit matter
one way or the other whetherone accordsthem validity or
regardsthem as fallacious?Take witchcraft again. Does it
makeany differencewhetherone believesin it or not, or can
onejust describehow a peoplewho believein it, thinkandact
aboutit, andhow thebeliefaffectsrelationsbetweenpersons? I
think it doesmakea difference.for if one doesnot think that
thepsychicassumptions on whichwitchcraft-beliefsarebased
are tenable.one has to accountfor what is commonsenseto
othersbut is incomprehensibleto oneself.re
If one believesthereare witches,then a setof ideasaboutwitches
and how to ward off the effectsof their evil powers will call for no
particularexplanationbeyondnormal perspicuity.If one doesnot
share this belief, one will be moved to ask how others come to
have it. What are the phenomenathat give rise to this beliefl
Answeringthis questionwill involve thinking aboutwitchcraft in
a way quite different from that of the native believers. It may be
placedinto a transculturalclassificationof socialpracticeswhich
would seem shockingor absurd to its practitioners.For Evans-
Pritchard,for instance,the belief in witchcraft he studiedamong
the Azande was an example of "ritual" behavior, which he de-
fined as "behavior accountedfor by mystical notions"-that is,
by notions ascribing to phenomenaproperties that they do not
have and that cannot be defendedfrom the point ofview of "our
body of scientific knowledge and logic. " The latter are "the sole
arbitersof what are mvstical. common sense.and scientific no-
U N t) TI? SIAN D INAN
G AL IENSOC IETY 39
tions" (althoughhc addsthat "thcir judgementsare neverabso-
lutc''1.zt t
But Evans-Pritchard himself has acknowledged, as Winch
points out, that the scientificthinking of Europeansjust as much
as the "mystical" thinking of the Azande represents"patterns of
thought provided for us by the societiesin which we live."2r
According to Winch, it is such patterns of thought that define
what is real and what is unreal. Sincethesepatternsare represent-
ed by langu&ge, "we could not in fact distinguish the real from
the unreal without understandingthe way this distinction operates
in the language." However,Winch holds, it is just on failing to do
this that Evans-Pritchard'sargumentdepends.The anthropolo-
gist wants
to work with a conceptionof realitywhichis not determinedby
its actualusein language.He wantssomethingagainstwhich
thatuseitselfcanbe appraised.But this is not possible;andno
morepossiblein thecaseof scientificdiscoursethanit is in any
other.We may ask whethera particularscientifichypothesis
agreeswith reality and test this by observationand experi-
ment. . But the generalnatureof the datarevealedby the
experimentcanonly be specifiedin termsof criteriabuilt into
the methodsof experimentemployedandthese,in turn, make
senseonly to someonewholly conversantwith the kind of
scientificactivitywithin which theyare employed.22
The excellenceof Evans-Pritchard'sethnologyof the Azande
has made this a standard case in the philosophy of the social
sciences;I interrupt the discussion,therefore, for a brief sum-
mary of his findings. The Azande are (to use the ethnographic
present) a people of the Sudan. Their lives are pervaded by
interrelated notions of witchcraft and magic. Although Evans-
Pritchardcalls their beliefs "mystical" in his sense,Zandedo not
experiencethe phenomenaof witchcraft as unnatural, awe-inspir-
ing, or mystical in our usual sense of that word. Rather they
experience the activities of witches as a normal part of every-
day life, as natural (although ill-understood) in their mode of
 SOCtnt KNOWIil)(,t
operation.witchcrali is hcld to bc an inhcritcdphysicalcondi-
tion, which is activated by ill-feeling of various sorts. Whcn
activated it is responsiblefor damageto the interest of the person
hated,envied,etc. If termitesdestroya man's house,or his crops
fail to grow, or he or his child becomesill, the questionimmedi_
ately arises, who is responsible, via witchcraft, for the disaster.
The witch responsibleis discovered by the use of a number of
oracles, with varying degrees of reliability. The most reliable
Evans-Pritchardcalls the poison oracle, as it consistsof feeding a
poison to chickenswhile asking questions,answeredby the death
or survival of the fowl. As the witch or witches responsiblecan
only be individualswho bear ill will to the victim, only the names
of suchindividualsare put to the oracle,which selectsone, more,
or noneof them. This leadsto a ritual in which the now-identified
witch is askedto call off his powers, and to further testing to seeif
he has done so. If witchcraft results in death, vengeanceis exact-
ed by witchcraft. In addition, magicalrites and medicinesmay be
used for protection against witchcraft.
In describing all this, Evans-Pritchardmakes a great effort to
be faithful to the zande point of view. Dealing with their concep-
tion of witchcraftas embodiedin a physicalsubstance,he goesso
far as to tell us, "I haveonly once seenwitchcrafton its path," in
the form of a bright light, without apparentexplanation, passing
in the night in the direction of a certain person's house. As it
turned out, the next morning found an inmate of that housedead!
In Evans-Pritchard's words, in Azande terms
This eventfully explainedthe light I had seen.I neverdiscov-
eredits realorigin, whichwaspossiblya handfulof grasslit by
someoneon his way to defecate,but the coincidenceof the
directionalongwhichthelight movedandthesubsequent death
accordedwell with Zandeideas.23
In the word "real" here we feel working the point of view to
which Winch objects as forcing the experience of the culture
studied into the categorial framework of the investigator.Indeed
Evans-Pritchardtells us that he himself "always kept a supply of
U N I) tR SIAN I) IN GAN AL IENSOC IETY 4 'I
l.roisonlirr thc uscol'nry householdand neighborsand we regulat-
cd our affbirs in accordancewith the oracles' decisions.I may
rcmark that I found this as satisfactory a method of running my
home and affairs as any other I know of. "24 At the same time,
however, he statesbaldly that "witches, as the Azande conceive
them, clearly cannot exist."25Given this, he asks a questionno
Zande would ask: why do people believe in them?
Could not the quarrel betweenthe anthropologist and the phi-
losopher be resolved by distinguishing explicitly between "real-
for-Azande" and "real-for-Englishmen, " with witchcraft falling
under the former but not under the latter? After all, the starting
point of anthropology is the fact of the differences between cul-
tures.As Winch himself presentsit, in an article taking issuewith
Evans-Pritchard along with some philosophical writers on the
question of "Understanding a Primitive Society," the aim of
anthropology is to presentan account ofthe beliefs and practices
of a very different people "that will somehow satisfy the criteria
of rationality demandedby the culture whose concept of ratio-
nality is deeply affected by the achievementsand methods of the
sciences,and one which treats such things as a belief in magic or
the practice of consulting oracles as almost a paradigm of the
irrational."26 The positive aspectof Evans-Pritchard'swork is
his attempt to sketch the rationality of the Azande, to communi-
cate some senseof how the world can be experiencedcoherently
through this particular pair of spectacles.
Evans-Pritchard'serror, according to Winch, is that he does
not stop there, but goes on to compare and contrast the Azande
view of the world with his own along the dimensions of truth/
falsity, reality/unreality, science/mysticism.But only by detach-
ing it from its cultural context, its "internal connection" with the
rest of Zande life, can the sentence"There are witches" uttered
inZande be equatedsemantically with the samesentenceuttered
in translation by an English anthropologistfor whom it represents
an untruth. This equation commits the sin of Pareto, identifying
an element of one culture with an element of another,thus identi-
fying them in terms that could not be acceptedby membersof the
 42 SOCINI KNO W II I ) G I
culture studied. "zandc notionsol'witchcrafi," Winch p.ints
out-in complete accordancewith Evans-pritchard-"clo not
constitutea theoreticalsystemin terms of which the Azandctry to
get a quasi-scientificunderstanding of the world."27 Therefore.
the attemptto construethem as suchis a mistake.But it is only in
the context of such an attempt that it makes sense to discuss
whetheror not witchcraft is real-for-Englishmen,or whetherit is
true-for-Englishmenthat there are witches. To compare ..real-
for-Azande" and "real-for-Englishmen" is to assume that
"real" itself meansthe samein both contexts.But this
is exactly
what winch wishesto deny; sincein his view "to give an account
of the meaningof a word is to describehow it is used"28and it is
irnpossible"to work with a conceptionof reality which is not
determinedby its actual use in language."2eHere we see that
cultural relativismis leadingto what we may call the incommen-
surability of languages:the view that there is simply no way in
which an assertion in zande about the reality of witches .un b"
compared,with respectto truth-value,with an assertionin Ens_
lish about the reality of witches.
To go into more detail, it's not so much that the Azandehave a
different explanation than we do for some class of phenomena,
but that they try to explain a classof phenomenathat we don't
recognizeto exist. Thus, Evans-pritchardtells us,
zande belief in witchcraftin no way contradictsempirical
knowledgeof causeandeffect.The world knownto theienses
is just as real to themas it is to us.30
Mystical causationis not invoked as a substitutefor natural cau-
sation, but as a supplementto it. If, for example, termites eat
away the wood posts of a granary so that it collapses and kills
people sitting beneathit, the only relationship we iecognize be-
tween the fact that the building collapsed and the fact that these
people were sitting beneathit is the accidentalcoincidenceof
two
space-timepaths at that place and moment.
We have no explanationof why the two chainsof causation
intersected
at a certaintime and in a certainplace,for thereis
no interdependence betweenthem.Zandephilosophycansup_
NN ALI I N SO CI EI Y
UNI ) I I ?SI NNI ) I NG
pl y thc r r r issinglink. 'l'hc Zandc knows t hat t he suppor t swer e
unclernrincdby termites and that pcople were sitting beneath
the granary in order to escapethe heatand glare ofthe sun. But
he knows besideswhy thesetwo eventsoccurred at a precisely
similar moment in time and space.It was due to the action of
w i tchcr af t . 3r
Similarly,
we shall not understandZande magic, and the differences be-
tween ritual behavior and empirical behavior in the lives of
Azande, unless we realize that its main purpose is to combat
other mystical powers rather than to produce changesfavorable
to man in the objective world.32
The point of magic, that is, is at least in part to influence the
experiencedcourseof events; but ' 'experience'' here includesthe
experienceof eventsas involving mystical phenomena.Indeed,
such phenomenaare part (to make a Winchian comment) of the
"objective" world of the Azande. The method of magic is not to
affect the course of nature directly but to affect other mystical
powers that are affecting nature.
What this revealsis that the Azande themselvesmake a distinc-
tion betweenritual and empirical action. They do not use Evans-
Pritchard's terms for this distinction, but they do distinguish
between the categories distinguished by those terms: "Azande
undoubtedly perceive a difference betweenwhat we consider the
working of nature on the one hand and the workings of magic and
ghostsand witchcraft on the other hand, though in the absenceof
a formulated doctrine of natural law they do not, and cannot,
expressthe difference as we expressit."33 They speak, for in-
stance, of witchcraft as the "second spear" in an event, on the
analogyof the secondspearthat may be thrown in the killing of an
animal. "Hence if a man is killed by an elephantAzande say that
the elephant is the first spear and that witchcraft is the second
spear and that togetherthey killed the man."3a
All this leavesus with the problem of understandingtheZande
 44 SO(_t^t KN()WII t)(:,t
asscrtion,"Thcrc arc witchcs." whcn Evans-pritchard tclls us
that "witches, as the Azande conceivethem, clearly cann.t cx-
ist," the subordinateclauseembodiesa claim to be contradicting
rhe zande assertion. And yet one sentenceaffirms the existence
of witches from within the mysticar world-view, while the other is
obviously rejecting the existenceof witches from the stanceof a
scientificallyminded Englishman.Since the Azande themselves
distinguish between empirical and mystical categories, Evans-
Pritchard's view seemsillegitimate. perhapsa more proper state-
ment would be somethingon the order of "witches cannotbe said
to exist in the same way as that in which we say that chairs and
electronsexist." As winch makes this point, with respect to
statementsaffirming the existenceof the God of the old resta-
ment, what the reality of God amountsto
can only be seenfrom the religioustraditionin which the
conceptof God is used,andthis useis very unlikethe useof
scientificconcepts.. . . The point is that it is within the reli-
gioususeof language thattheconception of God'srealityhas
its place,though,I repeat,this doesnot meanthat it is at the
mercyof whatanyonecaresto say;if this wereso, God would
haveno reality.35
Thus one might say,the Azandeand we agree,more or less,on
how to think about the class of phenomenawe call the realm of
empirical causation.In addition, the Azande treat of a realm of
phenomenawhose existencewe do not recognize. To say this is
not to saythat thesephenomenadon't exist, but to say that claims
or denialsof their existenceare madein terms of categorieswhich
we do not use.For an English anthropologistto say,i'witches do
not exist" is for him to say that the system of conceptsin which
"witch" has a central part plays no role in his culture.
This is
expressedvery well by Evans-pritchard himself:
I have often been asked whether,when I was among the
Azande,I got to accepttheir ideasaboutwitchcraft.This is a
difficult questionto answer.I supposeyou can sayI accepted
U N I) tI? SIAN I) IN (AN
; AL IEN SOC IEIY 45
thcrrr.I hlcl no choice. ln nry own culture . . . I rejected,and
rc.ject,Zande notionsof witchcraft. ln their culture . . I ac-
ceptedthem; in a kind of way, I believedthem. Azande were
talking about witchcraft daily, both among themselvesand to
me; any communicationwas well-nigh impossibleunlessone
took witchcraft for sranted.36
This is closeto Winch's "idea that the conceptsusedby primi-
tive peoples can only be interpreted in the context of the way of
life of those peoples."37 This statementtaken by itself, however,
is ambiguous. If taken to mean that analysisof a foreign culture
involvesstatingthe meaningsof that culture's conceptsin terms
of their relationshipsto other portions of the culture, few would
disagree. On the other hand, "interpretation" in this context
generally involves some notion of communicationbetweencul-
tures; it requiresthe idea of translation, the forging of a systemof
equivalencesbetween the concepts of one culture and those of
another. And this activity of translation is carried out by the
anthropologist, on the basis of his cultural system. For this rea-
son, translationis exactlywhat the Winchian is not interestedin.
As Winch's position is expressedby Sara Ruddick,
Understanding ratherthantranslation is in question. . pre-
dictingandtranslating arenottantamount the
to understanding
sensethey [the membersof the alienculture] seein what they
are doing. . . . Rather,as we realizein trying to understand
others'or our own societies,making senseis a prolonged
attemptto seethe realities"shown" in theconceptsof others,
i.e., to be ablenot only to saywhat they say,but to meanwhat
they mean.38
This is relatedto Winch's view of reality as revealeddifferently
in or through different cultural concept-systems;the attempt to
establishequivalencesbetween cultures is both an attempt to do
something impossible and a waste of the opportunity to expand
our experienceof the world. Thus, accordingto Winch, in study-
ing a culture S we are ideally "seeking a way of looking at things
 SOCIAI KN O W TI I ) G I
which goes beyondour prcvious way in that it has in sontc way
takenaccountof and incorporatedthe other way that nrembersol'
s have of looking at things. Seriouslyto studyanotherway of life
is necessarilyto seek to extend our own-not simply to bring the
other way within the already existing boundaries of our own,
becausethe point about the latter in their present form, is that
they ex hypothesiexclude that other."3e rhus it is not thatzande
conceptsshould be interpretedto match the distinctions we make.
Rather, "since it is we who want to understandthe zande cate-
gory, it appearsthat the onus is on us to extendour understanding
so as to make room for the zande category,rather than to insist on
seeing it in terms of our own ready-madedistinction between
scienceand non-science. "4o
While this may be only polite, it is difficult to see how our
categoriescan be expandedin this way, if winch's conceptionof
a culture is correct. Either we are dealingwith two systems,each
with its own criteria of rationality and truth, of which the ele-
mentsare meaninglesswhen dissociatedfrom the whole. or else
it is possibleto build conceptualbridges betweenelementsof the
systems,for us to understandaZande category.But suchbridges
go in both directions: if we can stretchour categoriesto meel a
zande concept, the latter hasipsofaclo been brought into congru-
ence with our ways of thought. ..I do not want to say." Winch
writes,
that we are quite powerlessto find ways of thinking in our
societythatwill helpus to seethe zandeinstitutionin a clearer
light.I onlythinkthatthedirectionin whichwe shouldlook is
quitedifferentfrom what [thoseI am criticizing]suggest.ar
Evans-PritchardevaluatesAzande magic as a semi-scientificpro-
cedure.But this is a mistake,sayswinch. compared to science,it
is irrational, but only becausethis is the wrong dimension for
comparison.Instead,zande practiceshouldbe put on the dimen-
sion of religion, andzande ritual compared to christian prayer.
I do not say that Zandemagicalrites are at all like Christian
UNI ) I RSI NNDI NG
AN ATI ENSO CI EI Y 47
l )rl rycf sol'suppliclt ionin t hc p<lsit ivcat t it udet o cont ingencies
which thcy cxpress. What I do suggestis that they are alike
in that they do, or may, express an attitude to contingen-
cies; one, that is, which expressesrecognitionthat one's life is
subject to contingencies, rather than an attempt to control
these.a2
It is ironic that Winch's argumenthas brought him here to do
exactly what, in his book, he attackedPareto for doing, when the
Italian sociologistidentified Christian baptism with Near-East-
ern blood sacrifice.No evidenceis given that the Azande would
accept the identification of their magic with Christian prayer,
even in the limited senseof Winch's comparison.And in fact,
Evans-Pritchard'sbook shows clearly that the Azande regard
magic as an attempt to control contingency.Checks for witchcraft
are carried out before any major decisionsare made or plans
executed.Illness is combattedby confrontationof the witch re-
sponsibleand by magical healing.
Indeed, Winch ends up giving a standardanthropological-so-
ciological explanationof ritual as a ' 'form of expression,'' or of
"drama," as "ways of dealing (symbolically) with misfor-
tunes"a3This is certainlynot how the Azandethemselvesthink of
their behavior; it is as much to put an interpretationdrawn from
our own culture on theirsas is the evaluationof witchcraftbeliefs
from the standpointof natural science.
Winch hasa retort availableto this argument.Any comparison
betweenculturesmust of courseinvolve referenceto somecom-
mon element.As we have seen,Winch rejectsthe possibility of
the direct identification of elementsof two different cultural
systems.Comparison is possible, however, from this point of
view, on the basisof somea priori-determinableset of "prime"
conceptsapplyingto all societies.Theseconceptscannot,by the
argumentsdiscussedin the earlier part of this chapter,be discov-
ered empirically (since any empirical comparisonpresupposes
commensurabilityof elementsfrom different cultures)but can be
discoveredthroughphilosophicalanalysis.Suchanalysis,Winch
claims. shows that
 S O CI A IK N OWID
I Gt N N AL IENSOC IETY
U N t) tI? SIN N D IN G

thc very concepti.n o1'hunranlil'c revcalsccrtain lunrlar'cntal []vans-Pritchard'saccount makes it clear that for the Azande
not ions . . . w h i c h ... i n a s e n s e d e ternti nc thc ..ethi cal "There are witches" expressesnot just a moral or religious
space" within which the possibilitiesof good and evil in human conception but also identifies certain persons as causal agents
life can be exercised.. . : birth, death, and sexual rela_ involved in certain unpleasantevents.47
tions. . . The specific forms which these conceptstake, the
It seems,then, as though Winch's effort to find a standpoint
particular institutionsin which they are expressed,vary very
outside of any culture, that would provide common terms of
considerably from one society to another; but their central
reference for the comparison of cultures, must be adjudged a
positionwithin a society'sinstitutionsis and must be a constant
factor. aa
failure. Morality and religion, along with birth, death, and sex,
must be cast into the same wildernessof cultural relativism as
rationality and truth. Indeed, how could it be otherwise?If con-
It is hard to avoid seeingin this ideaa closerelativeof pareto's cepts are defined by their role in cultural systemsof concepts,
residuesand derivations.All that distinguishesthem in principle how could there be a trans-culturally valid cultural analysis?We
is the claim of a priori knowledge, deriveclfrom ,.conceptual are thus squarelyfaced, as a consequence of Winch's relativism,
analysis." But can this claim of Winch's be supported?_even with an inescapableincommensurabilityof cultures.
apart from generaldifficulties with the ideasof a priori congni-
tion and conceptual analysis. perhaps birth, copulation, and
death are central concerns for all social systems, or even are
bound to be. This can hardly be considered"conceptuallyneces-
sary," however;one can easilyimaginea culture in which birth,
for example,is not particularly signalized.In any case, the no_
tions of birth, sex, and death don't seem to be the same for all
cultures.a5Similarly, when Winch speaksof ..the very concep_
tion of human life" he begs the question, whose conception?
What his argumentactuallyestablishesis that he, Winch, is able
to understandAzande witchcraft and magic by interpreting these
practicesand the conceptionsthey invoke as versionsof his very
different concepts of prayer, good and evil, and human life in
general.It is almostanticlimacticto point out that in doing this he
has totally departed from the sensethat the Azande seem to sive
their own conceptions, since what seems to characterize ttese
beliefs is an integration of ethical with empirical-manipulative
practices; so that witchcraft is for them a matter both of physical
action and of moral significance.a6winch objects to E'ans-prit-
chard's characterizationof Azande witch-talk as bad science,by
sayingthat it's not sciencebut religion, or ..ideasof good, evil,
and the meaning of life." But this separation of the relig_
ious/moral sphere from the scientific is quite ethnocentric.

Understonding
ond
Explonotion
The conceptof incommensurability,to which Winch's views on
the nature of social understandingbrought us, is borrowed here
from the philosophyof natural science.winch himself suggests
the analogy. Comparing magic to science,he writes, is ,,like
observingthat both the Aristotelian and the Galilean systemsof
mechanicsuse a notion of force, and concludingthat they there-
fore make use of the samenotion."rA purportedtranslationof a
certain Greek word by a certain (seventeenth-century)Italian
word does not establish equivalence of meaning. Curiously
enough, almost exactly this example has been employed by paul
Feyerabend,with T.S. Kuhn (at times) the best-knownadvocate
of the incommensurabilityof scientifictheories.In an article on
the relations between successivetheories of the same physical
phenomenon,Feyerabendbegins with the thesisthat ..the mean-
ing of a term is not an intrinsic property of it but is dependent
upon the way in which the term has been incorporated into a
theory"-its scientificuse, to employ the wittgensteiniancatch-
phrase.2From this it follows that "it is impossible to relate
50
AN D EXPL AN AIION
U N t) tI? SIN N I) IN G 5I
succcssivcscicntil'icthcoricsin sucha mannerthat the key terms
thcy provide firr the description of a domain D, where they
overlap and are empirically adequate,either possessthe same
meanings or can at least be connectedby empirical generaliza-
tions."3 They are literally incommensurablein the sensethat
measurementobservationsdefined in terms of one theory Iover
the domain D will not be translatableinto and so comparablewith
observationsdefined in terms of a theory 7' covering the same
domain.
Whathappenswhentransitionis madefrom a restrictedtheory
T' to a wider theory 7 (which is capableof coveringall the
phenomena whichhavebeencoveredby Z) is somethingmuch
more radical than incorporationof the unchangedtheory T'
into the wider contextof Z. What happensis rathera complete
replacement of the ontologyof T'by the ontologyof ?",and a
corresponding changein the meaningsof all descriptiveterms
of Z' (providingthesetermsare still employed).4
To completethe analogywith the anthropologist'sposition in the
face of witchcraft, as Winch construesit, we note that Feyerabend
explainsthat the incommensurabilityof theories"does not mean
that a person may not, on different occasions,use conceptswhich
belong to different and incommensurableframeworks. The only
thing that is forbidden for him is the use of both kinds of concepts
in the same argument."5
In the same way, as we have seen,on Winch's view Azande
witchcraft and the scientific views current in England at Evans-
Pritchard'stime are metaphoricallyincommensurable.6 There is
no way in which the two world-views might be jointly compared
with reality,because(to recall Winch's words) "both the distinc-
tion betweenthe real and the unreal and the conceptof agreement
with reality themselvesbelong to" a given languageor mode of
life. This is not to say (Winch maintains) that there is no distinc-
tion to be drawn between something's being true and its being
thought to be true. Thus "God's reality is certainly independent
of what any man may care to think, but what that reality amounts
 1;O(_lnt KNOWTil)Gl
to can only bc sccn liont thc rcligious tradition in wlrich thc
concept of God is used, and this use is very unlikc the usc of
scientific concepts, say of theoretical entities."T Truth, along
with rationality, is relative not to an individual's thought but to
the universeof discourse"in terms of which an intelligible con-
ception of reality and clear ways of deciding which beliefs are and
are not in agreementwith this reality can be discerned."s There-
fore, sciencecannotdisprove witchcraft beliefs, although a scien-
tific rvorld-view can displaceone in which witchcraft has a place,
just as in Evans-Pritchard's own experience the language of
witchcraft in certain contextsdisplacedhis scientific world-view.
Despite Feyerabend's claim that such an extreme epistemo-
logical relativism constitutes an "anarchist" theory of knowl-
edge,to me the conservativeimplicationslinking the view that (in
Feyerabend'sphrase)"anything goes" to the "mainstream" so-
cial sciencewinch wishesto criticize are evident.Like parsonian
sociology,Winch's view doesnot lend itself to the explanationof
change,either in ideasor in social reality, and containsan implicit
justification of any social dogma in terms of the standardsof truth
and rationality set by the relevant mode of life. perhaps this in
part explains the popularity of Kuhn's earlier work, along with
Feyerabend's,among political and social theorists, including
some writers who claim descentfrom the Marxist tradition. Such
views, I believe, indicatethe abandonmentof an earlier concep-
tion of human rationality as having the capacity to understandthe
world, including the problems and needsof human beings, and to
reshapethe social as well as the natural world the better to meet
those needs.
To say this is not to say that relativism is wrong. But a sign that
something is wrong with it is the fact that relativism is a member
of that classof principles, of which the best-knownmember is the
verification theory of meaningfulness,whosebeing true is incom-
patible with their being affirmed to be so. If the English can'r
understandthe Azande, there is no way for Winch to grasp that
the Azande have a different or any view of reality, and therefore
no way for him to argue that the scientific and the magical world-
AN t) L XPL AN AIION
U N I) II? SIN N I) IN G
vicws arc incornrncnsurablc! It seemcd(pp.44-a5 above)that the
anthropologistwas stuck in the positionof believingin witchcraft
among the Azande, disbelieving it back in England, and striving
not to let his two personaemingle and engage in cross-cultural
conversation.But when we spell out Winch's view, we seethat, if
such a situation could occur, the anthropologist could not be
aware of it, but would be condemnedto the existenceof a cultural
Dr Jekyll and Mr Hyde. For it follows from the incommensura-
bility of cultural systemsthat translation is not only insufficient
but impossible.There is simply no way for the anthropologistto
saymeaningfully,"The systemof conceptsin which 'witch' hasa
centralpart plays no role in my culture." This is becausethere is
no way for the Englishman to know the meaning of the Zande
words conventionally translated by "witch"-boro (ira)
mangu-as the rules of application of these words in Zande are
quite different (on Winch's analysis)from those of "witch" in
English. Living in a scientific culture, we cannot even use the
word "witch" as the inhabitantsof Salem, Massachusettsused it
in the seventeenthcentury. There is no way in which we can say
that the women executedthere for witchcraft were not guilty-
and furthermore no way in which we can say that they were. We
can only say that in Salemat that time it was appropriateto usethe
words "witch" and "guilty" in the samesentence;but we can't
say what these words mean.
This is the difficulty we encounteredin Winch's self-contra-
dictory attempt to extend the boundaries of one way of life to
encompassanother.Since (in Winch's words) "the point about"
our boundaries "is that they ex hypothesi exclude that other,"
there is no way to know how to extendthem or even what such an
extensioncould mean. It is for this reasonthat Winch is forced at
that stageof his argument to appealto a trans-culturala priori, the
set of "prime concepts" or "fundamental notions" of human
life. As we saw, however,this move is illegitimate. It dependson
a non-relativist understanding of "the very concept of human
life" or the "notions" of birth, death, and sexualrelations.
The double bind that Winch has createdfor himself is actually
 SOCIAI K NO W I I . I ) G I
visible in the statcmentof his rclativisticaccountol'truth itscll'.
"What is real and what is unreal shows itself in thc senscthat
languagehas," he says.eBut what is this "real" that.,shows
itself" in language?This way of speaking,which implies a reality
an sich, independentof language,which as it were functions as a
form of appearancefor some "real" content, is incompatible
with Winch's view that "the distinction betweenthe real and the
unreal and the concept of agreementwith reality themselvesbe-
long to" a given language.t0As Winch himself recognizes,..the
idea that men's ideas and beliefs must be checkableby reference
to somethingindependent-somereality-is an importantone. To
abandon it is to plunge straight into an extreme protagorean
relativism, with all the paradoxesthat involves." r r Thereforehe
wantsto insistboth that, within (e.g.) a religiouslanguagegame,
"God's reality is certainlyindependentof what any man may care
to think" and that "what that reality amountsto can only be seen
from the religious tradition in which the concept of God is
used . . ."t2It is evidentthat the victory over protagoreanpara-
dox is only apparent:while the individual possesses in his tradi-
tion a standardof truth, he has no standardto judge the tradition
itself. With respect to his religion, as opposedto a particular
religious perception,he is still in the position of Wittgenstein's
private language-speaker, who buys severalcopiesof the morning
paper to assurehimself that what it says is true.r3
Winch arguesthat there is no culturally neutralnotion of truth
or rationality: that is the only conclusion consistentwith his
premises,despite the attempt to avoid their relativistic conse-
quences.Sciencein particular, he claims, can lay no claim to
objectivity but is as much a matter of cultural subjectivity as
witchcraft. But what, then, is the statusof thesestatements-of
Winch's philosophy?Winch offers no reason why we should
accepthis implicit claim that philosophy,alone among modesof
thought, is not culturally relative, and that philosophersare more
equippedthan sociologistsor anthropologiststo win true knowl-
edge of social life.
Winch remains trapped within the alternativesof ,,obiectiv-
U N t) II? SIN N I) IN ( AN
; I) TXPL AN AIION
ity" and "sub-jcctivity",which arc only restatedin the alterna-
tivcs ofa transcendentally establishedscienceand rationality,on
the one hand, and relativism, on the other. Since it is the pair of
alternativesthat is the problem, neitherprovidesa solution. For
this reason,Winch's inconsistency,in trying to provide an inter-
pretationof Azande witchcraft in terms of elementsof his, Eng-
lish culture, is understandable. It is lesshis inconsistencythan the
principle he violates with it that is at fault. For if he were correct,
and the radical gulf betweencultures (as betweenthe mental and
physical pictures of the world) were so radical as to render them
unintelligibleexceptto insiders,then all communicationbetween
cultures would be impossibleor else miraculous.And, difficult
and laden with misunderstandingthough it is, such communica-
tion is clearly a reality. We are as justified in asking how, not
whether,anthropologicalknowledge is possibleas Kant was in
asking how natural scienceis possible. Winch himself knows
perfectly well that it is possible to grasp a good deal of the
workings of an alien culture, and this is why he tries to find
parallelsin his own societyto make senseof that of the Azande.
Where hasWinch gone wrong? One sourceof the problem can
be locatedin a certainpoverty of his categoriesfor the analysisof
cultural systems.Winch shareswith radically relativist sociolo-
gists of knowledgea tendencyto speakof "thought", "action",
"ideas", "language", in a fairly undifferentiatedway. This, for
instance,allows Barry Barnesto speakof scienceas just one of
numerous"belief systems." As G. Bardenhas shown, however,
even Evans-Pritchard'saccount involves a more complex and
subtledifferentiationof the conceptsof "thought" and "cultural
system" than this allows for, suggestingthat the analysis of
cultural categoriesshould lead us to speak "not generally about
mentalitiesbut about specificcongeriesof viewpoints,interests,
and patternsof experience."ra Thus, it is not only that (in Evans-
Pritchard's words) "theZande notion of witchcraft is incompati-
ble with our ways of thought," but also that
their intellectualconceptsof it are weakand they know better
what to do when attackedby it than how to explainit. Their
 S O ( - IN IKN O WIIX
I .I
rcsponseis action and not analysis.Thcrc is tttl olabtlratcltttd
consistentrepresentationof witchcraft that will accountin de-
tail for its workings . . . The Zande actualizesthese beliel.s
rather than intellectualizesthem, and their tenetsare expressed
in socially controlled behavior rather than in doctrines. Hence
the difficulty of discussing the subject of witchcraft with
Azande, for their ideas are imprisoned in action and cannot be
cited to explain andjustify action.r5
This is shown by the many casesEvans-Pritchard mentions ln
which Azande are totally undisturbed by contradictory elements
in the doctrine of witchcraft. " Azandedo not perceivethe contra-
diction as we perceive it becausethey have no theoretical interest
in the subject, and those situationsin which they expresstheir
beliefs in witchcraft do not force the problem upon them.
Azande are interested solely in the dynamics of witchcraft in
particularsituations."16As Bardendevelopsthis,
Azandefail to perceivea contradictionwhichspringsat onceto
themind of thealienEuropeaninterpreternot because theyare
deficientin logic but becausethey haveno theoreticalinterest
in the subject;they do not work out the implicationsof a
positionbecause their interestis in actionnot in theory.rT
In this way Evans-Pritcharddistinguishesbetweentheory, logi-
cally organized thinking that seeksto compare and systematize
different views or aspectsof a subject-matter,and unsystematic
thinking, which is tied to specificsituationsand actions,and does
not seek for generalization or systematizat-ion.
As Barden points out, witchcraft doctrine, though classifiedby
Evans-Pritchardas "mystical," becauseof its (in his eyes)false-
hood, falls for the Azande under the heading of "common sense"
thinking which, in contrast to science'sdependenceon experi-
ment and logic "uses experienceand rules of thumb."r8 If we
abstractfor the moment from the question of truth and falsity, we
can seethat witchcraft sharesthis feature with non-mystical com-
mon sense,both among the Azande and in our own culture. The
u N l ) i l ? sl n N l ) l N G n N t) txPL AN Al to N 57
rlescription ol' thc 'LanJc attitudc to witchcralt quoted on the
prcviouspagc fits, for instance,the attitude of most cooks, who
are unable to account for their procedures and recipes in any
systematictheoretical way. Another example is American patrio-
tism; while many Americansthink of themselvesas "patriotic",
a conceptionexpressedin a setofpracticesand attitudes,they are
unableifasked to give a coherentaccountofthis conceptionin the
context of their social activity as a whole. As Evans-Pritchard
puts it in a different context, "It is not so much a question of
primitive versuscivilized mentality [he is criticizing Levy-Bruhll
as the relation of two types of thought to each other in any
. a problem of levels of thought and experience."re
society,
We have discussedthis distinction earlier in this book, in the
form of the sort of sociological concept-formationof which
Winch approves.The contrastbetweentheoreticaleconomicsand
the conception of their own activity that workers in the stock
market may have is anothercaseof the distinction I have been
drawing here between common senseand theoretical thinking
about social life. In formulating the principles of Zande witch-
craft and magic, Evans-Pritchardwas playing the role of the
Martian investigator,in his words,
I hope I am not expectedto point out that the Zandecannot
analyzehis doctrinesasI havedonefor him. It is no usesaying
to aZande,"Now tell me whatyou Azandethinkaboutwitch-
craft" because the subjectis too generaland indeterminate,
bothtoo vagueandtoo immense,to bedescribedconcisely. But
it is possibleto extractthe principlesof their thoughtfrom
dozensof situations in which witchcraftis calleduponto ex-
plain happenings and from dozensof othersituationsin which
failureis attributedto someothercause.20
Here, it is to be noticed, Evans-Pritchardhas not oversteppedthe
boundsof legitimacy set by Winch for social analysis.But once
the variouspiecesof the witchcraft conceptionhave been assem-
bled into a system, even if the categoriesof this systemare, in
accordancewith Winch's criterion, logically relatedto the cate-
58 SOCIAT KNO W TI I ) G I
goriesof the culture itself, thc way is opcn to thc transccndcncc
ol
those bounds. Barden makes the point very elegantly:
The way in whichcommonsensedefinestheworld illuminates
the notion of world view, world picture, etc. For the world
definedby commonsenseis notnecessarily formulatedin other
termsthanrecurrentactions.It is not necessarily objectifiedin
words;neitheris it reflected upon.It is lived.It maybe reflect-
ed on subsequently, but first it is lived. The reflectionmay be
performedby membersof the communityor by outsiders,but,
whateverthe case,reflectionintroducesthepossibilityof radi-
cal critiqueand it is at this point that the distinctionbetween
truth anderror becomesrelevant.For reflectiontakesits stand
at theviewpointof scienceor theoryto enquireinto the nature
of commonsenseboth as contentand as Derformance.2r
Why is it that "reflection introducesthe possibilityof radical
critique"? By "reflection" Barden means the evaluationof a
systemof conceptsfrom the viewpoint of science.Here one can
imagineWinch objectingthat we are once again witnessingsim-
ple cultural imperialism, with "modern" culture set up as a
judge on all others. But, as Barden points out, the contrastbe-
tween Azande witchcraft and science "is not a contrast between
cultures but a contrast between patterns of experienceor modes
of meaning."22 What this meanscan be broughtout by a look at a
list, given by Evans-Pritchard,of factors supporting the Azan-
de's belief in witchcraft. These factors fall into three categories.
First, thereis a group which might be called sociological,suchas
"political authority supports vengeancemagic." Second is a
group of attributescharacterizingalmostany intellectualsystem,
such as the mutual support renderedby the different parts of the
systemand the generalfact that "aZande is born into a culture
with readymadepatterns of belief which have the weight of tradi-
tion behind them."
Both of these categoriesof factors hold for scienceas well as
for Azande belief. It is this which sives force to assertionssuch as
this of Barnes. that
U N I) II]SIAN I) IN GN N D TXPL AN AIION 59
naturalscicr.tcc
shouldpossess no specialstatusin sociological
thcory,and its beliefsshouldceaseto providereferencestan-
dardsin the studyof ideologyor primitive thought.The soci-
ology of scienceis no more than a typical specialfield within
the sociologyof culturegenerally.23
However, the picture changeswhen we look at the third group of
factors. For instance, Evans-Pritchard tells us, "Azande often
observe that a [magical] medicine is unsuccessful, but they
do not generalizetheir observations." "People do not pool their
ritual experiences" for purposes of systematic evaluative com-
parison. "Contradictionsbetweentheir beliefsare not noticedby
Azande becausethe beliefs are not all presentat the sametime but
function in different situations." "The place occupied by the
more important medicines in a sequenceof eventsprotects them
from exposureas frauds. Magic is made againstunknown witch-
es, adulterers,and thieves. On the death of a man the poison
oracledetermineswhetherhe died as a victim to the magic. If the
oracles were first consulted to discover the criminal, and then
magic were made against him, the magic would soon be seen
to be unsuccessful. "24 As a resultthe formal propertiesconstitut-
ing logical and methodological flaws are not apparent to the
Azande.
This observationgives an ironic resonanceto Winch's insis-
tencethat "ideas cannotbe torn out of their context" since "the
idea gets its sensefrom the role it plays in the system." This
position appearsplausible only so long as we forget that concepts
can play different roles in systems, and that there are different
sorts of systemsof categories.A given idea, that is, can be
situated in more than one context. Thus even if we set aside the
question of the truth or falsity of Azande witchcraft beliefs in
order to ask merely how they are related to other elements of
Azande life, we will be drawn back to the matter of truth in order
to answer this question. Skepticism is not presupposedby such an
inquiry, for one can certainly attempt a sociological study of the
rise of scientific theorizing in the last few hundred years without
believing that science is all mystification. The fact that natural
60 SOCIAT K NO W II I ) ( ; F
sciencehas made possiblean expansionol' hunranbcings' con-
scious interaction with nature, thanks to its ability to provide
reliable information about the world, will, however,play a role in
the explanation of the relation of, say, physics to the rest of
modern social life. Similarly, the unsystematic,illogical, unsci-
entific aspectof Azande witchcraft doctrine, its unsuitability for
providing such information and in fact its failure to do so, are
data for the investigation of "the role it plays in the system." The
difference between such different modes of thought is not gain-
saidby the recognitionthat concept-formationin both scienceand
magic is not a matter of transcribing into ideas data supplied by
some conceptuallyneutral experience,but one involving individ-
ual and social imagination. In Barden's words,
the constructs of imagination are not eithertrueor false;they
are createdimagesand as such go beyondsenseexperience
. . . But the imaginativecomplex,when it permeates a peo-
ple's entire world, is necessarily linked with the operative
desirefor truth and restson the popularassumptionthat it is
true. The truth or falsityof the positionis importantprecisely
because of thc implicit procedureandnot to be ableto makethe
distinctionis to fail ultimatelyto riseabovedescription. This
impliesthatthescientificgraspof anotherpositionis not iden-
tical with an understanding of the positionavailableto the
original actors.It is to know how they know, and what they
know,but not as they know.25
The problem we have been examining here has been discussed
by J.L. Mackie under the rubric of "ideological explanation."
He defines an ideology, in a manner clearly derived from Marx,
AS
a systemofconstructs, beliefs, and valueswhich is characteris-
tic of somesocialclass(or perhapsof some other socialgroup,
perhaps even of a whole society) and in terms of which the
membersof that class(etc.) seeand understandtheir position in
and relation to their social environment and the world as a
whole, and explain, evaluate,and justify their actions, and
especiallythe activitiesand policiescharacteristicoftheir class
AN I) tXI'IN N ATION
U N I) TIISIN N D IN G 6I
(ctc.). . Al leastsonrcof the bcliel.sand concepts in the
systell)arefalsc,distorted,or slanted,andat leastsomeof the
activitiessustained and guidedby that ideologyhavea real
tunctiondiffering from that which, in the ideology,they are
seenas having.26
Zande witchcraft can be consideredan ideology under this defini-
tion, which has obvious similaritiesto Evans-Pritchard'sdefini-
tion of "mystical" belief. It might be thought preferable to omit
the referenceto a putative "real function" of ideology, given the
difficulties long recognized in functionalist theory. I have left it
in, because it indicates the need for some explanation of the
maintenanceof false beliefs, which must play a role in social life
different from the one they are thought to play by people who
believe them to be true.
In terms of his definition Mackie distinguishestwo sensesof
"ideological explanation": (l) "an explanationgiven within and
in terms of an ideology by some of those whose ideology it is"
and (2) an explanation, given by someonewho does not sharethe
ideology, either of the ideology itself or of actions as resulting
from people's holding the ideology. In short, the distinction is
that betweenwhat we have called the insider's and the outsider's
views of a culture; in Harris's terminology we could refer to emic
and etic explanation.I will illustratethesemodesof explanation
with a few aspectsof Zande witchcraft for which explanations
rnight be asked.
1. Why, among the Zande, is a suspectedwitch often ap-
proached in a conciliatory way rather than with open anger?
2. Why are men not accusedof bewitching their wives?
3. Why do the Zande believe in the phenomenon of witch-
craft?
Ad 1. The Zande answer to the first question is that since
witchcraft is activated by ill-feeling on the part of the person in
whose body witchcraft-substanceresides, the last thing a victim
ought to do is provoke more ill-feeling. An anthropologist,noting
this in his ethnography of the custom, would probably give the
sameexplanation. The questionconcernsthe inner articulation of
the concept of witchcraft, and its answer must be stated in terms
62 SOCINL KNOWLTDGT
of that concept, which is the samc lirr outsidcr and insidcr
alike.
Ad2. Zandeexplainthat men don't practicewitchcraftagainst
their wives because "no man would do such a thing as no one
wishes to kill his wife or cause her sicknesssince he would
himself be the chief loser."27 The anthropologistM.D. Mcleod,
in contrast,explainsthis fact as a caseofthe generalprinciplethat
accusationsof witchcraft are "confined to oracle owners" in a
situationin which "control of women and controlof oracles[are]
in the handsof roughly the samepeople."28Women, that is, are
in the first place not in a position to make authoritativeaccusa-
tions against their husbands.Secondly, Mcleod argues that the
practiceof witchcraftaccusationas a whole is part of a systemof
competitionbetweenmen and kin-groupsof men for socialpow-
er, which is to a large extentembodiedin power over women. The
first part of this explanationdoes not contradict the native's
explanation;so far it seemscompatiblewith Winch's criterion of
propriety in anthropologicalaccounts.But the secondpart of the
explanationshows that it is not really in accord with Winch,
becauseit involves an accountof the witchcraft systemas a whole
which, if comprehensibleto a Zande, would be unacceptableto
him or her, sinceby explainingthe distributionof accusationsas a
function of socialpower relationsit conflictswith nativeexplana-
tion in terms of the empirical reality of witchcraft. Implicit in
Mcleod's discussionis the assumptionthat the correlation of
power relations and witchcraft beliefs is not coincidental. and that
in some manner,to be determined by anthropological theory, the
former explains the latter.
Ad 3. For the Zande the answerto the third question is simple:
they believe in witchcraft becausethe evidencefor its reality is all
around them. Evans-Pritchard'sexplanationhas already been
quoted in part:
Witches,as the Azandeconceivethem, clearly cannotexist.
None the less,the conceptof witchcraftprovidesthem with a
naturalphilosophyby which the relationsbetweenmen and
unfortunateeventsare explainedand a readyand stereotyped
UNI)IIISIANI)INCJ,
AND IXPLANATION
nlcansol' rcactingto suchcvcnts.Witchcratiideasalsoem-
bracea systernof valueswhich regulatehumanconduct.2e
For cxamplc, Evans-Pritchardanalyzesthe explanation of death
as causedby witchcraft by pointing out that "death is not only a
natural fact but also a social fact. It is not simply that the heart
ceasesto beat . . . but it is also the destruction of the member of
a family and kin, of a community and tribe. . . . Among the
causesof death, witchcraft is the only one that has significancefor
social behavior. The attribution of misfortune to witchcraft
. . . gives to social eventstheir moral value."30
As with the previous example, this explanation depends on
implicit reference to an anthropological theory-to a particular
conception of culture and of the role and mechanismof symbol-
ization in social life. Whatever we may think of this theory, this
example servesto make clear the logic of the outsider's explana-
tion. Given that the native's own explanation("Witches exist")
cannotbe true, how are we to explainhis beliel'/ And it is neces-
sarilyexplainedin terms foreign to the native'sview of the world,
in terms derived from the culture, general and scientific, of the
observer. The meaning of witchcraft-belief which is being expli-
cated is not the meaning for Azande, but the meaning for the
anthropologist. After all, the explanationhere provided isn't one
that the native even feels any needfor. The meaninghere assigned
"witchcraft" is not the semantic meaning of the term within
Zande culture, but spells out the relation between the idea of
witchcraft as a social fact and other facts about Zande society.
Ideological explanation of the second type appears, in
Mackie's words,
radicallyfaulty becauseit explainssomethingotherthanwhat
it purportsto explain.It purportsto explainactionsand social
behavior:but the ideologyofthe relevantclass(society,etc.) is
a constituentof thoseactionsandbehavior-theyareessential-
ly howevertheagentsseethemto be-so anyexplanationfrom
outside,which treatsthe ideologymerelyas a fact, causally
relatedto actionsand behaviordescribedin otherterms, will
invariablyneglectsomething essential.3r
 SOCINI KNOWLTDGT:
As I havealrcadysuggested, the error in this linc ol'rcasoninglics
in the slide from the fact that ideology is a constituentof actions
and behavior to the assertionthat the latter are therefore "essen-
tially however the agentsseethem to be. " And as we saw,there is
a sensein which this is true. as a result of the fact that human life
is organized by culture. Accounts of pre-capitalist goods trans-
fers that describetheseactions as though they were market trans-
actionswill fail to be accuratesocial descriptions;ethnographyof
the Azande begins with the description of certain practices as
involved with witchcraft. But it is also true that the behavior of
social actors (and here is Mackie's correction of the view de-
scribed above)
hassomepartialcauses of whichtheyareunaware,thatit will
haveeffectsthatare not includedin their purposes,and indeed
that it may havean unknownfunctionin the sensethat these
unknownpartial causesincludethe fact that suchactionstend
to producetheseeffects which were no part of the agents'
purposes.Thus there is a sensein which the agentscan be
wrong aboutwhat theyare doing.32
Azande engaged in warding off witchcraft are (apart from the
"placebo effect") not really curing illnessor ensuringsuccessful
hunting, although this is the (ideological) description they would
give of their activity. The way is at leastopen to ask what elsethey
might be doing, without being awareof it, and this questioneven
must be posed once we recognize that "witches, as the Azande
conceivethem, clearly cannot exist." In any case,
a full descriptionof whatis goingon mustandcantakeaccount
ofboth aspects,both ofhow their actionsappearto the agents
andof whatis notapparentto them:it canrecognizedistortions
as part of what there is without itself becomingdistorted.33
Mackie's account takes us far beyond the bounds set by
Winch's criterion of faithful interpretation, becauseit placescon-
cepts clearly in their real context, that of action. This is a crucial
tJN l ) tl ? SIn N t) l N ( n
' N t) t:Xt'tAN AIION 65
point: that action is a contcxt,and not an equivalent,for concepts
and language;or ratherthat action is a more generalterm, cover-
ing thought and languageas well as other forms of action. For the
shallownessof Winch's theory derives in part from his simple
identification of conceptsand behavior. For Winch, actions em-
body concepts;if they speaklouder than words it is becausethey
are like words, forms for conceptualcontents. The context from
which ideascannotbe torn is for Winch a "system of ideas." For
instance.Winch tells us that "social relations betweenmen exist
only in and through their ideas" or again that "our languageand
our social relations are just two different sidesof the same coin.
To give an account of the meaning of a word is to describe the
social intercourseinto which it enters."3aThis is why "the rela-
tion betweenidea and context is an internal one," to be under-
stood by conceptualanalysis.3sUnderstandinga society is, on
this view, nothing other than grasping the meaning of ideas in
their systematicinterconnection.
To take a concreteexample, "war" for Winch is not a concept
devisedto signify a type of social behavior; rather it is the use of
this concept which makes the behavior the kind of behavior it is.
These are not, however, the only alternatives. True, war cannot
be characterizedas a kind of "behavior" but only as a cultural
institution. But it does not follow from this that the institution is
correctly understodby the natives whose institution it is, nor that
actions always correspond to ideologies, so that acts of war al-
ways have the character claimed for them by actors.
To begin with, as the Azande witchcraft examplebrings out, to
claim this is ultimately to agree with Hayek that "any.knowledge
which we may happen to possessabout the true nature of the
material thing, but which the people whose action we want to
explain do not possess,is as little relevant to the explanation of
their actions as our private disbelief in the efficacy of a magic
charm will help us to understandthe behavior of the savagewho
believesin it." But we have already seen that an unbeliever's
account of magical practices will be quite different from a be-
liever's. Whatever the linguistic reality of witchcraft, it either
doesor doesn't causecrops to fail, make for bad hunting, or cause
66 SOCTAI KNOWTil)Gt
children to die from f'evers;and dcciding this qucstionnray bc a
matter of life or deathfor the peopleinvolved.The persistenceof
belief in methodsthat don't work calls for someexplanation,even
though we normally can't expect those for whom those methods
embody basic tradition to heed that call.
There is also a further, and most important, point, which I have
already hinted at. Social intercourse is richer than language. It
comprises a vast range of objects and practices which may lie
outside the field of verbal representationor even contradict it.
Even in those areasof social action which are organized through
language, if we identify the meaning of a word or idea with the
socialbehaviorin which it plays a role, it may easilyturn out that
meaning in this sensewill differ from the meaningselicitable
from natives. How do we, then, "describe the social inter-
course" in which a concept has its place? If we do it in the
languageof the natives, we have learned nothing; if we do it in
our theoreticallanguage,we will describesocialactiondifferent-
ly from a native and perhaps in a way a native might not recog-
nize.
Winch's view assumesa doublecoherenceof world-view: both
internal coherenceand coherencein its relation to action. As
Alasdair Maclntyre has pointed out, howeveq
The criteriaimplicit in thepracticeof a societyor of a modeof
social life are not necessarilycoherent;their applicationto
problemssetwithin thatsocialmodedoesnot alwaysyield one
clear and unambiguous answer.When this is the casepeople
start questioningtheir own criteria. They try to criticize the
standards of intelligibilityandrationalitywhich theyhaveheld
hitherto.On Winch's view it is difficult to seewhat this could
mea n.3 6
This is unrealistically optimistic: actually it takes a good deal
more than conceptualambiguity or incoherenceto spur people to
criticism of their conceptual schemes.These features have per-
haps even been typical of such schemes;the main point is that,
whether or not people are questioning their own criteria, the
latter may well be open to question.
uNt ) lI t SlnNt ) r N( ,ANt) t xPt ANn
I t oN 6/
E,rncstGcllncr providcsa conrplcxcxamplcin his description
o1'the institution of the agurram of Moroccan Berber society.
This is a man who is held to be selectedby God to be a spiritual
leader,adjudicatorof inter-groupdisputes,etc. Selectionis indi-
cated by endowment with magical powers, as well as by the
combinationof generosity,prosperity,pacifism, unconcernwith
material wealth, and other traits. Winch might take exception to
Gellner's statementthat "in reality" election of an agurram is
not by God but by his fellow tribesmen, as to Evans-Pritchard's
statementof his disbelief in witchcraft. But he must admit that an
agurram cannot really have at once all the characteristicswith
which he is credited and on which his identification as an agur-
ram depends. Lack of concern with material wealth in reality
spells the end of prosperity and the possibility of generosity.Were
this internal contradiction to be recognized, the social category
would disintegrate,just as it would were the reality of non-divine
election admitted, since this would rule out the neutrality within
the tribal systemessentialto the social role of the agurram.31
This brings out anotherflaw in Winch's identificationof social
concept and social action: its neglect of the fact that "concepts
generally contain justifications of practices, and hence that one
misinterprets them grossly if one treats them simply as these
practices, and their context, in another dress."38 As Gellner
concludes,
One might sumall this up by sayingthat nothingis more false
thantheclaimthat,for a givenassertion,itsuseis itsmeaning.
On the contrary,its usemay dependon its lack of meaning,its
ambiguity,its possession of wholly differentand incompatible
meaningsin quitedifferentcontexts,and on thefact that,at the
sametime, it as it were emitsthe impressionof possessing a
consistent meaningthroughout-onretaining,for instance,the
aura of a justificationvalid only in one contextwhen usedin
quite another.3e
For example, the social use of the conceptof "nobility" in West-
ern Europe dependedon the ambiguity betweenthe identification
of certain highly-valued behavior traits-courage. generosity,
 SOCIAL KNO W ITt ) G I
loyalty,etc.-and the identificationof a ccrtain socialposition,
which allowed people in that social position ipsofacto to lay claim
to the traits. As we will seein the next chapter,Marx tries to bring
out similar featuresof the ideology of economics, attempting to
show the role of the essential ambiguity of key concepts in the
continuing existenceof the current social system. (Thus "capi-
tal" for Ricardo-as for subsequenteconomic theory-signifies
at once raw materials, tools, and other meansof production, and
money invested with the aim of realizing a profit.)
To say all this is, in a way, to restatethe position of Winch with
which we began this chapter, only to draw from it consequences
very different from his. Winch is quite right to emphasizethat
scientific understandingof a culture involves the interpretationof
the meaning of its elements in terms other than those of the
culture itself. It is in his attempt to prove the impossibility of such
interpretation that Winch involves himself in his uncomfortable
relativism.
Winch is, I think, correct in contrastingtranslationwith under-
standing.Gellner is like many in describingthe anthropologist's
problem as one of translation:
The situationfacing a social anthropologistwho wishesto
interpreta concept,assertion,or doctrinein an alienculture,is
basicallysimple.He is, say,facedwith an assertionS in the
local language.He hasat his disposalthelargeor infinitesetof
possible in hisown language.
sentences His taskis to locatethe
nearestequivalentor equivalentsof S his own language.a0
in
Now, clearly, this task cannot be accomplishedwithout an under-
standing of enough of the culture to pick a justifiable equivalent
for,Sin the home language.But-and this is the dilemma Winch's
argumentbrings out-understanding seemsto require translation
just as translation requires understanding.With this dilemma
Winch is stuck: the cultural relativity of concepts spells their
incommensurability.
The origin of this problem lies in construal of the problem of
interpretation as one of preservationof meaning in the first place.
Construed in these terms, the problem is insoluble, since an
I'N I) II? SIN N I) IN ( AN
; I) IXPIAN AIION 69
outsidcrcannotsharcthc insidcr'scxpcrienceof his world view
and so the rneaningof its conceptualconstituentsfor him. The
hermeneuticcircle will be unbroken.A solutionappears,howev-
er, as soon as we look at translation as both basedon and enabling
a reconstruction of the culture in its non-linguistic and non-
conceptual aspects-what I referred to above as the context of
action. This, as suggestedby the Wallacearticle cited in Chapter
2, is what goes on in real casesof radical, and non-radical,
translation.If we ask in what form translationis given,
It may seemthat the answerQuine and Davidsongive, that
translationis a recursivecorrelationof expressions with ex-
pressions,andultimatelyof sentences with sentences, is obvi-
ously correct, even tautological.There may be a senseof
"translation"in which their answeris tautological and true.
But if we meanby theform of thetranslationtheform in which
the meaningof the [Linear B] tabletsis explainedto us, their
answeris wrong. The real explanationshave the form of a
translationin theQuine-Davidson senseplusa gloss.. . . The
scholarsin this businessnever give the meaningof a tablet
simplyby a wordfor wordtranslation: thereis alwaysa gloss.
And whenyouthinkof it, in ordinarylife thereis almostalways
a glosswhenwe saywhatsomeone meantor saidor thought.4'
The gloss,of course,refersto the non-linguisticcontext;sinceit
is in the translator'slanguage,and is not part of the translated
material, it dependsupon the making of cross-cultural compari-
sons as a means to knowledge of the alien culture.
My intention here has not really beento re-enter the problem of
translation, but rather to stressmy conclusion that translation, or
more generally the preservation of meaning between cultures, is
not the prime operation of social study. The anthropologist may
even find a meaning for cultural featuresthat the natives do not
understand. To say this is not to contrast "social-conceptual"
with "physical" description. As we saw in Chapter 2, social
action is itself observable; it is open to alternative descriptions.
One may therefore dispute a social actor's account of his own
behavior.But this doesnot meanthat a social scientific analysisof
 SOCIAL KNOWLTDGI

witchcraft ritual in terms of power relationsantongthc Azandets

a descriptionof a different realm of phenomenathan the native's
own description in terms of witchcraft. This is shown by the fact
that the anthropologist's analysisemploys the native's concepts,
Foundotion
ond
although not only them. Both accounts, the insider's and the
outsider's, are descriptionsand explanationsof the sameactions.
Superstructure
Theseactions will not then have the samemeaning for the outsid-
er as for the insider, but they will not therefore become meaning-
less, becausethey will not have lost their connection with social
life as a whole. Possibly,it is only in the anthropological account
that that connection will be revealed; in any case, it will be
spelled out in terms different from thoseof the natives. This does
not mean that the natives' experienceof their own culture be-
comes of no account. Just the opposite: it is among the phenom-
ena with which social theory must grapple. It is not by its ability
to substituteitself for a culture that a social theory is judged but Recognizing that culture provides the structure of social action, I
by its ability to explain it. gave reasonsin Chapter 2 for abandoningthe use of "subjectiv-
ity" to single out the character of social facts, becauseof the
misleading natureof the contrastwith "objectivity. " Similarly, in
the last chapter I argued for dropping "meaning" as referens of
the social scientist's object of study. The paradoxesof Weberian
sociology and its variants, as well as of Winch's philosophical
approach, turn out to arise at least in part from the difficulty of
reconciling the meanings (for insiders) of cultural elementswith
the demand for (outsider's) scientific truth-an opposition again
conceivedof as one of "subjectivity" to "the objective."
An excellent epitome of the position I have been opposing is
provided by a recent historical article on European witchcraft.
Basing himself on Winch's work, Stuart Clark argues that
Renaissancedemonology is to be understoodin terms of the total
world-view within which it had its place, and that

part of whatwe meanwhenwe speakof a "world-view" at all

is surelythat its constituentsneedno other explanationthan
their coherence with oneanother.Theprimarycharacteristicof
demonolosicaltexts as historicalevidenceis not their

71
/2 SOCIAL KNO W I TDG T
s uppos ed[ ] u n v e ri fi a b i l i tyb u t th c i r re l a ti onshi pto w hat J.L.
Austin called a "total speechsituation": their mcaningfirr the
historian may be thought of as exactly symmetrical with thcir
original meaning as linguistic performances.r
Clark admits the cost of this preservation-through-reconstruc-
tion of meaning: "A contextualreading of Renaissancedemonol-
ogy may not help us answer the major questionsabout the genesis
or decline of the European'witch craze' . . ."2 But presumably
what historians, along with anthropologistsand social theorists
generally, are after is the explanation of social phenomena, in-
cluding those of meaning. Ideological explanation in the first of
Mackie's two senses-explanationsgiven in the terms of a given
ideology-is not enough; we want also explanation in his second
sense,explanationo/the ideology. Exactly becausethe conceptu-
al aspectof culture has a systematiccharacter,though as we saw it
is not as consistentas Winch imagines, satisfying the demand for
explanation of this type will involve relating concepts to non-
conceptualelementsof social life, sinceotherwise we will remain
within the circle of concepts.
Clark objects to "the explanation of . . . witchcraft beliefs in
terms of social and socio-psychologicaldeterminants"3on the
ground that this bypassesmeaningby reducingthesebeliefs to the
status of epiphenomenaof non-ideological social factors. If we
are to make some senseof the concept of ideological explanation
we must recognize the justice in this point of view, despite its
ultimate untenability. To begin with, we may reject explanations
of (e.g.) witchcraft that claim that witch-talk is "really" talk
about social relationships, morality, etc. Such explanations re-
main on the terrain of "meaning" and on this terrain they fail
exactly as does the similar effort to explain moral terms as "real-
ly" expressionsof feeling or descriptions of social conventions.
"Murder is wrong" does not mean "l don't like murder" and
"This is the work of witches" spokenamong the Azande does not
mean "There is bad will at work in this or that social relation-
ship." As Martin Hollis expressesthis,
I O UNI ) ATI OAND
N SUPTI I SI I ?UCI UI]?E
3
A Zandc. who bclicvcshc has beenbewitched,surelydoesnot
believc that he has of-fbndedsome social authority or other. If he
did, he could perfectly well say so. He surely believes, rather,
that he is the victim of supernatural interference.a
Similar objections can be made to a classicalschemaof ideologi-
cal explanation, that which operates in terms of social "func-
tion." Clifford Geertz describes this schema wittily:
A patternof behaviorshapedby a certainset of forcesturns
out, by a plausiblebut nevertheless mysteriouscoincidence, to
serveendsbut tenuously related to those forces.A group of
primitivessetsout, in all honesty,to pray for rain andendsby
strengthening its social solidarity . . . The conceptof latent
functionis usuallyinvokedto paperover this anomalousstate
of affairs, but it rather namesthe phenomenon. . than ex-
plainsit...5
This remark holds true of both the chief forms of anthropologi-
cal functionalism. To begin with, the idea that items in a culture
are to be explained by demonstrating the interactions between
them (e.g. the correspondence betweena hierarchyofgods and a
social hierarchy) does not take us very far. If we ask to what
questions such functional analysescan supply answers, we find
only questionsof the form, "Is there a pattern of interaction
betweenthe variouselementsof a culture?" This is a questionof
some interest. The origin of functionalist theory lay in the rejec-
tion of speculative historical explanations of cultural traits.
Against this functionalisminsisted,first, that traits could not be
understoodin abstractionfromtheir systemiccontext,and second'
that explanationshad to be given in terms of real information'
gathered in fieldwork, and not in the form of guesses,more or
less inspired, about (unknown) earlier periods. But this form of
explanation cannot supply answersto questions about the origin
and development of the culture as a whole.
 /4 SOCIAL KNOWLTDGT
These lirnits of this form of functionalismare avoiclcclby a
secondversion, in which social traits are to be explainedby their
causal efficacy in satisfying certain biological-psychological
needs (Malinowski) or in maintaining the existenceof the social
system (Racliffe-Brown). Aside from problems peculiar to each
of thesepatternsof explanation,both suffer from the flaw that (in
Maurice Mandelbaum's words) such "generalizations permitted
no deductive consequenceswith respect to the specific nature of
the practicesof the peopleswhich they investigated.',e
It is interesting to note that Marx made a similar point in
criticizing Feuerbach's speculative anthropological study of
Christianity. Feuerbach believed that his book, The Essenceof
Christianity, demonstratedthat the Christian idea of God repre-
sentedan hypostatrzationof the social nature of the human spe-
cies. This he argued for by showing that the attributes of God are
the attributes of humanity. The content of the argument, despite
its philosophicalform, is very much that of the symbolic interpre-
tation of culture. Against this mode of analysisMarx commented
that "it is . . . much easier to discover by analysisthe earthly
kernel of the misty creations of religion than to do the opposite,
i.e. to develop from the actual, given relations of [social] life the
forms in which thesehave been apotheosized."T To do this would
be to meet Mandelbaum's challengeto functionalism: to explain
the particular categoriesof a culture on the basisof their relation
to others of its features. How did Marx think this can be accom-
plished?
The following passage,from Marx's and Engels'searly work
againstthe "German ideology," again has Feuerbachin view:
In directcontrastto Germanphilosophywhich descends from
heavento earth, here [in their own work] it is a matter of
ascendingfrom earthto heaven.That is to say,not of setting
out from what men say, imagine,conceive. . . in order to
arriveat menin theflesh;but settingout from real,activemen,
and on the basisof their real life-processdemonstratingthe
development of the ideologicalreflexesandechoesof this life-
process.8
IOU N T) AIIONN N D SU PTITSIIIU C IU R T/5
In thc d iscussion ol'Winch wc saw t he inadequacyof f or m ing an
inragc of a social system on the basis of its natives' self-image.
However, how can we speak of "real, active men" without
speaking of what they "say, imagine, conceive"? With this we
return to the problem raised in the introductory chapter of this
essay, and to Marx's assertion, there cited, of the determination
of consciousnessby social being. To quote the core of the 1859
Preface again:
In the socialproductionof their existence,men inevitablyenter
into definite relations, which are independentof their will,
namely relationsof productionappropriateto a given stagein
the development of their material forces of production. The
totality of theserelations of production constitutesthe econom-
ic structure of society, the real foundation on which arises a
legal and political superstructureand to which corresponddefi-
nite forms of socialconsciousness. The mode of productionof
materiallife conditionsthe generalprocess of social,political,
and intellectuallife. It is not the consciousness of men that
determines their existence,but their social existencedeter-
mines their consciousness.e
It will be usefulto distinguish two, variouslygeneral,pointsat
issuein this famouspassage.The metaphorof foundationand
superstructure is used,first of all, to expressMarx's oppositionto
a particularapproachto history.Accordingto Hegel,the institu-
tions of the political stateprovidethe organizingframeworkfor
sociallife as a whole.In the traditionof Enlightenment philos-
ophyof history,Hegelheldthat the development of the statewas
to be explainedin terms of a logic of concepts-thedialectic-
exhibitedin the history of "spirit," or the world of cultural
categories. Marx's "critical reexamination of the Hegelianphi-
losophyof law" led him "to the conclusionthat neitherlegal
relationsnor political forms could be comprehended by them-
selvesor on the basisof a so-calledgeneraldevelopment of the
/6 S O Cr ntK N OW Itr)Gr
humanmind, but that on thc contrarythey originatcin thc nratcri-
al conditions of lif-e . ." The "anatonty" of thesc matcrial
conditions,Marx discovered,"has to be soughtin political econ-
omy."10
Howevel Marx entitled the work in the Preface to which the
above-quoted passagesappear, a contribution not to political
economy but to its critique. When he summarizesthe "general
conclusion" to which his economicstudiesled him. we find not
the description of a new economic theory but the assertion that
Justasonedoesnotjudgeanindividualby whathethinksabout
himself, so one cannotjudge . . . a periodof transformation
by its consciousness,but, on the contrary,this consciousness
must be explainedfrom the contradictionsof material life,
from the conflict betweenthe socialforcesof productionand
the relationsof production.rr
This suggestsan idea of greater generality than the rejection of
legalistor mentalistexplanationsof socialchange.It suggeststhat
in general the social theorist can and must distinguishbetween
"the economic conditions of production, which can be deter-
mined with the precisionof naturalscience," and the "ideologi-
cal forms" with which people represent these conditions and
their changesto themselves.l2And throughoutthe writings that
constituteMarx's critique of political economy-Ihe Contribu-
tion, the Grundrisse , Capilal , Theories of Surplus Value, and the
mountain of notesand studiesMarx left besidethem-we discov-
er that the categoriesof economics,whether in the form of econo-
mists' conceptsor as they appear in "the everyday consciousness
of the agentsof production themselves,"r3are treatedas ideo-
logical forms, to be explainedby referenceto the social relations
of production. Towards the end of the manuscript edited by Eng-
els as Volume III of Capital, for instance, Marx refers to this
ordinary consciousnessas the "religion of everyday life."ra
While what Marx called "vulgar economics"-the ancestorof
today's "neo-classical" economics-"actually does nothing
more than interpret, systematize, and turn into apologetics the
TOU N I) N IIONN N t) SU PTIi STIIU C IU I? E//
notionsol'thc agcntstrappedwithin bourgeoisrelationsof pro-
duction"r5 these notions were to an extent critiqued by those
whom Marx called the classicaleconomists,above all by David
Ricardo. But even these "remain more or less in the grip of the
world of illusion which their criticism had dissolved . . "r6 This
raises two interrelated questions: what is the relation of the cri-
tique of political economy to that science itselfl and what are
"economic relations of production" if they are not to be iden-
tified with the referents of the catesories of economic the-
ory?
In his discussion of pre-capitalist society in the Grundrisse
Marx observesthat "the fact that pre-bourgeoishistory, and each
of its phases,also has its own econom} and an economicfounda-
tion for its movement, is at bottom only the tautology that human
life has since time immemorial restedon production, and, in one
way or another, on social production, whose relations we call,
precisely,economic relations."rT "Production," here equated
with economy, is what Marx in his methodological introduction
to the Grundrisse called an abstraction, a concept which with
respect to a group of phenomena"brings out and fixes the com-
mon elementand thus savesus repetition." In so far as it does
this, sucha concept is "rational" (verstiindig). But it must not be
forgotten that such abstractionsare tools for organizing the his-
torical materials.Any particularcaseof production-of the "me-
tabolism betweenman and nature"-will be essentiallycharac-
terized by a multitude of features. Some of thesefeatureswill be
common to all systems of production, but
nevertheless, just thosethingswhich determinetheir develop-
ment, i.e., the elementswhich are not eternaland common,
must be separated valid fbr pro-
out from the determinations
duction as such so that in their unity-which arisesalready
from the identity of the subject,humanity,and the object,
nature-their essentialdifferenceis not forgotten.rs
Capitalism and feudalism both fall under the abstraction, "sys-
tem of production," but this is not to say that the properties of
 7B SOCIAI KNOWLII)GI IOUNDAIION AND SUPERSIT?UCIURE79

either systemare to be elucidatedin terms of sonrcunil'icdset of of'casesfalling under an abstractionby their differing complexes
characteristics.Each must be explained by referenceto the par- of determinationsis illustratedby his remarkson the categoryof
ticular constellationof features("determinations," in Marx's "population." This term can be takento refer to a "concrete "
vocabulary)characteristicof it as a unique historical object. i.e. the empirical subjectof social activity. But theoriesof popuiu-
Marx's "abstractions" are not "generalizations" suitable for tion, exactly becausethis abstraction is so general as to refer to
the framing of universal laws linking social facts; their function is the membersof every social system, shedno light on any particu-
rather to provide a framework for comparison of social systems lar one. "Population"
with respectto their differences. It may be useful in particular to
distinguish Marx's idea of "abstraction" from Max Weber's
is anabstraction
if I leaveout, for example,theclassesof which
"ideal type" concept. The latter is defined as an imaginary, pure
it is composed.Theseclasses in turn are an emptyphraseif I
caseof some pattern of motivation-the profit motive, for exam-
amnotfamiliarwith theelements on whichtheyrest,e.g.wage
ple, or the "Protestant Ethic." It is formed, in Weber's words, labor,capital,etc. Theselatter in turn presuppose exchange,
"by the synthesisof a great many . more or less present and divisionof labor.22
occasionally absent concrete individual phenomena, which are
arranged according to these one-sidedly emphasizedviewpoints
The "presupposition" and "resting" here, reminiscentof the
into a unified analytical construct."re Having no direct counter-
relation of foundation to superstructure,are certainly not meant
part in reality, they are "primarily analytical instrumentsfor the
ontologically;the relationsMarx has in mind are explanatory.As
intellectualmasteryof empirical data."20 Marx's abstractions,in
he later says, each involvesthe others: "the simplesteconomic
contrast, are intended to refer to elements of social reality, al-
category,e.g. exchangevalue,presupposes population,moreover
though their construction may involve processesof idealization.
a populationproducingin specifiedrelations;as well as a certain
With respect to these concepts Marx is a realist as against We-
kind of family, or commune,or state,etc. It can neverexist other
ber's instrumentalism. "Capitalism," for instance, refers for
than as an abstract,one-sidedrelation within an already given,
Marx to an actual social structure. At any given moment (for
concrete,living whole."23 Ary human population,that is, exists
example, at the time when Marx developedthe concept!) there
ordered by specific social relations. To understanda particular
may be no social systemwhich as a whole instantiatesit. But this
society, therefore, the system-specificdefinitions of these rela-
term refers to a developing system: the condition for its applica-
tions must be studied.Furthermore,in sucha study,the meanings
tion is the existenceof certain social relations which are develop-
of the terms for theserelations are given by their interrelation as
ing over time. The concept of "capital" in fact assumesthe
elementsof the whole system.
existenceof non-capitalistor pre-capitalist social relations, since
The same holds true, in Marx's eyes, for "production": the
it is an historical concept; therefore it takes account in advanceof
identity of all systemsof production, or their elements-labor
the fact that in any real casecapitalist relations will be in interac-
and means of production-is proved by economists only
tion with non-capitalistrelations. All the category requires for its
application is that the capitalist elementswill dominate, either at
by holding fast to the featurescommonto all processesof
the time or over a time.2r
production,while neglectingtheir specificdifferentiae.The
What Marx has in mind when he speaksof the differentiation identityis demonstratedby abstractingfrom thedistinctions.2a
BO SOCIAL KNOWLIDGI
Marx makes the same argunlcnt fbr tcrtns dcnoting particular
socialinstitutions,when theseinstitutionsare common to scveral
societies.Although money, for instance,plays a crucial role in
capitalism, it also existed "before capital existed, before banks
existed, before wage labor existed, etc."2s Money is analyzedby
Marx as a commodity which plays the role of the form by which
exchangevalue is represented;this allows it to fulfill the func-
tions of (to take two) the medium of exchangeand the store of
value. Both of thesefunctions are to be discovered in pre-capital-
ist systems,but with a difference. In capitalism,money has the
role of the "universal equivalent": all goodswith a central role in
the reproductionof social life, including the basic good, labor
power or the ability to produce, can be and are representedby
sums of money. This is not true in pre-capitalist systems,where
most goods are not produced for sale and where money does not
mediate the social circulation of goods. Although both are cases
of money, an explanation of money that fits one casemay not be
very helpful in answering questions about the other.
We thus encounterin Marx's critique of political economy a
core elementof Winch's critique of social science:the argument
for the fallacious nature of equating features of different soci-
eties, made in terms of the systematic nature of social institu-
tions. The unbridgeable gap between the two critiques would
seemto be Marx's apparentdisregard,in his desirefor a scientif-
ic analysisof society, of the consciousnessof the individuals
constituting it.
On the other hand, we must note that Marx defines labor, the
active factor in production, as "purposive activity aimed at the
production of use-values."26Productive activity is of course
itself regulatedby consciousness; Marx stressesthat his discus-
sion of the labor process in Cnpital presupposes
labor in a form in which it is an exclusivelyhumancharacteris-
tic. . . [W]hat distinguishes the worstarchitectfrom the best
of beesis thatthe architectbuildsthecell in his mind beforehe
constructsit in wax. At theendof everylaborprocess,a result
emergeswhichhadalreadybeenconceivedby theworkerat the
beginning,hencealreadyexistedideally.Man notonly effectsa
chanseof form in the materialsof nature;he alsorealizeshis
I O UNT) NI I ONND
N SUPTI ?SI I ?UCI UIBI
?E
()wn pulposc in t hc r nat cr ials.And t his is a pur posc he is
consciousof, it determinesthe mode of his activity with the
rigidity of a law, and he must subordinatehis will to it.27
Furthermore. Marx stressesthat the aims of labor are socio-
culturally and not naturally defined: "Hunger is hunger, but the
hunger gratified by cooked meat eatenwith a knife and fork is a
different hunger from that which bolts down raw meat with the
aid of hand, nail, and tooth."28 "Food" itself, to develop this
example, is too abstractto serve as a generalname for the product
of the relevantbranchesof production, while to treat it as mean-
ing "satisfier of biological need" is exactlyto ignore the social
determinationof the humankind-naturemetabolism."Food" oe-
notes a use-value,it is true: that defined by the satisfactionof
hunger.But in capitalism,accordingto Marx, use-valuesare the
aim of productiononly in so far as they are commodities,objects
saleableat a profit on the market. Food which cannotbe sold will
be destroyed.It is therefore produced not as foodper se but as the
commodity food, as a "unity of use-valueand exchange-value. "
The generalgoal of production in capitalismis what Marx calls
"valorization," the expansionin value of investedcapital. Thus
the labor processwhich abstractlyappearsas "an appropriation
of what exists in nature for the requirements of man, . . . the
universal condition for the metabolic interaction between man
and nature"2e in capitalism "is only the means where by the
valorization processis implemented."30
The distinction between abstraction and concrete example is
twofold. On the one hand, since "the requirementsof man" are
in part culturally as well as biologically defined, "labor" in
capitalism includes practices of no directly biological value-
castinghoroscopes,for example,or adjustinginsuranceclaims.
On the other hand, many activities, suchas brushingone's teeth,
or even cooking one's own dinner, do not fall under the definition
of labor relevant to the capitalist market system. In so far as the
two-humankind-nature metabolism and valorization process-
do coincide, capitalism and the physical requirementsof human
life are compatible; but if and when they don't, either the system
or life must go, since "production" at any time refers only to one
82 SOCTA| K NO W |t t ) G t
particular instanceof it or another.
Marx constructs a similar argument fbr the category of
"capital," persistentlyconfusedby economistswith tools and
materials in general. "Whatever the social form of production, "
Marx writes in Volume II of Capital,
workersand meansof productionalwaysremainits factors.
. . For anyproductionto takeplace,theymustbe connected.
The particular form and mode in which this connectionis
effectedis what distinguishesthe variouseconomicepochsof
the socialstructure.In the presentcase,the separationof the
free workerfrom his meansof productionis the given starting
point,andwe haveseenhow andunderwhatconditionsthetwo
cometo be unitedin the handsof the capitalist-i.e.,as his
capitalin its productivemodeof existence.3r
Thus, to take the classicaleconomists'conceptof "fixed capi-
tal," "means of labor are fixed capitalonly where the production
processis in fact a capitalist processand the meansof production
are thusactuallycapital,i.e. possessthe economicdetern-rination,
the social character of capital; secondly, they are fixed capital
only if they transfer their value to the product in a particular
way."32 The abstractdefinition of "means of production" as
"objects usedin the processofproducing other objects" doesnot
hold true for "fixed capital," in the strict sensethat being such an
object is necessarybut not sufficient for such an object's being
capital. A tool can only function as capital if its usemakespossi-
ble the exploitation of labor power and the eventualgenerationof
a sum of money exceeding that invested in the production pro-
cess. Thus the abstraction,"means of production," applies to
capital only in so far as "production" is here taken in its special
form as the production of capital.
But such definitions, which Marx is opposingto those of the
economists,fundamentally involve referenceto cultural categor-
ies: as Marx puts it in the passagejust cited, "capital" (for
instance)is a social characteror determination. "Capital" is
defined as self-expanding"value," and "value" in terms of the
socially necessarylabor time needed to produce commodities.
IOU N T) AIIONAN t) SU PFIISIR U C IU IIT 8 3
This labor tinrc appcarsto ref'er to a natural phenomenon,the
hours of actual work put into producing a given object, and is
generallytakento do so by economistscommentingon Marx, but
it does not. Marx makes quite a point of distinguishing between
labor as use-value-creating,as natural phenomenon,and labor as
value-creating,as socially-defined "abstract labor." The latter is
defined only in the context of the sale of goods on the market.33
' ' Commodity, ' ' once it becomesthe general form of goods, turns
out not to be disentangleablefrom the category of "capital,"
while "exchange" itself requires-as noted in the first chapter of
this book-reference to a social mode of behavior in which people
recognizea categoryof "ownership" and "behave in such a way
that each does not appropriate the commodity of the other, and
alienate his own, except through an act to which both parties
consent.' ' 34
The conceptof "capital" is thus ideological in the sameway as
the conceptof "nobility" discussedin the previouschapter:by
conflating the specific, historical form of the meansof labor with
"means of labor" as a trans-culturalabstraction,the particular
institutions of capitalist society are presentedas universal, as
present,even if in different disguises,in all societies.We will
return to this examplebelow. Here I want to point out that Winch
is wrong to assimilateMarx's concept of ideology to Pareto's
"derivations." This is becauseMarx's abstractionsare not "resi-
dues," real contentshidden by ideologicalappearances. For Pa-
reto, the various casesthat representthe same residue are equiv-
alent to each other, except for their superficial appearances.This
is exactly what Marx wishes to deny in the case of abstractions.
The differencesthat are extinguishedin the formation of abstrac-
tions are what we must appeal to in any attempt to explain the
featuresof actualsocieties.
The materialsandmeansof labor . . . play their part in every
labor processin everyageand in all circumstances. If, there-
fore, I label them "capital" . . then I haveproved that the
existenceof capitalis an eternallaw of natureof humanpro-
duction.andthattheKirehiz who cutsdownrusheswith a knife
84 SOCTALKNOWItt)Gt
he hasstolenfrorna Russianso as to weavcthcnltogcthcrto
makea canoeisjust astruea capitalist asHerr vonRothschilcl.
I could prove with equalfacility that the Greeksand Romans
celebrated communionbecause theydrankwine andatebread,
and that the Tirrks sprinklethemselvesdaily with holy water
like Catholicsbecause they washthemselves daily.35
This argumentholds even for a caselike that of "labor, " which
unlike "capital" is an abstractionlike "production," denoting
human effort in metabolizing nature, and so is valid for all social
systems.On the other hand, it has a specific use for the analysisof
capitalism which it does not have for any other system. It is only
in this society that the idea of a "labor force," the individual
members of which "can with ease transfer from one labor to
another"36 has operationalmeaning,sincein earlier systemsthe
division of labor was more strictly regulatedby custom.In addi-
tion, all forms of labor are treatedas equivalentin capitalism,in
the sensethat their embodiments in products are equally trans-
formable into money. In contrast, when applied to pre-capitalist
systems,"labor" changesits significance.It can not function asa
category for the specific form that production may take in any
given societyeither. We must distinguish, that is, between "la-
bor" as an abstractiondenotingproductiveactivity in any society
and "labor" as a categoryfor suchactivity in a particular society,
capitalism.A betterknown example,which hasprovedequallyas
troublesome for commentators on Marx, is the concept of
"class." In describing"all previoushistory" as "the history of
classstruggles," Marx uses"class" to designatean abstraction:
social grouping defined by its control over the means of social
production. When describing the evolution of capitalism out of
pre-capitalistsociety, inthe German ldeology, however,he draws
attention to the difference between class and estateas forms of
social organization, pointing out the historical specificity of the
former to capitalism.
Marx seems indeed to have held a conception closer to
Winch's idea of a socialsciencethan the latter suspects.But here
we must remember the double use Marx makesof his metaphorof
IOU N I) AIION AN I) SU I}FI? SIR U C IU IIF8 5
Ioundationand supcrstructurc:to make not only the specificcon-
trast between processesof (consciously regulated) interaction
with nature and other activities, but also the more general con-
trast between social institutions and people's understanding of
them. To say that the processesof social life are structured by
culturally-determined categories is not-as we saw earlier-to
say that people's understanding of these processescannot go
seriouslyawry. Although it is by their belief in witchcraft that the
Azande create an object of study for the outsider anthropologist,
this does not mean that the phenomena categorized as witchcraft
and magic are correctly describedor understoodby the insiders.
As Zande witchcraft for Evans-Pritchard,so does capitalist eco-
nomics serveas the subjectmatter of Marx's study,which is thus
an example of ideological explanation in the secondof Mackie's
sensesand not the first (see p. 61 above).
In effect, Marx's work threatensa classicdistinction between
anthropology and the other analytic social disciplines. S.F. Na-
del, for example, finds no difference in principle betweenanthro-
pology and sociology, but distinctions in method and concept-
formation arising from the character of the objects of
anthropologicalinvestigation-the "strangenessof primitive cul-
tures, their separationfrom our civilization":
Sociologyaswell ashistorypresuppose a certainfamiliarityof
the observerwith the natureof the datahe studies:bothdisci-
plinesimply that certaincommonlyusedcategories are imme-
diatelyapplicable,andthat the probabilitiesandconsequences
of actionsare eitherreadilyunderstoodor deduciblewith self-
evidencefrom the generalbackgroundof experience.New
items . . or new problems. . . fall into place againstthis
backgroundof an experience whichis commonto observerand
observed.This is not true of primitive society,which is unfa-
miliar andoftenstrange. . . Socialcategories mustbe evolved
ad hoc; they cannotsimply be takenfor granted.3T
As Nadel observes, such an approach need not be limited to
strangecultures; one may go further to ask how it could be, if the
86 S O Ct nt K N O W ttt)c t
goal is somethingwc would call a "scicntil'ic" unclcrstanding ol'
social life. The contrastwith the anthropologicalpoint of vicw
can be brought out if we look back to the Martian whom Fritz
Machlup had investigating the stock market (see Chapter 2
above).In Nadel's terms, Machlup's Martian is not doing anthro-
pology but economics. He is trying to explain the functioning of
the stock market within the context of the economic and social
systemas a whole, and indeed the concept-systemand culture of
capitalismitself. This is implicit in Machlup's statementof the
Martian's goal as "to know the economicfunctionsof the stock
market, particularly its role in the utilization of investiblefunds
and in the formation of capital.' '38Sucha goal is imaginableonly
if there is capital investment,and all thatgoes with it, on Mars!
Machlup's questionsare anthropologicalones only as the de-
mand for ideological explanationof Mackie's first sort is, as
explorations of the inner articulation of the culture studied.
Machlup imagines these to be anthropologicalquestionsof a
deeper sort perhaps becausehe is himself an economist, for
whom economictheory consistsof a systemof " 'universallaws'
or fundamentalhypotheses" involving "constructs of idealized
human action basedon [assumed]objectivesto maximize profits
and satisfactions."3eIn line with this Weberiandefinition of his
field, Machlup seesthe theorizing necessaryfor understanding
the market as consisting "mainly in constructing ideal types of
motivatedconduct of idealizeddecision-makers,and examining
them in abstractmodelsof interactions." So, for example,busi-
nesscompetitionis to be explainedin terms of a "profit motive."
Of course,thesemotivationscan only exist in a social systemin
which the requisite categoriesof action-market exchange,cap-
ital, profit-exist; Machlup focuses on the goal-defined ideal
types ofaction becausehe takes the social categoriesfor granted.
His anthropologist from Mars is really an economist from Earth
in a spacesuit.ao
Were he actually to make the effort, in Nadel's words, "to treat
a familiar culture as though it were a strange one," the anthro-
pologist would go beyond the analysisof a given institution or
IOU N I) AIION N N D SU PTI? SIR U C IU I? EB]
phcnonrcnon in tcrrnsol'capitalism'sown categories("profit,"
"nrarket," "capital," etc.) to an explicationof thesecategories
themselves.He would in this processbe forced to evolve new
categories "ad hoc," in the attempt to develop a theory about
those structuredactivities that in contemporary society we insid-
ers think of as "the economy," a theory which might ultimately
provide an explanation and analysis of this classification itself.
Thus an anthropologist would be led to ask why instruments of
production should be construedunder the rubric of "capital,"
i.e. as equivalentto certain sums of money,just as Evans-Prit-
chard wonderedabout the equivalenceof birds and people among
the Nuer and Willett about the Yoruba'sidentificationof cowrie-
shell boxes with their headsor souls.
For the economic theorist the most abstract categories are
those of his own culture; thus economic anthropologistshave
struggled-without notable success-to impose the concepts of
neoclassical economicson their field studiesof tribal societies.In
reality, in other cultures (for instancethat of medieval Europe),
the category of "capital"-and therefore a concept like "profit-
maximization"-plays no real part at all, any more than the
concept of inherited witchcraft plays a role in the culture of
capitalism.The anthropologist'stask (as statedconciselyby Ju-
lian Pitt-Rivers), if he is not "to accept the 'natives' view' of
their own world as if it bore the credentialsof science." is "to re-
order that which they order and discoverbehind their ordering a
more abstractorder that explainsit. "4r It is the inability to do
this-to adopt the outsider'spoint of view in the study of their
own culture-which typically mars the curious attempts of an-
thropologists to apply their techniquesto capitalist society. Al-
most alwayssuchstudiesexaminean area,a neighborhood,or an
occupationalgrouping.a2But if the classictopicsof anthropology
are such as the Nuer, the Mbuti, the Nayars, or even "tribal
society" globally, then the equivalent frame of reference for
studying modern society should also be the whole culture-the
capitalist world system. For this reasonone may speakof Marx's
work as anthropological in approach, rather than as sociological
BB SOCIAI KNO W I . I I ) G I
or economic(althoughit would bc rnislcadingto spcakol'Marx as
an anthropologist,since for him scientific understandingof cap-
italism was to be only a means to its revolutionary abolition).
As a matter of fact, Marx contrastshis theoretical procedure
with that of the economists in the same terms as those we have
used to contrastanthropologicalinvestigation with the researches
of Machlup's pseudo-Martian.The starting point of Marx's re-
searchprocedure in Capital is an attempt to seethe categoriesof
his own culture as foreign and strange. "A commodity appearsat
first sight an extremely obvious, trivial thing. But its analysis
brings out that it is a very strangething, abounding in metaphysi-
cal noveltiesand theologicalniceties."a3 The languageof "meta-
physics" and "theology," and above all the dominant metaphor
of "fetishism" is not accidentalor purely literary: Marx meansto
suggest that the analysis of economic ideology is a case of the
same procedure as is representedby the analysisof religious or
philosophical ideology. The "strangeness" of the commodity is
not noticed by the economistswho study it. This is because,in
Marx's words,
Reflectionon the forms of humanlife. hencealso scientific
analysisof thoseforms, takesa coursedirectlyoppositeto their
realdevelopment. Reflectionbeginspostfestum,andtherefore
with the resultsof the processof development alreadyto hand.
The forms which stampproductsas commoditiesalreadypos-
sessthe fixed qualityof naturalformsof sociallife beforeman
seeksto give an account,not oftheir historicalcharacter,
for in
his eyesthey are immutable,but of their contentand mean-
ing...
The categoriesof bourgeoiseconomicsconsistpreciselyof
formsof thiskind. Theyareforms of thoughtwhicharesocial-
ly valid andthereforeobjective[are "social facts," one might
saylfor therelationsofproductionbelongingto thehistorically
determinedmode of production,i.e., to commodityproduc-
tion. The whole mysteryof commodities. . vanishesthere-
fore as soonas we cometo other forms of productionaa
-i.e., as soon as we put them into a comparativeperspective.
IOU N D AIION N N D SU PII? SIIIU C IU I? I 8 9
This is ol'coursconc of thc clichesol'anthropology,that cross-
cultural comparisondemonstratesthe particularity of what ap-
pears to the cultural insider to be universal. But, as Winch ob-
jects, how can we compare that which is culturally specific? To
quote Pitt-Rivers again,
A comparisonof socialstructuresmustbe concernedwith the
elementsof socialstructure:hierarchy,authority,power,sanc-
tity, the division of labor, the definition and solidarity of
groups,thetransmission of property,the rulesof descent,etc.,
not with approximations to a stereotype,that is to say,a par-
ticularconjunctionof featureswhichmay frequentlybe associ-
atedin factbut haveno demonstrated necessary relationshipto
one another.a5
Here lies the role of the Marxian abstraction:the subordination
of "capital" (e.9.) to the complexof abstractions"tool use" and
"exploitation" (to take the most important elements of Marx's
explication of this institution) exposesthe particularities of the
capitalist mode of production. One can even say that it is, as
Winch recognizesin his attack on transculturalanalysis,the avail-
ability of comparison that defines the outsider's point of view (as
opposed to being an actual traveller from a different society).
How can this be? How, that is, could anyoneand how can Marx
in particular be said to have had an outsider's point of view on
what was after all his own social system? The answer for the
general case lies in the fact that modern society certainly, and
most societiesto some extent, far from being the monoliths sug-
gestedby our use of the term so far (which follows Winch's) are
complexesof different interests,or (to use the jargon expression)
subcultures.Marx wrote as a member of the socialist movement
of the second half of the nineteenthcentury, which by that time
had already developedcategoriesdefining a new form of social
life-based on the abolition of wage labor and the substitution for
it of the "free associationof producers"-incompatible with the
institutions of capitalism. The novelty of capitalism in world
history is that it is a classsociety in which one classrepresentsan
90 SOCIAI K NO W TII X; t I O UNDATI OAND
N SUI 'I I I S] I ?UCI URt9I

entire potential social systcnr;it is this which was cxprcsscdby historical pcriod possesscsits own laws . . 'fhc old econo-
the socialist movement. In this context Marx was an outsidor nrists rnisunderstoodthe nature of economic laws when they
within the system itself, in that he could look at it from the likened them to the laws of physics and chemistry. A more
perspectiveofa society yet to be created,but ctlreadyrepresented thorough analysis of the phenomenashows that social organ-
by new categories of socictl action. isms differ among themselves as fundamentally as plants or
Similarly, Marx claimed that the categoriesof capitalist culture animals. Indeed, one and the same phenomenon falls under
come to be used to understand other (earlier) forms of society quite different laws in consequenceof the different general
only when capitalism itself is seenas an historically specific form structure of these organisms, the variations of their individual
organs, and the different conditions in which these organs
of society, rather than as either the equivalent of other forms or
function. a8
the endpoint of historical development.a6The scientific strength
of classicalpolitical economyderived also from its peculiar his-
Marx is therefore with Winch in opposition to the idea of social
torical position at a period when a new form of society, capital-
scienceas aiming at the discoveryof generallaws of social life.
ism, was still in conflict with remainsof the pre-capitalistsystem.
What general "laws" are frameable in this realm are reducible
(This is visible, for example,in Smith's polemic againstmercan-
"to a few very simple characteristics,which are hammered out
tilist policy and for laissez-faire.)Even thoughthe classicalthink-
into flat tautologies."aeSimilarly, the relation between social
ers, true representatives of the Enlightenment,made their argu-
consciousnessand relations of production, between superstruc-
ments in terms of the "naturalness" of capitalist institutions,
ture and foundation, cannot be statedexcept by the most abstract
they still in practice contrasted those institutions with the
of formulas, so that "in order to examinethe connectionbetween
leftovers of the earlier order. This implicit conflict of points of
spiritual production and material production it is above all neces-
view, then, brought into focus the specific structure of the new
sary to grasp the latter itself not as a general category but in
system;so that, "although encompassed by [the] bourgeoishori-
definite historical form. ' 's0Agreement with Winch ends, howev-
zon, Ricardo analyzesbourgeoiseconomy,whose deeperlayers
er, at the conception of this "definite historical form." For
differ essentially from its surface appearances,with such theo-
Winch cultural forms are wholly defined by the meanings they
retical acumenthat Lord Broughamcould sayof him 'Mr Ricardo
representfor natives, so that understandingsociety is "grasping
seemedas if he had dropped from another plane1."'+t
thepoint or meaning of what is being done or said."5r For Marx,
Whether or not this explanation for the anthropological char-
in contrast, "grasping the point or meaning" of cultural items for
acter of Marx's study of economicsis correct, it is certainlytrue
the natives is only the prelude to the knowledge he is seeking:it is
that this is one way of explaining why he called his writings a
only to identify that which is to be explained,that whose meaning
critique of political economy.His position was, in Marx's own
for the scientific investigator is to be established.
opinion, ably summed up by I.I. Kaufmann, the first Russian
For Max Weber, as we saq complete explanationrequires the
reviewer of Capital, who wrote that the position that "the general
conjoining of a motivational interpretationwith observedstatisti-
laws of economic life are one and the same, no matter whether
cal regularity of behavior. As we noted in Chapter 2, Winch quite
they are applied to the present or the past,"
correctly points out that if we construe meaning as something
"subjective" it is hard to see how correlations with statistical
is exactlywhat Marx denies.Accordingto him, suchabstract regularities canjustify interpretations.Therefore, of necessity,in
laws do not exist . . . On the contrary,in his opinion, every
attempting to describea social situation in natural law-like terms,
92 SOCIAL KNOWIEDGI
Weberceasesto usethe notionsthat would be appropriatcto an
interpretiveunderstanding of this situation.Insteadof spcaking
of the workers in his factory of being paid and spendingmoney,
he speaksof their being handedpiecesof metal, handing those
piecesofmetal to other people and receiving other objectsfrom
them; he does not speak of policemen protecting the workers'
property, but of "people with helmets" coming and giving
back the workers the pieces of metal which other people have
taken from them; and so on. In short, he adopts the external
point of view and forgets to take account of the "subjectively
intended sense" of the behavior he is talking about: and this, I
want to say, is a natural result of his attempt to divorce the
social relations linking those workers from the ideas which
their actions embody: ideas such as those of "money," "pro-
perty," "police," "buying and selling," and so on. Their
relations to each other exist only through these ideas and
similarly these ideas exist only in their relations to each
other.52
Marx wouldbe in full agreement with thebulk of this passage.
Gold is certainlynot (in his eyes)to be identifiedwith money,nor
areall authoritieswith helmetspolicemen.On the otherhand,he
strenuouslydeniesthat socialrelationsexistonly throughideas,
sincethe socialrelationsidentifiedby their idealrepresentations
amongthe natives,being observable,are opento different de-
scriptionsby outsideobservers.The error sharedby Winch and
Weberlies in the assumption thatthe alternativesof "objective"
description and''subjective"interpretation exhaust thepossibili-
ties of socialunderstanding. As we sau this is not true.
In thepreviouschapterI sketcheda distinctionbetween"com-
mon-sense"and "theoretical" thinking about social life. The
first is represented by Azandethinking aboutwitchcraft in re-
sponseto Evans-Pritchard's questions,and by the stockmarket
workers'conceptions of their rolesin theeconomy.The secondis
represented by Evans-Pritchard's systematization of Azandedoc-
trine and by economictheory. Here Winch draws the line of
acceptability,for thethird sortof thinkingaboutsociallife, which
Marx calledspecificallyscientifictheorizing,represents a theo-
IOUNDA]ION AND SUPII?STI?UCTUI?E
93
rctical constructionon thc native categorieswhich the natives
themselveswould not recognizeas valid.
Such an approach is justified by Marx as affording answersto
questions that insiders will not ask. Vulgar economy, as Marx
seesit, merely "interprets, systematizes,and defends" the going
conceptual schemeof (social) things. "But all sciencewould be
superfluousif the form of appearanceof things directly coincided
with their essence."s3Here Marx adaptsthe distinction of es-
sence and appearanceto the social world: the phenomenato be
explained by social theory include the appearancesofthe system
in the eyes of its inhabitants.The distinction betweenthe two, as
Marx conceived it, can be seenin his discussionof the dual nature
of Adam Smith's work in The Wealth of Nations:
On theonehandhe attemptedto penetrate theinnerphysiology
ofbourgeoissociety,but on theother,he partlytriedto describe
its externallyapparentforms of life for the first time, to show
its relationsastheyappearoutwardly,andpartlyhe evenhadto
find a nomenclatureand correspondingmentalconceptsfor
thesephenomena . . for the first time. . . . The one task
interestshim as much as the other, and sinceboth proceed
independently of oneanothel this resultsin completelycontra-
dictory ways of presentation:the one expresses the intrinsic
connections moreor lesscorrectly,theother . . expresses the
apparentconnections[betweeneconomicphenomenal
-apparent, that is, "to the unscientific observer just as to him
who is actually involved . . . in the processof bourgeoisproduc-
tion. "5a Smith was engagedin systematicdescription of different
aspectsof the capitalist social system, in terms of the everyday
categoriesof "value," "price," "rent," etc. At the sametime,
when he usedthe concept of "labor" to explain the phenomenon
of value, he made use of a category which, as used in his theory,
was not part of the vocabulary of everyday life.
But, in Marx's opinion, despite the efforts of the classical
economists, even they "remained more or less trapped in the
world of illusion andnothingelseis possiblefrom the bour-
94 SOCIAT KNOWLTDGI
geoisstandpoint;they all fell thcrefbrcurorcor lcssinto inconsis-
tencies,half-truths,and unresolvedcontradictions."5sTb go bc-
yond the second level of social thought requires the posing of
questionsthat do not arise for the social insider. Just becausein
capitalism the values of commodities are determined by "the
market," and confront the owners of the commodities as facts
limiting their choicesof action, the determinationof exchange
values appearsas the outcome of natural laws of economics. The
fundamentalcategoriesof theselaws-"commodity," "value, "
and (for classicaleconomics)"labor"-appear to the economists
as basic properties of the social world.
Political economyhas indeedanalyzedvalue and its magni-
tude, howeverincompletely,and has uncoveredthe content
concealedwithin theseforms. But it hasneveronceaskedthe
questionwhy this contenthas assumedthat particularform,
that is to say,why labor is expressedin value . .56
Why, that is, is labor representedin this system by price, the
exchange-valueof commodities relative to the money commod-
ity? Marx's answeris to explain the categoryof value, the proper-
ty of goods measuring their mutual exchangeability,in terms of
specific features of capitalist society. Social labor appears as
value, i.e., is organizedthrough the market, (a) when labor has
becomeradically social, in the sensethat production is carried on
for society at large rather than for the producer or immediate
social groups; (b) while control over property remains individ-
ual; (c) and labor itselfis treatedas a good to be exchangedin the
market against other things. This condition requires the separa-
tion of the laboring population from the means of production,
without which they cannot produce for themselves.We may dis-
tinguish two sidesto this explanation.First, the category,"val-
ue," is explainedas the representationin this culture of a certain
organizationof social production. Second, this organization is
explained as the outcome of a particular historical process, and
not asa consequence of some"inner natureof productionitself.''
While it is an explanation, such an account is not a substitute
IOU N D N TIONN N I) SU PL I? SII? U C TU IIT
95
l i l r rl r a tr anslat ion ol't hc ccononr ic cat egor ies. Just as lif e in
Zande society required Evans-Pritchard to talk in terms of magic
and witches, despite his disbelief in them, life in capitalism re-
quires that one does sell one's labor-power, or buy someone
else's, and so think of labor as a commodity. And
IfI statethat coatsor boots stand in a relation to [gold] because
the latter is the universal incarnation of abstract human labor,
the absurdity of the statement is self-evident. Nevertheless,
when the producersof coatsand boots bring thesecommodities
into a relation with . . . gold or silver . . as the universal
equivalent,the relation betweentheir own private labor and the
collective form of societyappearsto them in exactly this absurd
form.5 7
"Value" does not mean "socially necessary abstract labor
time," althoughthat, accordingto Marx's theory, is what it is.
The latter concept cannot replace the former in economic life; it
would have no significancethere. It is because,under the existing
conditions of private appropriation of the social product, the
abstract or social character of labor can have no direct form of
representation(as it would, for example, in a system in which
decisions about social production were collectively made), that
social labor can only be represented,as "value," in the form of
money.
The Marxian analysis of money, as "the commodity whose
natural form is also the directly social form of realization of
human labor in the abstract,"58 is not open to the objection made
by Winch to Weber's attempt at an "objective" description.
Money is defined in terms of the cultural institutions referred to
by the categoriesof "commodity," "price," and so on. At the
sametime, Marx's accountalso escapesthe alternativesposedby
Winch, of cultural relativism or a priori truths of social life, by
explaining both what the essenceof this phenomenonis and why
it appears as it does to the natives of capitalism.
To illustrate this, I will return to the analysisof the categoryof
"capital," money used to make money. According to Marx,
96 SOCIAL KNOWLIDGI
"c apit al is not a th i n g , a n y m o rc th a n mo n e y i sathi ng.l n capi tal ,
as in money, certain social relations appear as thc natural
properties of things in society."5e "Capital" defines things as
they fit into a pattern of social action.
Under certain circumstances a chair with four legs and a
velvet covering may be used as a throne. But this samechaiq a
thing for sitting on, does not become a throne by virtue of its
use-value.The most essentialfactor in the labor processis the
worker himself, and in antiquity this worker was a s/ave. But
this does not imply that the worker is a slave by nature . .,
any more than spindles and cotton are capital by nature just
becausethey are consumed nowadays by the wage-laborer in
the labor process. The folly of identifying a specific socicl
relationship of productiorz with the thing-like qualities of cer-
tain articles is what strikes us most forcefully whenever we
open any textbook on economics and see on the first page how
the elements of the process of production, reduced to their
basic form, turn out to be land, capital, and labor. . . . By
confusing the appropriation of the labor processby capital with
the labor processitself, the economiststransform the material
elementsof the labor processinto capital simply becausecapital
itself changes into the material elements of the labor process
among other things.60
Marx does not stop, as Winch would wish, at noting this social
definition. First, Marx judges this descriptionof the economists
as ideological, in Mackie's sense.It containsthe error of treating
a social character as a natural property. Marx assimilatesthis
error to that of attributing an independentpower to manmade
idols by speakingof the "fetishism of commodities," i.e. the
treatmentof the results of a stageof social developmentas ineluc-
table laws to which humans can only submit.
Second, he explains the ideology, by showing what aspectsof
capitalist social experience (as described in his own theoretical
vocabulary) call for this ideological interpretation. In capitalism,
TOU N D AIIONAN D SU PIIISII? U C IU IIE 9 /
R aw nr at er ialsand t hc objcct ol'labor in gener alexist only t o
ubsorb thc work of others, and the instrumentof labor serves
only as a conductor, an agency, fbr this process of absorption.
. . Since work creates value only in a definite useful form,
and since every particular useful form of work requires materi-
als and instruments with specific use-values, . . labor can
only be drained off if capital assumesthe shapeof the meansof
production required for the particular labor process in ques-
tion, and only in this shapecan it annex living labor. This is the
reason, then, why the capitalist, the worker, and the political
economist, who is only capableof conceiving the labor process
as a processowned by capital, all think of thephysical elements
of the labor process as capital just because of their physical
characteristics. That is why they are incapable of detaching
their physical existenceas mere elements in the labor process
from the social characteristicsamalgamatedwith it, which is
what really makes them capital. They are unable to do this
becausein reality the labor process that employs the physical
qualities of the rneansof production . . . is identical with the
labor process that converts these self-same means of produc-
ti on i nt o capit al. 6-
The sorry state of economic theory mentioned in Chapter I of
this essay, in combination with the most severe economic difficul-
ties since the thirties, has led to a crisis of economics serious
enough for the Nobel and Rockefeller foundations to invest mil-
lions in a study of the conceptual bases of economic policy,
described by Dr Victor Urquidi as "theoretical concepts that
have nothing to do with the present or looking to the future."62
We may note also a rash of books and articles on the insufficien-
cies of economics.However, the economistsare neither resigning
their academicand governmental posts nor being askedto do so.
As onejournalist commented, "The economistshave a great deal
invested in their present 'scientific' approach, both intellectually
and practically-in professorships,learnedjournals, possession
of graduatestudents,positions in government and industry, hon-
ors, up to and including the Nobel Memorial Prize in Economic
Science."63 And their employers must find the thought of being
9 [J SO(-INI KN O W It t X; t I( JU N T) N IIONAN I) SU I'[R SII? U C IU I? T 9 9

withoutscicntil'icaclvicc,cottnscl.andpolicicsto spollsorcqr-rally as Evans-Pritchard: thc cxplanationof a conceptualschemethat is

rlisagrccablc. functionallyindispensable to the life of a culturedespiteits incon-
Truc thoughthis cxplanationis, thereis more to the matter than sistencies and absurdities.
this: wc must explain the material successof economic theory, The adequacy of Marx's theory of capitalist society and its
given its scientific failure. A factor that springs to mind is the forms of consciousnesscannot be a topic for the present essay.
apparent successobtained by Keynesian economics during the But it should be noted that sincethis theory aims not at interpreta-
nineteen-fifties and sixties in managing the businesscycle. But tion of capitalist culture, in the senseof a translation of capitalist
not only did this successturn to stagflationaryashesin the mouths into non-system-specificcategories,but at an explanation of the
of its pundits; more significant is the fact that throughout the continuing existenceof this social system, its adequacyis not to
heyday of deficit financing neoclassicaleconomics continued to be judged by its conformity to "the ordinary consciousnessof the
be taught and acceptedas an integral part of economic science, agentsof production themselves." Rather, the theory's strength
despite the proof of its own inadequacypresentedby the Great is to be measured, first, by its ability to explain that "ordinary
Depression,which had launchedan earlier "crisis of economic consciousness"and, second,by Marx's successin explainingand
theory." This situation, indeed, is reminiscent of nothing so predicting the developmentof a social system structured by the
much as the disinterest of the Azande in exploring and resolving categoriesof capital and wage-labor. As I have indicated, I find
the contradictions in their accountof witchcraft. "Azande do not Capital, judged in these terms, to contain the most important
perceivethe contradiction as we perceive it, becausethey have no well-confirmed theory within social science.Whatever the diffi-
theoretical interest in the subject, and those situations in which culties from which Marx's theory may prove to suffer, however,
they expresstheir beliefs in witchcraft do not force the problem they cannot be shown to derive either from neglect of the inten-
upon them," wrote Evans-Pritchard. Here, however, we see the tional character of social categoriesor from a failure to observe
over-optimism in Barden's idea that theoretical thinking has a the canons of scientific objectivity.
natural tendency to become critical: while for the economists
their subject is a theoretical one, they nonethelesshave little
interest in its methodologicalpresuppositions,given that the aca-
demic or policy advisory situations in which they expresstheir
beliefs in economics do not force such problems upon them.
Marx's theory of capitalist society, as we have seen, goes far
towards explaining this state of affairs. If economic categories
play as central a role in the definition of capitalist social relation-
shipsas witchcraft categoriesdo in the life of the Azande,it is not
hard to understandthe resistanceof this network of conceptsto
the effects of theoretical incoherenceand empirical disconfirma-
tion. What drew Marx's attention to political economy as an
object for critical analysiswas the combination of its theoretical
inadequacy and its dominion over professional intellectual and
ordinary thinking alike. He was thus faced with the sameproblem

Scienceond Society
I have argued in this essay that the "meaningful" character of
"social facts" does not put them beyond the reach of scientific
explanation. On the contrary, it is the very possibility of scientific
understandingof society that makesit important to reject many of
the claims currently made to this title. The many failures of
official social science are due not to an inherent resistanceto
scientific inquiry on the part of social phenomena, but to the
would-be scientists' failure to include their own culture in the
domain of investigation.
Of course, it does not seem to them that they have done so.
Economics, for instance,has explicitly sought to discover highly
abstract laws applying to all social systems,just as sociologists
have striven to unveil generalprinciples of human behavior appli-
cable to the visible spectrum of such behavior. Yet the former
discipline, to take what is generally regardedas the "queen of the
social sciences," remains "in the grip of the world of illusion"
constitutedby the categoriesof everyday life. This judgement of
Marx's has the virtue of explaining the explanatory sterility of
economics and its failure as an empirical predictor and basis for
policy choice. Fundamentally, as we have seen, the root of the
ro o
SC - IIN ( - IAN I) SOC IIIY IOI
problctn is to bc lound in thc goal of discoveringgenerallaws of
social lil'e. These general laws turn out upon examination to be
merely the projection upon the totality of social experienceof the
categoriesof modern society. It is the inability to see what Karl
Korsch called the "historical specificity" of thesecategoriesthat
has blocked even recognition of Marx's discovery of the princi-
ples regulating capitalism as a social system.
It hasbeenheld-despite his approvalof his Russianreviewer's
criticism of the idea-that Marx himself maintained a general
theory of society ("historical materialism"), of which his ac-
count of capitalism is then the application to a particular case.In
the 1859 Preface to Zur Kritik Marx affirms in the tradition of
European"universal history" that "Asiatic, ancient,feudal,and
modern bourgeois modes of production can be designated as
progressive epochs in the economic formation of society," with
bourgeois society as "the last antagonistic form of the social
processof production. "r This way of speakingderives from the
Enlightenment's periodization of history, as preservedand trans-
formed-aufgehoben-by Hegel. While for the Enlightenment
human nature was constant throughout history, with the advance
from one historical stage to another marked by the growth of
knowledge of that nature and the laws governing it, Hegel devel-
oped the Romantic idea of a progressionof the human spirit itself,
through a seriesof civilizations, each with its own unifying prin-
ciple. This view is relatedmore closely than its predecessorto the
modern anthropologicalconceptof culture; it is even an ancestor
of Winch's view of cultures as monadlike systemsof meaning'
Commensurability of cultures is fundamental to the Hegelian
conception, however, since not only their elements but the cul-
tures themselvesfor him have meaning as elements in the over-
arching Culture of human history as a whole. As a result, later
stagescan comprehend earlier ones; and in fact, according to
Hegel, a culture can be scientifically (wissenschafilich)under-
stood only when its day is done: "The owl of Minerva spreadsits
wings only with the falling of the dusk."2
Elementsof this view have certainly been preserved in Marx's
orientation to history, as the phrase quoted above demonstrates.
I02 SOCIAL KNOWLEDGE
At the sametime, however,Marx's view is rnarkcdby an abarr-
donment of the teleology essentialto the Hegelian schemc. For
Marx the great importanceof Darwin's Origin of Specieswas that
it dealt the "deathblow" to teleology in the natural sciences,
while at the same time offering an empirical explanation of its
"rational meaning"-i.e., explaineda teleologicalpattern as the
result of a causalmechanism.3This certainlyrepresentedMarx's
attitude towards teleological explanation in human history as
well. He had already renouncedthe Hegelian vision in 1845,
writing that
History is nothingbut the succession of the separategenera-
tions, eachof which uses. . the productiveforceshanded
down to it by all precedinggenerations, and thus,on the one
hand,continues the traditionalactivityin completelychanged
circumstances and. on the other. modifies the old circum-
stanceswith a completelychangedactivity.This canbe specu-
lativelydistortedsothatlaterhistoryis madethegoalof earlier
history . . . while what is designatedwith the words "des-
tiny", "goal", "germ", or "idea" of earlierhistoryis noth-
ing morethanan abstractionfrom laterhistory,from theactive
influencewhich earlier history exerciseson later history.a
Thirty-odd years later, Marx wrote one of his Russiandisci-
ples to strcss that the "historical inevitability" which Capital
ascribedto the developmentof capitalism "is expresslylimited to
the countriesof WesternEurope."s Similarly, in reply to Mikhai-
lovski, who had criticizedhim for a theoryof "necessarystages"
in history, Marx replied that it was only his Russian critic who
mustabsolutelytransformmy sketchof the genesisof capital-
ism in WesternEuropeinto an historical-philosophicaltheory
of the generalmarch,fatallyimposedon all peoples,whatever
the historicalcircumstancesin which they find themselves, to
arriveat longlastat thiseconomicformation[socialism]which
will ensure,with thegreatestprogressof theproductivepower
of sociallabor,themostintegraldevelopment of man.But I beg
his pardon.This is to do me at oncetoo greathonor and too
sreat shame.
:icttNCt nNt) socr[IY l03
Onc can ol' coursc. Marx continues. come to some more
or less general conclusions about history by comparing the
results of detailed studies of different areas and times.
But an understanding of social history will never be achieved
"with the master-key of an historico-philosophical the-
ory whose supreme virtue consists in being supra-histori-
cal.' '6
We must conclude, then, that despite his use of Hegelian
phraseologyMarx did not conceive of the series "asiatic, an-
cient, feudal, and [capitalist]modesof production" as ordering
all the world's cultures in one connectedhistory. The sequence
appcarsuniversal in characterbecausecapitalism,which devel-
oped from European feudalism, itself developedon the ruins of
"ancient" society,in becominga world systemtransformsall the
social forms it overwhelms. As Marx put it in The German
Ideology, the more capitalism expands,breaking down the isola-
tion of hitherto relativelyautonomouscultures, "the more histo-
ry becomesworld history."T
Even if we disregard the appearancesof a teleological concep-
tion, however, can we not identify "what deservesto be called a
theory of history, which is not a reflective construal, from a
distance,of what happens,but a contribution to understandingits
inner dynamic"? That we can is the claim of G.A. Cohen in his
I
influential book, Karl Marx's Theory of History, in which he at-
t
I tempts to reconstruct"parts of historical materialism as a theory
a
I or infant science."8 Cohen's version of the theory supposedly
I stated in the 1859 Preface, and developed in the seriesof works
I
i constituting the critique of political economy,predicatesof histo-
t
I ry "a perennial tendency to productive progress, arising out of
rationality and intelligence in the context of the inclemency of
nature." The basisof history is thus developmentof the forces of
production, as measuredby the ability to produce a surplus above
current consumption. The basic situation of mankind is said to be
one of "scarcity," meaning that "given men's wants and the
character of external nature, they cannot satisfy their wants un-
lessthey spendthe better part of their time and energy doing what
,i
they would rather not do, engaged in labor which is not experi-
,{ enced as an end in itself. " Human intellisence leadsto the devel-
l;
t:
t,

IO4 SOCIAL KNOWLEDGE
opmentof technologiesincreasinglyableto mect the challcngcol'
scarcity; human rationality in turn dictates that "when knowl-
edge provides the opportunity of expanding productive power
[people] will tend to take it." Productive forces exist and are
utilized only within one or another set of social production rela-
tions, which therefore develop in such a way as to permit the
developmentof productive forces. The forces-the embodiment
of rationality's responseto scarcity-are the primary causal ele-
ment in history; "the property of a set of productive forces which
explains the nature of the economic structure embracing them
is their disposition to develop within a structure of that na-
ture."e
Setting aside the difficulties of the "functional" mode of ex-
planation Cohen employs, this is an extremely implausible recon-
structionof Marx's views, as well as an implausibleunderstand-
ing of history. As we have seen,Marx explicitly rejectsappealsto
features of the human mind, including therefore its alleged ra-
tionality, as a basis for historical explanation. At the sametime,
the concept of "scarcity", drawn-like the characterizationof
work as inherently painful-from the ahistorical economicsMarx
was attacking,has no place in Marx's view. For Marx, "wants"
are precisely not "given"; the "creation of new needsis the first
historical act."r0 This "creation of needs" is itself conditioned
by social experience,in Marx's view; thus he held that "the key
to the riddle of the unchangeabilityof Asiatic societies" is sup-
plied by "the simplicity of the productive organism in theseself-
sufficing communities which constantlyreproducethemselvesin
the same form" and "on an unalterabledivision of labor."tr
Here the nature of the social relations inhibits the progress of
productivity, just as in capitalism the nature of society has led to
its acceleration.
I chose Cohen to discuss becausehis work has been highly
praised among recent contributions to the understanding of
Marx. But I believe that investigation of any other attempt to
develop a general theory of history from Marx's work would lead
to the same result. It would otherwise be remarkable that the
S( - i l N C I n N I) SOC l l l Y l05
argunrcntin Cupitul tnakcsno uscol'a generalsocialtheoryof the
sort we have beendiscussing.We have already seenthat Marx's
general categories, as abstractions,are used not to generateex-
planations for particular phenomenabut to organize the material
in a comparative way. The "guiding thread" in his theoretical
fabric, as he makes clear in an interesting referencein Capital to
the Preface to Zur Kritik, is that "each particular mode of pro-
duction" is the foundationof "definite forms of socialconscious-
ness."r2 For this reasona generaltheory of ideology is impossi-
ble, as we saw in Chapter 4.
It is the rejection of the possibility of general theories of soci-
ety that lies behind the two-fold critique of bourgeois social
thought exhibited in Marx's distinction between "classical" and
"vulgar" economists.The former were scientific in so far as they
sought to constructexplanationsof economic phenomenain "the
real internal framework of bourgeois relations of production,"
while the latter only "flounder around in the apparentframework
of those relations"-apparent, that is, in "the banal and compla-
cent notions held by the bourgeois agents of production about
their own world, which is to them the best possibleone."r3 Here
"the forms of appearanceare reproduceddirectly and spontane-
ously, as current and usual modes of thought; the essentialrela-
tion must first be discoveredby science."l4 In anotherpassage
Marx invokes an analogy with physical science:a scientific anal-
ysis of capitalism is possible only in terms of new theoretical
categories,not part of the system we are trying to explain, "just
as the apparent motions of the heavenly bodies are intelligible
only to someonewho is acquaintedwith their real motions, which
are not perceptibleto the senses." l5 Everyone can seethat the sun
rises and sets, moving around the earth; only by looking at the
solar system from a more abstract point of view-i.e., one that
sets the earth-eye view alongside those from other spatial posi-
tions-can the "essential relation"-the motion of the earth
around the sun-be seen. Once we know the real motion we are
able to explain the apparent motion of the sun around the earth.
Similarly, Marx argued, the origin of profit, which appears to

I06 SOCIAL KNOWLEDGF
result from the investrnentof capital, can be undcrsttxrdonly
from a more abstractpoint of view that seescapitalism as one way
of organizing an exploitative production process.
It is to be noted that for Marx neither its lack of historical
consciousnessnor its theoretical inadequaciesdisqualified classi-
cal economicsfrom the title of "science." Scienceis not to be
identified with correct theories only, but rather with the employ-
ment of certain standardsof theory-constructionand evidencein
the explanationof some domain of phenomena.Yet a theory that
regards capitalist social relations as natural, inevitable forms
"can only remain a sciencewhile the classstruggle remainslatent
or manifestsitself only in isolatedand sporadicphenomena,"r6
since the growth of opposition to capitalism bears witness to the
spiralling crisis cycle characteristicof the developedsystem,and
so to the historicallimit to its existence;any theory worthy of the
name will have to be able to explain this cycle and the implica-
tions of capitalism'sfinitude. Economicsat its best was riding to
an inevitablefall as a result of its appeal,explicit or implicit, to
the conceptof nature in the analysisof social institutions.
While classicalpolitical economywas scientific, in distinction
to vulgar economics, in explaining the phenomenaconstituting
the economic system, the two were alike in viewing "the capital-
ist order as the absoluteand ultimate form of social production,
instead of as a historically transient stage of development."rT
Since capitalism does not in fact representthe emergenceto clear
view of social processesand institutions essential to any and
every form of society,this viewpoint was bound to lead to errors
and an inability to explain fundamentalcharacteristicsof capital-
ism, such as the tendency toward depression and crisis. Thus
"classical political economy's . . . uncritical acceptanceof the
categories[of capitalist society] led it into inextricableconfusions
and contradictions . . ., while it offered a secure base of oper-
ations to the vulgar economistswho, in their shallowness,make it
a principle to worship appearancesonly."r8
By revealing the natural laws of the production and distribu-
tion of wealth political economy was supposed to clarify the
:;( l l N ( | n N l ) :;( x- l l l Y lo/
rangc ol'choiccs opcn t() social policy Inakcrs.Outsidc these
choiccsthcrc was nothins.
Thebelatedscientiflcdiscoverythattheproductsof labor,in so
far astheyarevalues,aremerelythematerialexpressions of the
humanlaborexpended to producethem,marksan epochin the
but by no meansbanishes
historyof mankind'sdevelopment,
the semblance ofobjectivity possessedby the socialcharacter-
isticsof labor.Somethingwhichis only valid for thisparticular
form ofproduction . . appearsto thosecaughtup in the rela-
tionsof commodityproduction(andthisis truebothbeforeand
after the above-mentioned scientificdiscovery)to be just as
ultimatelyvalid asthe fact thatthe scientificdissectionof the
air into its componentpartsleft theatmosphere itselfunaltered
in its physicalconfiguration.re
Justas the chemicalanalysisof the atmospheredid not alter but
only explainedthe composition of the air, so the classicalanalysis
of value in terms of sociallabor time did not alter the reality thus
described.This is of courseexactly what we should expect from a
scientific analysis.However, while Marx's explanationof wfty
labor time appearsin the form of value also does not alter the
reality explained,it doessuggestthe possibilityof suchchangein
a way the classicalanalysisdoesnot. By describing"value" as a
categoryof socialbehavior,as a conceptin terms of which goods
are deaLtwith, Marx showsthat it is not just descriptivebut also
constitutive of the social system. While natural properties of
goods exist independentlyof social action, a property like value
that is itselfa productofsocial behaviorcan ceaseto existevenas
it it was once brought into existence.
l. On the one hand, Marx criticizes the portrayal of the institu-
tions of modern society, by economists and by social theorists
generally,as naturalgivens.He emphasizes,in contrast,not only
the transitory, historical characterof social structuresbut also the
fact that they are products of human activity rather than of a
"logic of history." On the other hand, Marx describeshis own
work on the pattern of natural science,comparing it explicitly to
I08 SOCIAL KNOWLEDGT
'lhis paradox
the physicists' researchinto natural processes.20
reflects, not a self-contradictionin Marx's understandingof his
enterprise, but the nature of the object of enquiry-capitalism-
itself. It is a feature ofthis social systemthat the consequencesof
people's continuing adherenceto the rules and institutions defin-
ing it affect them as forces over which they have no control, and
which indeed the imaginative effort of science is necessaryto
understand. Truly, says Marx, the determination of value by
social labor time "asserts itself as a regulative law of nature "
but-quoting his fiiend Engels-"What are we to think of a law
which can only assert itself through periodic crises? It is just a
natural law which dependson the lack of awarenessof the people
who undergo it."2! As we have seen,the law would remain in
effect even ifpeople did understandit: beyond understanding,the
world requires actual changing. But it would no longer appear
natural.
Thus Marx suggeststhat social sciencecomes into existence
when society appears as a force (or structure) dominating the
experienceof those who make it up. In particular, Marx suggests
that it is capitalism itself which creates the conditions for a
science of society. It is certainly the case, despite the work of
Plato, Aristotle, and Ibn Kaldun, that the social sciencesas we
know them have their origin in the seventeenthcentury. Joseph
Schumpeter attempted an explanation of this historical fact:
commonsense knowledge,relative to scientific knowledge,
goesmuchfartherin the economicfield thanit doesin almost
any other.It is perferctlyunderstandable, therefore,that eco-
nomicquestions,howeverimportant,took muchlongerin elic-
iting specificallyscientificcuriositythandid naturalphenom-
ena.Natureharborssecretsinto which it is excitingto probe;
economiclife is the sum total of the most commonand most
drab experiences.22
It obviously did not occur to him that the subject-matter of
economics-the capitalist market economy-not only was not
banal but did not exist until the modern era. As Marx said.
N( . 1 ANt ) socll lY
1, ( , ll l( ) 9
'l'lrc valuc characlcr ol'thc productsol' labor becomesfirmly
cstablishcdonly when they act as magnitudesof value. These
magnitudes vary continually, independentlyof the will, fore-
knowledge, and actions of the exchangers.Their own move-
ment within society has for them the form of a movement made
bythi ngs. . . . 23
It was this actual experience of a world of autonomous prices. an
historical novelty, which suggested the need for a science of this
new sphere of experience, the market economy. Given the rapid
rise of the languageof natural scienceto its position as privileged
mode of discourse, reflecting both the break with the theological
ideology of the ancien rtgime and the technological needs of
industrial capitalism, the extension of scientific thinking to the
study of society-in Hume's words, "An attempt to introduce the
experimental Method of Reasoning into Moral subjects"-was
inevitable.
Indeed, it was first in the eighteenth century that the word
"society" took on its modern meaningas the body of institutions
and relationships,characterizedby special laws, in which the
individual finds himself.
Only in the eighteenthcentury,in "civil society,"do the var-
iousforms of socialconnectedness confrontthe individualasa
meremeanstowardshis privatepurposes,asexternalnecessity.
But theepochwhichproducesthis standpoint, thatof theisolat-
ed individual,is alsopreciselythat of the hithertomostdevel-
opedsocial. . relations.2a
Hence the subjectivist strain in social science,seeingthe essence
of social life in individual consciousness;hence also the problem
of explaining how the interaction of many individuals produces
results affecting eachperson as the working of an Invisible Hand,
and the conception of social processesas a domain of social facts.
It seems to follow from this understandingof the social sci-
ences as products of capitalism that socialist revolution can be
IIO SOCIAL K NO W LEDG T
expectedto bring with it, along with the abolition ol'wagc labor
and the state,the disappearanceof the subject-matterof econom-
ics and political science-more generally, in G.A" Cohen's
phrase, "the withering away of social science." This is already
suggestedin Marx's comparison of the mystery of commodity
relations with the absenceof economic facts in other forms of
society.Thus in medieval Europe, "precisely becauserelationsof
personaldependenceform the given social foundation, there is no
need for labor and its products to assumea fantastic form differ-
ent from their reality." In this society people neverthelessexperi-
enced social reality ideologically, in terms of such concepts as
kinship, honor, loyalty, and the divine will-"the motley feudal
ties that bound men to their 'natural superiors"' that capitalism
(as Marx put it in the CommunistManifeslo) "has ruthlesslytorn
asunder." On the other hand, in the future socialist society as
Marx imagined it, "an associationof free men, working with the
meansof productionheld in common, and expressingtheir many
different forms of labor power in full self-awarenessas one single
labor force," the relations of the producerswith each other, their
labor, and its products, would again be "transparent." The gap
betweenappearanceand reality filled today by economic science
would no longer exist.25Marx believed that a tendency in this
direction was to be seeneven within capitalism,as in the course
of the classstrugglethe workers discoverthe historicallylimited
character of the laws of economics, so that the continued effec-
tivenessof theselaws becomesa matter of political action.
Thusas soonasthe workerslearnthe secretof why it happens
that the more they work, the more alien wealththey produce,
andthat the more the productivityof their labor increases,the
moredoestheir very functionasa meansfor thevalorizationof
capitalbecomeprecarious; assoonas,by settingup trade
unions,etc.,theytry to organizeplannedco-operation between
the employed and the unemployed in order to obviate or to
weakenthe ruinous effectsof [the] natural law of capitalist
productionon their class,so soondoescapitaland its syco-
phant, political economy,cry out at the infringementof the
1 ;( .ttN ( .t n N t) :i ( x_ tttY ill
"clcn)i.rl''
anclsoto spcak"sacrcd"lawof supplyanddcmand.
H,vcrycon'rbination betweenemployedand unemployed dis-
turbsthe "pure" actionof this law.26
With the triumph of communism, furthermore, Marxian the-
ory itself would ceaseto have meaning. In so far as the aim of this
theory is the unmasking of the social reality beneaththe appear-
ances of capitalist society, it would have no place in a world in
which "the practical relations of everyday life betweenman and
man, and man and nature,generallypresentthemselvesto him in
a transparentand rational form."27 Marxian science,in other
words, will of necessitysharethe fate of its bourgeoisantagonist.
This would not mean an end of scientific thinking about soci-
ety, but only that of bourgeois ideology and its critique. The
disappearance of thosephenomenarooted in the peculiaritiesof
capitalismwould not bring a suddensolutionof all the mysteries
of social behavior: the nature of grammar and its relation to
spoken language, for instance, or the processesby which an
infant becomesa human being would still remain invisible to the
non-scientific eye. In addition, the attempt to construct a demo-
cratically and consciouslyregulatedsociety would require exactly
that questioningof assumptionsand procedureswhich plays an
essentialrole in scientifictheorizing.To be scientifichere would
meanconsciouslyto attemptto placeexistingand proposedinsti-
tutions and proceduresin the context of the past, on the one hand,
and in that of the imagined and desired future, on the other. The
goal of democratic decision-makingwould require attention paid
to comparison and mediation between the variety of cultural
forms constituting societyon the world-wide and local levels. But
one imagines a flavor here quite different from that of today's
social science,as the aim would be explicitly not so much the
discoveryof necessitiesas the expansionof the realm of choice.
Such a future implies a transformation in the conceptionof
social knowledge which goes far beyond the discarding of the
illusionsof the socialpseudo-sciences. Two aspectsof this trans-
formation-thc abandonmentof "value-freedom" as a scientific
ideal, and an altcration of the idea of science itself-may be
 S( - IIN C I AN D SOC ITIY
\12 SOCIAL KNO W LEDG I

discussedhere, however briefly, as they may be discerncd tn tainly thc twcnty ycars' rcncwed expansionof capitalism that
lbllowed World War II were accompaniedby a renewed faith in
trends of the presenttime, bound up with the continuing develop-
the power of scienceto understandand control the world, social
ment of the existing social system. The two are intimately con-
nected in the complex idea of "nature" and of science as the and natural-this was the age both of the atom bomb and of
"macroeconomic fine-tuning." In this context the dominance of
"objective" study of it. Basic to the concept of "nature" in-
positivist philosophy of science is not surprising. But the opti-
volved here is the idea of a "natural order," representedby a
mism of capitalism's earlier periods could not survive the century
system of laws, existing independentlyof the human observer,
of total war, and the social relativism produced in part by the loss
that it is the aim of science to discover. This is a world of
"objective" facts, as opposedto the ("subjective") conventions offaith in progressas ofthe essenceofcapitalism and in part by
and values of human invention. the apparently stable confrontation with state-run societies has
It is in accordancewith this set of conceptsthat social scienceis made possible the development of "post-modern" (if not yet
post-capitalist)philosophiesof science.30
taken to be the study of human nature, or of the nature of society,
One consequenceof this intellectual developmentwas a change
construedaSa Setof drives, instincts, innate mental structures,or
in the understandingof laws of nature. The concept of "natural
social necessities in terms of which complex systems can be
law," applied for most of its history both to the investigation of
explained. A classic version of the ideology of modern science
physical natureand the evaluativestudy of social institutions, is
imagined a unified science, in which sociology, individual psy-
of course itself a metaphor taken from social reality. The medi-
choiogy, biology, chemistry, and physics study the properties of
the same basic stuff at varying degrees of complexity. But the eval Christian picture of God as lawgiving sovereign of creation
gradually gave way to the image of God as engineer,designer of
relevant concept of nature may be shared by those, like Durk-
heim, who think in terms of irreducibly social facts. the world machine. The disappearanceof the divine rule-setter
Brun van Albada suggeststhat not just social sciencebut this from today's science has left the metaphor as designation of a
..science" itself is apartof thedevelopmentofcapitalism, form of description, or as a means to the explanation, of natural
ideaof
.'born of the presumptionof that classwhich had the illusion of phenomena.
having solved all social problems and believed it had the power to We are coming to see that just as what seemed to be social
givens are not universal, so truth need not be understood as
dominate nature as it dominated society.Thanks to the bourgeoi-
somethinggiven and waiting to be discovered,but can be con-
sie the world was now in ordeq and in good order, in harmony
struedas somethinghistorically and socially constructed,without
with the demandsof reason.It ought to run regularly, like a well-
for all that becoming arbitrary or "subjective." The concept of
designedand well-oiled mechanism."28 Van Albada continuesby
"natural order," similarly, can now be seento draw its meaning
pointing out that the shocksto the bourgeoisie's senseof cosmic
from the fact that we are able to construct theories ordering and
order and its place in it occasionedby the economic and social
upheavalsof the latter half of the nineteenthcentury were reflect- explaining our experience,just as it is only by our construction of
categoriesand theory-like systemsof thinking that our experience
ed in the displacementof the confident materialism of the eigh-
which, with- has shapeat all. We now have learned that just as on the individ-
teenth century by "a semi-idealistconception
world, could ual psychological level there is no perception without representa-
out entirely denying the existenceof the material
tion, so scientific experimentsare meaningful only in the context
assignto man only the role of powerlessand fearful observer,"2e
of sets of theoreticalconstructs.
a conception exemplified by Mach's philosophy of nature' Cer-
]1 4 SOCIAL KNO W LEDG E
Science,in a word, is being perceivedmore and morc on the
model of a product of social labor, both as a final product and as a
tool for future production, although few are as clear about this as
is J.R. Ravetz in his description of scientific work as a type of
"craft" exercisedon "artificially constructedobjects" (intellec-
tual constructs)and his conclusion that "the special character of
achieved scientific knowledge"-its justifiable claim to truth-
"is explained by the complex social processesof selection and
transformationof the results of research."3r This new under-
standing of the place of science in the social labor process has
been no doubt in part suggestedby the actual social position ofthe
scientist, far different today from that of the eighteenthor even
the nineteenth century. In the age of "big science," involving
massivestateand private investment, he can no longer be seenas
a disinterestedand independentseekerof truth. This situation has
led to a change
arti-
as radicalas that which occurred. . when independent
sanproducersweredisplacedby capital-intensive factorypro-
ductionemployinghired labor. . . . With his lossof indepen-
dence, the scientistfalls into one of theseroles: either an
employee,workingunderthe controlof a superior;or an indi-
vidualoutworkerfor investingagencies,existingon a succes-
sionof smallgrants;or hemaybea contractor. managing a unit
or establishment which produces researchon a large scaleby
contractwith agencies.32
One casualtyof this situationis the image of the scientistas an
individual mind seekingtruth through analysisof its experienceof
the "external world. " The alternative is a conceptionof scientif-
ic truth not in terms of a correspondenceof mental with natural
objects or structures,but as socially communicable and confirm-
able knowledge.
This conception of sciencealso erodesthe distinction between
science as search for truth, on the one hand, and its technical
application, a non-scientific matter to be decidedby referenceto
values, on the other. The impossibility of sustainingthis distinc-
tion in the twentieth century was perhaps demonstrated most
SC IIN C l AN D SOC ITIY II5
clcarly by E,nricoFermi's useof it in explaininghis feelingsabout
thc atomic bombing of Hiroshima: while unfortunate from the
human point of view, he explained, it was a wonderful piece of
physics.33The classicalformulation of this ideal for the social
sciencesis that of Max Weber, in his addresseson science and
politics as vocations. For Weber the "scientific attitude" implied
freedom from the evaluative stanceessentialto political activity.
This attitude reproduceswithin the social sciencesthe opposition
between the (naturally) given and (culturally) chosen or created
aspects of society, as with the difference between the laws of
economics and policy choices evoked by the title of Arthur
Okun's book, Equality and Efficiency. The Big Tiade-Off.3a
The sameattitude appearedwithin Marxism, as exemplified by
Bernstein's distinction between socialism as ethical ideal and
Marxism as a scientific theory of the capitalist economy.35For
this Bernstein seemingly could claim ancestry in Marx's charac-
terizationof his own viewpoint as one "from which the develop-
ment of the economic formation of society is viewed as a process
of natural history." Marx praised the English factory inspectors
for their freedom "from partisanship and respect of persons,"
and observed,in anticipationof Capital's critical reception,that
In thedomainof politicaleconomy,freescientificinquirydoes
not merelymeetthe sameenemiesasin all otherdomains.The
peculiarnatureof the materialit dealswith summonsinto the
fray on the opposingsidethe most violent, sordidand malig-
nantpassionsof the humanbreast,the Furiesof privateinter-
est.
In the face of this, he welcomed "every opinion basedon scientif-
ic criticism."36
Yet Marx had no wish to pretend that his theoretical labors
were "value-free. " His aim was to serve what he consideredthe
practical needsof the working classin its struggle againstcapital-
ism. But this for Marx meant not abandonmentof the claim to
scientific truth, but the opposite. Those who wish to control their
social (as their natural) conditions of life need to understandthe
I16 SOCIAL KNOWLEDGE
situationsin which they find themselvesand the possiblcchoiccs
of action within thesesituations.In fact, it seemsclear that Marx
anticipatedthe picture of sciencewhich only now is gaining more
generalacceptanceas the inadequacyofthe older picture becomes
inescapable.With the supposedconflict between objectivity and
values Marx abandoned both the idea that there is an eternal
socialorder to be discoveredand the scientist'sclaim to a special
position of social expertise, basedon his supposedcharacterof a
seeker after truth, above the passionsof the laity.
Thus by "scientific socialism," as Marx put it in reply to
criticism by Bakunin, he meant-in opposition to "utopian so-
cialism which seeksto foist new fantasiesupon the people"-
"the comprehensionof the social movement createdby the peo-
ple themselves. "3TThe historicalprocessMarx was interestedin,
as a socialist theoretician, would consist precisely in people's
attemptsto changethe society in which they find themselves,just
as that society's continuedexistenceconsistsin their continued
obeisanceto the laws of economics.Scientificwork, in leadingto
a better understandingof society and so of the tasks involved in
changing it, should serve as an element of these attempts.
"The productionof ideas,of conceptions,of consciousness,"
Marx says in The German ldeology,
is at first directlyinterwovenwith the materialactivityandthe
materialintercourse of men,thelanguageof real life. . . . The
sameappliesto mentalproductionasexpressed in thelanguage
of thepolitics,laws,morality,religion,metaphysics of a peo-
ple. Men aretheproducers of theirconceptions, ideas,etc.-
real,activemen,astheyareconditionedby a definitedevelop-
ment of their productiveforcesand of the intercoursecorre-
spondingto these. . Consciousness can neverbe anything
elsethanconsciousexistence, andthe existenceof menis their
actuallife-process. 38
Once consciousnessis construed as the organization of human
like its opposite,can
activity, then revolutionaryconsciousness,
be understoodas the systemsand quasi-systemsof conceptionsby
SC IIN C T AN D SOC IITY II7
nrcans of' which pcople organize their revolutionary (or non-
rcvolutionary) behavior. Like everyone, revolutionaries think
about what they are doing: the theoreticians among them are
those who try to explore, systematize, and explain the social
interactionsthat constitutethe social statusquo and the movement
against it.
This attitude was reflected in Marx's conceptionof the tasksof
intellectualsin the socialistmovement. He put his writing skills at
the service of the First International, preparing position state-
ments, official communications,and so forth. In addition, we
should note the project of an Enqu?te Ouvribre, a questionnaire
which Marx published in the Parisian Revue Socialiste in 1880,
and had reprinted and distributed to workers' groups, socialist
and democratic circles, "and to anyoneelse who askedfor it" in
France. The text has the form of 101 questions,about working
conditions, wages, hours, effects of the trade cycle, and also
about workers' defenseorganizations, strikes and other forms of
struggle,and their results.Though this might be describedas the
first sociological survey, its prefaceurges workers to reply, not to
meet the data needs of sociologists or economists, but because
only workers can describe"with full knowledgethe evils which
they endure," just as "they, and not any providentialsaviors,can
energetically administer the remedies for the social ills from
which they suffer." The role which Marx intendedto take on was
that of the collection, organization,and transmissionof this in-
formation; thus, the results of the Enqu?tewere to be analyzedin
a series of articles for the Revue and, eventually, a book.3e
The main task which Marx set himself as a revolutionary intel-
lectual, however,was the task of theory: the elaborationof a set of
concepts, at a fairly abstract level, that would permit a better
comprehension of the struggle between labor and capital. He
prefaced the French serial edition of the first volume of Capital
with an expression of pleasure, that "in this form the book
will be more accessibleto the working class-a consideration
which to me outweighseverythingelse."40The function of the-
ory was to hclp the movement as a whole clarify its problems and
IIB SOCIAL KN O W LEDG E : x. t tNCI nNt ) Soclt lY

possibilities;it did not, in Marx's vicw, placc thc thcorist in a pcrrnitting rcal progrcssin thc dircction of social knowledge. I
dominating (or "hegemonic," as the currently fashionablceu- have hoped to defbnd the powers of scientific rationality against
phemismhas it) position vis-d-visthe movement,but was rather barriers supposedly discovered through philosophical analysis.
what he had to contribute to a collective effort. But my wish is not so much for the improvement of the social
That effort, however,was to be abandoned.Capital was trans- sciencesas academicdisciplines as for the imaginative rejection
formed by the history of the Secondand Third Internationals into of the conditionsof present-daysocietyby their victims. It is this
the "Bible of the working class"-an ironic fate for the product refusal of the laws of capitalist reality which will make possible
of Marx's scientific efforts. Today,in the absenceof a working- both their understanding and their abolition.
classmovement,Marxism has becomea specializationfor intel-
lectuals,and Marx's critique of the ideologyof capitalistsociety
grist for the social-scientific mills that continue to produce that
ideology. Basic elements of Marx's approach to the analysis of
social action and consciousnesshave increasinglycome to be
adoptedby the social sciences-strikingly enough, with the great-
est successand the least resistanceat the points farthest from the
critique ofpresent-dayreality,suchasthe studyofclassicalantiq-
uity or the anthropologyof "simpler societies." But in their
inability to understand the culture that gave birth to them, the
social sciences,and in particular economics,have not advanced
since Marx's day.
Commentingon a recentbook on this stateof affairs, a review-
er noted the author's

firmly heldbeliefthatthenotoriousabsence of a universally

acceptedconceptualcore in socialscience,and of commonly
agreedstandards by which suchcore-concepts couldbe select-
ed,hasbeenengendered no
in small measure by confusionasto
the exactepistemological statusof social-scientificconcepts;
andthattheresultingdisarraycouldbe largelyrectified,if only
socialscientistswere to embracethe right philosophyof sci-
enceand recognizethe true natureof scientificconcepts.ar

Despite my own sympathy for anything suggestingthe social

utility of employment of philosophers,the weaknessof this sug-
gestion is obvious. What needsexplanation is the persistentfail-
ure, throughout the history of these "folk sciences" (as Ravetz
calls them), to embrace conceptions of society and of science
 N OIES I2 ]

Arrrcrican put it, "Thcrc's a suckcrborn cvery

prol)lolcrantlsl]0wrtran.
r r r i n u( "c.

See also the works cited in Chapter 5, note 63.

4. lalling C. Koopmans, "The Constructionof EconomicKnowledge,"
pp. 528-41 in May Brodbeck (ed.), Readings in the Phibsoplry of the Social
S c i e n c e s ,pp .5 3 7 - 8 a n d 5 4 1 .
Notes 29.
5. T. B. Bottomore, Sociology, A Guide to Problems and Literature,p.

6. Bottomorc,
ibid.,p. 32.
7 . B r o d b e ck ( e d .) , p . l .
8. Maurice Natanson,"A Study in Philosophyand the Social Sciences",
p.273.
9. Ernest Nagel, The Structure of St'ient'e, p. 484.
10. Max Weber, The Theory rf Social ctnd Et'onomic Organiz.ation,p.88.
ll. K a r l M a r x, C a p i ta l , Vo l u m e l , p .9 2 .
12. Karl Marx, A Contribution to the Critique of Political Eunomy, pp.
20-21.
13. E. P. Thompson, The Poverty of Theory arul Other Es.sa-ys,
p. 6-5.
14. Marx, Capital, l, pp. 178-79. My emphasis.

2. "Subjective"ond "Objective"
L Introductory
1. F. A. Hayek, The Counter-Revolution rf Scient.e,
p.28.
l. FrederickSuppe(ed.), The Structureof ScientificTheories,p. v. 2. Ibid. , pp. 28, 29, 26.
2. RichardJ. Bernstein,TheRestructuringof Socittland PoliticctlTheory, 3. For a representative selectionof articleson methodological
individual-
p. x ii. ism, seeBrodbeck(ed,.),Readings,Chapters13-18;for a critiqueseeAlan
3. LeonardSilk, writinginTheNew YorkTimes of 27 July1983,reported Garfinkel, Forms of Explanation: Rethinkingthe Questirnsof Social Theory
(New Haven:YaleUniversityPress,l98l).
recentresearchon economicforecasting,by Victor Zarnowitzof the National
Bureauof EconomicResearch: 4. Weber,Theory,p.89.
5. Seetheeditors'introduction to H. H. GerthandC. WrightMills (eds.),
After examining the forecastsof 79 individual economistsand firms made From Mar Weber:Essaysin Sociobgy(New York: Oxford UniversityPress,
bctween the fourth quarter of 1968 and the first quarter of 1969, Mr. 1946),pp. 57- 58.
Zarnowitz found that "no single forecaster has been observed to earn a long
6. H a yek,p. 30.
rccord of superior overall accuracy, and indeed nothing in the study would
7. W eber Theor
, y,p. 93.
cncourage us to expect any individual to reach this elusive goal.
"For most of the people, most of the time," he found, "the predictive 8. H a yek,p. 31.
lcvel is spotty, but with transitory spells of relatively high accuracy," a 9. Weber,Theory,p. 101.
finding that might apply to the gambling tables at Las Vegas or Atlantic 10. S e eibid. , p. 99.
City. . . . 11. Ibid.
The economist's strategy, if he is to gain fame and higher rewards for his 12. ErnestNagel, "On the Method of Verstehenas the Sole Method of
ability to beat the standardodds, must be to capitalize on what Mr. Zarnowilz Philosophy,"p. 264.
calls the "transitory spells ofrelatively high accuracy" by taking risks in the 13. Weber,Theory,p. 97.
first place and then heavily advertising his occasionaltriumphs to keep them in 14. rbid.
public consciousness.
15. See,in particular, DonaldDavidson,"Actions,Reasons, andCauses,"
The risk-taking forecaster has the advantageof dealing with a public that
seemsto have a very good memory for forecasting successesand a very poor
Journal of PhilosophyLX:23 (7 November1963),pp. 685-700.
one for failures. This nonsymmetrical public memory appearsto be correlated 16. PeterWinch, TheIdea of ct SocialScienceand its Relationto Philos-
with general credulousness and grecd. As Phineas T. Barnum, the great oph1,,p. ll3.
17. Ibid. , pp. 123, 88.
18. Seeibid., pp. 49-50.

120
,122 N OIIl ; t 23
SOCIAL KNOWLEDGE

1 9. Ib id., p . 11 7. on AlienSociety
3. Understonding
20 . Ib id., p p. 1 09- 110.
21. Ernest H. Hutten, The ldeas of Physics,p. 3. For a detailedaccountof l . W eberTheor
, y,p. 99.
observation in physics, see Dudley Shapere, "The Concept of Observation in 2. Affred Schutz,The Phenomenology of the Social World,p. 234.
Science and Philosophy," Phibsophy of Science 49 4 (1982), pp. 485-525. 3. Winch,Idea, p. 108.
22. John Wallace, "Translation Theories and the Decipherment of Linear 4. Ibid. , p. 90.
B ," p p. ll2-3 . 5. Fritz Machlup,"If MatterCouldTalk," p. 54.
23. Weber is not a psychological behaviorist. Thus he would accept intro- 6. Winch, Idea, p. 58.
spective reports as evidence for the evaluation of motivational models. In his 7. BronislawMalinowski,TheSexualLife of Savages in North-WestMela-
view, however, except for the rare laboratory cxperiment direct psychological nesia(New York: Harcourt,Brace& Jovanovich , 1932),pp. 425-29, cited in
accessis not possible and we must construct and evaluateour models from the AlasdairMaclntyre,"The Ideaof a SocialScience,"in Alan Ryan(ed.), ?'fte
"outside." Philosophyof SocialExplanation,pp. l5-32, p. 16.
24. Wallace, op. cit., p. 123. 8. Winch, Idea.p. 89.
25. Ibid., p. 131. 9. Ibid., pp. 108-9.
26. See, for example, Jean Piaget and Barbel Inhelder, The Psychology ttf 10. Ibid. , p. 107.
the Child (New York: Basic Books, 1969), pp. 4-6. 11. E. E.Evans-Pritchard, TheoriesoJ PrimitiveReligion,p.98.
27. Clyde Kluckhohn, Culture and Behavior, pp. 25, 52. Note that 12. JackGoody,ComparativeStudiesin Kinship, p. xiii.
"learned behavior" does not have here the meaning it has in behaviorist 13. SeeWinch, Idea, p. 104.
learning theory. 14. GopafaSarana,TheMethodologyof AnthropologicalComparisons,p.
28. Kluckhohn, op. cit., p. 63. 15.
29. Alasdair Maclntyre, "A Mistake about Causality in Social Science," 15. E. E. Evans-Pritchard, "The ComparativeMethodin SocialAnthropo-
pp . 6l-62 . The non sequitur in Maclntyre's argument-lack of the concept of a logy," p. 35. The skepticismderivesfrom his sensethat socialfactsmay well
game need not spell inability to recognize rule-following-does not affect his be "so totally differentfrom thosestudiedby the inorganicand organicsci-
essentialpoint. encesthat neitherthe comparativemethodnor anyotheris likely to leadto the
30. The problem offormulating categoriesfor the observation of "subjec- formulationof generalizations comparableto the laws of thosesciences.We
tive" phenomenacan be seenin perhaps its purest form in archaeology,where haveto dealwith values,sentiments, purposes, will, reason,choice,aswell as
the human constructorsof meaning have long since vanished.For an interesting with historicalcircumstances"(p. 33).
account of the problem in this context, see Edwin Wilmsen, Lindenmeier: A 16. H ar r is,p. 41.
Pleistocene Hwting Society (New York: Harper & Row), pp. 45-58 and 17. Winch, Idea, p. ll5.
passim. r8. l bid. , p. 88.
31. Barry Barnes, ScientiJicKnowledge and Sociological Theorl'-,pp. 66' 19. E.E. Evans-Pritchard,Witchcrafi, Oracles, and Magic among the
A zande,p. 224.
32. Winch , Idea, p. 84. 20. Ibid. , p. 229.
33. Ibid., p. 87. 21. Cited in PeterWinch, "Understanding a PrimitiveSociety,"p. 80.
34. Marvin Harris, Cultural Materialism: The Struggle for a Science of 22. Ibid. , p. 82.
Culture, p. 32. Harris's use of the emicietic distinction is a subjectof debate 23. Evans-Pritchard, Witchcrafi,p. 11.
among anthropologists;see Anthony F. C. Wallace's review of Cultural Mate- 24. Ibid., p. 126.
rialism in American Anthropologist 822 (1980), pp. 423-26, and especially 25. Ibid. , p. 18.
Lawrence E. Fisher and Oswald Werner, "Explaining Explanation: Tension in 26. Winch, "Understanding,"pp. 78-9.
American Anthropology," Journal of Anthropological Research 34:2, pp. 21. Ibid. , p. 93.
194-218. The position taken in this paper is much closer to my own than is that 28. Winch, Idea, p. 123.
of Harris. However, the passageI quote here from Harris does not commit the 29. Winch, "Understanding,"p. 82.
faults againstwhich Fisher and Werner argue; since Cultural Materialism was 30. Evans-Pritchard, Witchcrajl,p. 25.
published after their article, perhaps this indicates a responsive change in 3r. Ibid. , pp. 22- 3.
Harris's position. I am grateful for these referencesto Kenneth M. Kensinger, 32. Ib id. , p. 199.
of Bennington College. 33. Ibid. , p. 31.
35. Harris, op.cit., pp. 34-35.
 N o IIS 125
124 SOCIAL KNOWLEDGE

34. Ibid., pp. 25-6. lO. lltid.

35. Winch, "Understanding,"p. 82. l l . Ib id. , p. t Jl; scealsoRuddick,p. 43.
36. Evans-Pritchard, Witchcrafi, p. 244. 12. Winch,"Under st anding,pp. " 8l- 2.
37. Winch, "Understanding,"p. 95. 13. SeeLudwig Wittgenstein,PhilosophicalInvestigations,section265.
38. SaraRuddick,"ExtremeRelativism,"pp. 317-8' 14. GarretBarden,"Method and Meaning,"p. 128.
39. Winch, "Understanding,"p. 99. 15. Evans-Pritchard, Witchcraft,pp. 3l-2.
40. Ibid., p. 102. 16. Ibid., p. 4.
4 r . I bid. , p. 10 3 . 17. B a r den,p. l12.
42. Ibid., pp. 104-5. 18. Evans-Pritchard, Witchcraft,p. 229.
43. Ibid. 19. Evans-Pritchard, Theories,p. 91. It mustbe saidthatEvans-Pritchard's
44. Ibid., p. 107. definitionsof "mystical," "commonsense,"and "scientific" notionsare not
45. An effectiveWinchiancritiqueof just suchan attemptto universalize unproblematic. The first, for example,aredefinedas "patternsof thoughtthat
the conceptof "kinship" can be found in Ruddick,pp. 42 ff. It is hardto see attributeto phenomena super-sensible qualitieswhich,or partof which,arenot
why suchan argumentas shegivesherewouldn'tapplyequallyto attemptsat derivedfrom observationor cannotbe logically inferredfrom it, and which
theydo not possess' ' (Witchcrafi, p. 229). Givenwhatwe know aboutthe role
generalizingcategorieslike "birth," etc.
46. SeeEvans-Pritchard, Witchuail, p. 53. of unobservables in science,and the far from straightforwardconnectionof
47. In Evans-Pritchard's words, "For if the Azandecannotenunciatea scientificconcept-formation with observation, thisdefinitionwouldmakeit no
theoryof causationin terms acceptable in an
to us, they describehappenings easytask to discriminatebetweenmysticalnotionsand scientificerrors.
idiom that is explanatory. Theyare awarethat it is particularcircumstancesof 20. Evans-Pritchard, Witchcra.ft , p. 23.
eventsin their relationto man, their harmfulnessto a particularperson,that 21. Barden,p. I 16. To anticipate,compareMarx on the relationof social
constitutesevidenceof witchcraft." Witchcrafi,p. 24. theoryto socialpractice:"Reflectionon the forms of humanlife, hencealso
scientificanalysisoftheseforms, takesa coursedirectlyoppositeto their real
development. Reflectionbeginspostfestum, andthereforewith the resultsof
ond Explonotion
4. Understonding the processof developmentreadyto hand" (Capital, Vol. I, p. 168).
22. Barden,p. ll7.
23. Barnes,p. 43; seeWinch, "Understanding,"p. 89.
l. Winch, Idea, p. 107. 24. Evans-Pritchard, Witchcraft,pp. 202-4.
2. P.K. Feyerabend, "Explanation,Reduction, andEmpiricism,"p. 68. 25. Bar den,p. 126.
3. I bid. , p. 81 . 26. I.L. Mackie,"IdeologicalExplanation,"p. 184.
4. Ibid., p. 59. For theexampleof "force" in AristotelianandNewtonian 27. Evans-Pritchard, Witchcrafi, p. 9.
physics,seep. 57 and p. 89. 28. MalcolmD. Mcleod, "OraclesandAccusations amongtheAzande,"
5. I bid. , p. 83 . pp. 170, 176; the point had alreadybeen made by Evans-Pritchard:"The
6. This is thesituationpicturedby theSeligmans in theirstudyof thetribes poisonoracleis a maleprerogativeand is one of the principalmechanisms of
of theSudan:on thesubjectof magic,' 'theblackmanandthewhiteregardeach male control and an expressionof sex antagonism"(Witchcrail,p. 130).
otherwith amazement: eachconsiders thebehaviorof theotherincomprehensi- 29. Evans-Pritchard, Witchcrail,p. 18.
ble, totally unrelatedto everydayexperience,and entirely disregardingthe
knownlawsof causeandeffect." C.G. andB.Z. Seligman,PaganTribesof the it' 30. Ibid. , p. 25.
31. Mackie,p. 187.
Nilotic Sudan(London:Routledge,1932),p.25,citedin StevenLukes,"Some 32. Ibid. . pp. 193- 4.
ProblemsaboutRationality, " pp. 194-213in BryanWilson (ed.),Rationality, 33. Ibid., p. 194.
p . 198. 34. Winch,Idea, p. 123.
7. Winch, "Understanding,"pp. 8l-2. The last two assertions
sentence
in this
are highlydisputable.AlasdairMaclntyrehasarguedthe contraryof
li
I
35. Ibid. , p. 107.
36. Maclntyre,"Is Understanding ReligionCompatiblewith Believing?"
the first on lines similar to thoseof my argumenthere, in "Is Understanding I pp. 67-8 .
ReligionCompatiblewith Believing?";RobinHorton hasdisputedthe second 37. SeeErnestGellner,"Conceptsand Society," p. 43.
in "African TraditionalThoughtandWesternScience,"pp. 131-171in Wil- 38. Gellner,pp. 44-5.
son( ed. ) . 39. Ibid. , p. 45.
8. Winch, "Understanding,"p. 83. 40. Ibid., p. 24.
9. Winch, "Understanding,"p. 82. 41. Wallace,pp. 126-7.

126 SOCIAL KNOWLEDGT
ond Superstructure
5. Foundotion
l. stuart Clark, "Inversion, Misrule, and the Meaning of Witchcraft,"
p. 12 7.
2. Ib id.. p p. t2 6- 1.
3 . Ib id., p . 9 8.
4. Martin Hollis, "Reason and Ritual," p. 38'
5. Cliffbrd Geertz, "Ideology as a Cultural System," p. 56.
6. Maurice Mandelbaum, "Functionalism in Social Anthropology,"
p. 324. To my taste, Mandelbaum's phrase suggestsan unwarranted attach-
ment to the covering-law model of scientific explanation. His point can
be extended, however, since functionalist generalizations also permit no
theoretically warranted consequenceswith respectto specific cultural phenom-
ena.
7. Karl Marx, Capital, Volume l, p- 494.
8. Karl Marx and Friedrich Engels, The German ldeology, p. 36'
9. Marx, Contribution, PP. 20-1'
10 . Ibid ., p . 20 .
ll . Ibid . , p . 21 .
t2. Ibid.
f 3. Karl Marx, Capital, Volume III, p. 117.
t4. Ibid., p. 969.
1 5. Ib id., p. 9 56 .
t6 . Ib id., p. 9 69 .
17. Karl Marx, Grundrisse. Foundatittnsof the Critique of Political Econo-
llr,1'(Rough Draft), p. 511.
'
18. Ibi(t., p. 85; see Evans-Pritchardon anthropological"laws": "Such
generalization.s,supported by a few selectedillustrations, are either so general
as to be devoid of significanceor, where more precisely formulated, rest on too
slendcr a base of evidence and fail to take into account negative evidences ...
The more the universality claimed, not only the more tenuous does the causal
interpretation become but the more it losesalso its sociological content" (''The
Comparative Method in Social Anthropology"' pp.24-5)-
19. Max Weber, The Methodology of the Social Sciences,p. 90. In Weber's
own opinion, "all specifically Marxian 'laws' and developmentalconstructs-
insofar as they are theoretically sound-are ideal types," heuristically valuable
but "pernicious" "as soon as they are thought of as empirically valid or real"
(ihid., p. 103).
20. Ibid., p. 106. It should be noted that weber takes this instrumentalist
position with respect to all "precise" concepts.
21. See Leszek Nowak, "Weber's Ideal Types and Marx's Abstraction,"
Neue Hefie fur Phibsophie 13: Marx' Methodologie (Gottingen: Vandenhoek
& Ruprecht, 1978), pp. 8l-91; for the relationshipbetween conceptsand
h.istory in Marx, see Otto Morf , Geschichte und Dialektik in der politischen
bkonomie (Frankfurt am Main: Europaische Verlagsanstalt' 1970).
22. Marx, Grundrisse, P. 100.
23 . Ibid .. p. l0 l.
N OIIS l2/
2.1.K arlM alx. "Rcsult sol t hclr r r r r r ct liat ol'Pr oduct ion,p.
Prcoccss " 982.
2-5.Marx, Grundrisst,, p. 102.
26. Marx, Capital,VolumeI, p. 290 (my emphasis).
2'7. Ibid., p. 284.
28. Marx, Grundrisse, p.92.
29. Marx, Capital, VolumeI, p. 290.
30. Mar x, "Result s, "p. 991.
31. Marx, Capilal,VolumeII, p. 120.
32. Ibid., p. 303. SeeCapital,VolumeI, p. 909.
33. SeePaul Mattick, Jr., "Some Aspectsof the Value-PriceProblem"
tr"r"*irt ,i-i,triirii (C"ni"rsde I'ISMEA, SerieS) l5:6,7 (1981), pp.lz5-
781.
34. Marx, Capital,VolumeI, p. 178.
35. Marx, "Results," p. 998-99.
36. Marx, Grundrisse,p. 104.
37. S.F. Nadel, ZfteFoundationsof SocialAnthropology,p. 5.
38. Machlup,"If MatterCouldTalk," p.297.
39. Ibid., p. 300.
40. The distinctionis apparentlyclear at leastto Martian academics,as
witnessan editorial,entitled"The Martian's View," from the International
Herald Tribuneof 5 February1982:
The man from Mars dropped into Earth orbit, just in time. He had heard
that the first Reagan budget was about to appear. What, he asked, was the
focus of public attention? . . . The Martian wanted to know whether the
[national deficitl number turned out to be accurate. We laughed, and ex-
plained that everyone knew it would be pretty fake from the beginning. . . .
"If everybody knew the number was bent," the man from Mars asked,
"why did the president go to such lengths to produce it?" He kept asking
questions like that. You could tell he was from Mars.
"It makes people feel better," we patiently explained. "That's why Mr.
Reaganis going to struggle so hard . . . to keep his deficit figures under $ 100
billion. It's a matter of paying respect to the proprieties, like the medicine
man doing the rain dance.
If the budget office puts out a fairly reliable set of numbers, the Martian
asked, why does the president put out different ones?
"That's polilics, " we said, "which would hardly interest a serious econo-
mi st l i ke y ou."
"l'm not an economist," the Martian indignantly exclaimed. "I'm an
anthropologist. I'm writing a book on the tribal habits ofthe smaller planets."
"O n t he Wor d 'Cast e', "pp. 23l- 2.
41. JufianPit t - River s,
42. Eric Wolf's positionis typical:"The anthropologist's studyof complex
societiesreceivesits majorjustificationfrom thefact thatsuchsocietiesarenot
aswell organizedandtightly knit astheir spokesmen wouldon occasionlike to
makepeoplebelieve.. . . [The] formal frameworkof economicand political
power existsalongsideor intermingledwith variousother kinds of informal
structurewhich are interstitial,supplementary, parallelto it. . . . The anthro-
pologisthas a professionallicenseto study such . structuresin complex
societyand to exposetheir relationshipto the major strategic,overarching
institutions"("Kinship, Friendship,and Patron-ClientRelationsin Complex
I2B SOCIAL KNOWLEDGE
Societies," pp. l-22 in Michael Banton (ed.),'l-he Sociul Atthrttpttlttgyttf'
Complex Societies, New York: Praeger, 1966). "Professional liccnscs" are
clearly parcelled out by reference to academic fields, rather than by the nature
of social subject matter; anthropology's claim to some bits of modern social
reality are necessitatedby the disappearanceof the "simpler" societies that
were its original preserve. On a more conceptual level, what we see here is the
projection onto modern societyofcategories evolved (within that society) in the
study of "simpler" ones, just as economic anthropologistsattemptedto exam-
ine "primitive" economies with the categories of marginal utility theory.
43. Marx, Capital, Volume I, p. 163.
44. Ibid., pp. 168-9.
45. Pitt-Rivers, p. 250.
46. "[B]ourgeois economics arrived at an understandingoffeudal, ancient,
oriental economies only after the self-criticism of bourgeois society had be-
gun" (Marx, Grundrisse, p. 106). "The owl of Minerva spreadsits wings only
with the falling of the dusk."
47. Marx, Contribution, pp.60-1.
48. Marx, Capital, Volume I, p. 101.
49. Marx, Grundrisse, p. 86.
50. Marx, Theories of Surplus Value, Part I, p. 276.
5l . Winch, Idea, p. ll5.
52 . Ibid .. p p. ll7- 8.
53. Marx, Capital, Volume III, p. 956.
54. Marx, Theories of Surplus Value, Part II, p. 165.
55. Marx, Capital, Volume III, p. 969.
56. Marx, Capital, Volume l, pp. 173-74.
57. Ibid., p. 169. Compare Evans-Pritchard's remarks in Witchcraft,
p. 15l: "It is difficult for us to understandhow poison, rubbing-board,ter-
mites, and three sticks can be merely things and insectsand yet hear what is said
to them and foreseethe future and reveal the presentand the past, but when used
in ritual situations they ceaseto be mere things and mere insects and become
mythical agents." This is of course a prime example of the phenomenon of
fetishism, of which Marx believed the power of gold to incarnate human labor
was another.
58. Marx Capital, Volume I, p.241 .
59. Marx, "Results," p. 1005;seep. 1003: "Man can only live by produc-
ing his own means of subsistence,and he can produce these only if he is in
possessionof the meansof production, of the material conditions of labor. It is
obvious from the very outset that the worker who is denuded of the means of
production is thereby deprived of the meansof subsistence,just as, conversely,
a man deprived of the meansof subsistenceis in no position to create the means
of production. Thus even in the first process, what stamps money or commod-
ities as capital from the outset, even before they have really been transformed
into capital, is neither their money nature nor their commodity nature, nor the
material use-valueof thesecommodities as meansof production or subsistence,
but the circumstance that this money and this commodity, these means of
production and these means of subsistenceconfront labor-poweq stripped of all
material wealth, as autonomous powers, personified in their owners."
N OITS t29
60. Ibid., pp. 997 {t.
61. Ibid. , pp. 10078.
62. "EconomicTheoriesFaceTest." New YorkTimes.2l June1983.
63. Leonard Silk, "Economic Scene," New York Times,S July 1983.
CompareLesterThurow'sarticle, "It's AI1TooEasyto Be a Critic," ibid., 19
June1983.
Among the most importantcriticismsof economictheoryfrom within the
profession,seeSidney Schoeffler,The Failures of Economics.A Diagnostic
Study(Cambridge:CambridgeUniversityPress,1955);OskarMorgenstern,
On the Accuracyof EconomicObservations (Princeton:PrincetonUniversity
Press, 1963) and "Thirteen Critical Points in ContemporaryEconomic
Theory," pp. 1163-l189 in Journal of EconomicLiteraturel0:4 (December
1972);BenjaminWard, Wat Is Wrong with Economics?(New York: Basic
Books, 1972);Guy Routh, TfteOrigin of Economicldeas (White Plains:Inter-
nationalArts and SciencesPress,1975);E. Ray Canterberyand R. J. Burk-
hardt,"Economics:The Embarrassed Science,"NationalScienceFoundation
EVISTResearch Report,1979;andAlfred S. Eichner(ed.), Wy EconomicsIs
Not Yeta Scienc'e (Armonk,NY: M. E. Sharpe,1983).
6. Scienceond Society
l . Ma r x, Cont r ibut ion,p. 21 .
2. G.W.F.Hegel,Phiktsophyof Right (trans.T.M. Knox, Oxford: OUB
1952),p. 13. Ehud Sprinzakhasobservedthat this belief of Hegel'swas the
ancestorof Marx's argument"that an insightfulview of societyin history is
bestto be gainedonly in retrospect.Classicalpoliticaleconomy,accordingto
this conception,can very well understandall the . . stagesof history that
anticipatedits own, by virtue of its inclusivecategories,"suchas thoseof
labor,economy,etc. This explainsMarx's unwillingnessto write "recipesfor
the cook-shops of the future": in Sprinzak'swords,"Marx's objectiveknowl-
edgecould not be extendedto the socialistfutureage,but had to limit itselfto
the critiqueof bourgeois society.. ."(8. Sprinzak,"Marx's HistoricalCon-
ceptionof Ideologyand Science,"pp. 409,410). This is why "scientific
socialism" could not be a scienceof the future societybut only a critical
understanding of the presentday movementto establishone.
3. Marx to Lassalle,16 January1861,in Marx and Engels,Selected
Corresytndencc,p. 151.
4. Marx and Engels,Germanldeology,p. 50.
5. Marx to VcraZasulich,8 March 1881,in Marx andEngels,Selected
Corrasytrulcnce, p. 412.
6. Marx to the editorialboardof the Otechestvennye zapiski,November
1871,in M. Rubel(ed.),Marx Oeuvres,vol.2, p. 1555;and seeM. Rubel
(ed.), Marx/Engels:Die russischeKommune.Kritik einesMythos(Munich:
Carl Hanscr.197211, pp. 49ff.
7. Marx and Engels,Germanldeology,p. 51.
8. G.A. C<rhen, Karl Marx's Theoryof History,p.27.
9. C ohcn,op. r 'it . , pp. 155, 152, 153,161.
10. Marx and Engels,Germanldeology,p. 42. This is the basisfor two
I3O SOCIAL KNOWLEDGE N Oi l S l3l

importantpointsmadeby Marx in his critiqueof politicalcconorny. First,hc r n c r r u r i r so l ato n r i c p h ysi ci sts,a s w cl l a s b y th c co n ti n u cd w o r k o t'so n te
attackedthe biologicalsubsistence conception of wages;in contraclistinction hunclredthousanclscientistson weapons research. In the words ofJacob Bron-
both to the classicaleconomistsand to early socialistslike Lasalle,for Marx crwski. in his book Sciencearul Human Valueslsicl, "science has nothing to be
"the numberandextentof [theworker's]so-callednecessary requirements, as ashamedof even in the ruins of Nagasaki. The shame is theirs who appeal to
alsothe mannerin whichtheyare satisfied,arethemselves productsof history other than the human imaginative values that science has evolved" (London:
. . ." (Capital,Vol. I, p.275). Second, he insistedagainstMalthus(and,after Hutchinson, 1961,p. 83, cited in Ravetz,op. cit., p. 65 n' 41)-values which,
him, thegeneralityof economists) thatit is incorrectto speakof a naturallaw of it need hardly be pointed out, did nothing to prevent the bomb-designersfrom
population;he insistedthateveryform of societyhasits own lawof population. proceeding with their work.
^
11. Marx, Capital,Vol. I, pp. 479,478. SeealsoLawrenceKrader,Marx's 34. A.M. Okun, Equality and Efficiency. The Big Tradeoff(Washington:
EthnologicalNotebooksand The Asiatic Mode of Production (both strikingly The Brookings Institution, 1975)'
missingfrom Cohen'sbibliography,which showslittle sign of acquaintance 35. SeeEduard Bernstein, wie ist wissenschalilicherSoz.ialismusmbglich?
with the marxologicalliterature). (Berlin: SozialistischeMonatshefte, 1901). Bernstein shared Max Weber's
12. Marx, Capital,Vol. I, pp. 175-6,n.35. Seealsoibid., p.273 f . neo-Kantianism. Not too much should be made of this, however, fbr our
13. Ibid., pp. 174-5,n. 34. purposes here, as this philosophical orientation, like Kant's itself, was only a
14. Ibid., p. 682. particular manifestation of a general cultural attitude.
15. Ibid. , p. 433. 3 6 . M a r x, C a p i ta l , Vo l . I, PP.9 1 ,9 2 ,9 3 .
1 6. I bid. , p. 96. 37. Marx. Notes on Bakunin, in Marx and Engels, Werke,Yol' 18' pp'
t7. Ibid. 597-642; translation quoted from Richard N. Hunt, The Political ldeas of
t8. Ibid., p. 679. Marx and Engels (Pittiburgh: University of Pittsburgh Press, 1914), p' 326'
19. Ibid., p. 167. 38. Marx and Engels, German ldeobgy, pp. 13-14'
2 0 . S eeibid. , p. 90 . 39. T. Bottomore and M. Rubel (eds.), Karl Marx: Selected Writings in
2 1. I bid. , p. 168n . SociologyemclSocial Philosophy (Harmondsworth: Penguin, 1963)' pp' 210-
22. J.A. Schumpeter, History of EconomicAnalysis(New York: Oxford 11.
UniversityPress,1954),p. 53. 40. Marx, Capital, Vol. I, P. 104.
23. Ibid. , p. t67 . 41 . Zygmunt Bauman, review of William Outhwaite, Concept Formation
24. Marx, Grundrisse. Seethe articleon "Society", pp.243-47 in R. in Sociat-siience(London: Routledge and Kegan Paul, 1983). inTimes Literary
Williams, Keywords(New York: Oxford UniversityPress,1976). Supplement4,118 (29 April 1983)' p. 441.
25. Marx, Capital,Vol. I, pp. 170, lll.
26. Ibid., pp.793-4. Marx goeson to pointout that the bourgeoisie are
quitehappyto violatethe "sacred" law of supplyand demandthemselves, rn
the caseof labor shortages.
2 1. I bid. , p. 173.
28. Brun van Albada,"Sciencesde la Natureet Soci6t6."
29. Ibid.
30. The phraseis borrowedfrom StephenToulmin'sarticle, "The Con-
strual of Reality: Criticism in Modern and PostmodernScience" (Critical
Inquiry9:l (1982),pp. 93-l I l), expressing a viewon therelationofnaturalto
socialsciencesomewhat similarto my own.
31. J.R. Ravetz,ScientificKnowledgeand its SocialProblems,p.72.
32. Ibid., p. 44.Ravetzquotesthe SovietphysicistKapitsa:
The year that Rutherford died (1938) there disappearedforever the happy
days offree scientific work which gave us such delight in our youth. Science
has lost her freedom. Sciencehas become a productive force. She has become
rich but she has become enslaved and part ofher is veiled in secrecy.I do not
know whether Rutherford would continue nowadays to joke and laugh as he
used to do (ibid., p. 32).

33. That such attitudes were hardlv Deculiar to Fermi is shown bv other
 w ( ) t? KS( _ i l i l ) t3 3

l i c y c n r b c r r tl ,I) .K. "l r xp l l r tl ti o tt. I{ctl u cti o tt,l r td l Jl l p i r i ci sttr ," i n H e r b e r t

ticigl arrrlGrovcr Maxwell (cds.) Mlrurcsotu Sttrtlicsin the Phiktsophl'ttJ
5 <i t , r r tr ,.Vr l . l l l , M i n n ca p o l i s: U n i ve r si tyo f M i n n e so taPr e ss, 1 9 6 2 , p p .
28 9'7.
Gecrtz, Clifilrd. "Ideology as a Cultural System," in David E. Apter (ed.),
WorksCited Itleologv untl Discontenl, New York: Free Press, 1964, pp. 47-76.
Gellner. Erncst. "Concepts and Society," in Wilson (ed.), pp. l8-49.
Grxrdy, Jack. Comparative Studies in Kinship. Stanford: Stanford University
P r e s s .1 9 6 9 .
Harris, Marvin. Culnral Material.ism: The Stuggle for a Scienceof Culture.
New York: Random House, 1979.
Hayek, F.A. The Counter-Revolutionof Science. New York: Free Press, 1955.
Hollis, Martin. "Reason and Ritual," in Ryan (ed.), pp. 33- 49.
Hutten, Ernest H. The ldeas of Physics.Edinburgh: Oliver & Boyd, 1967.
Kfuckhohn, Clyde. Culture and Behavior. New York: Free Press, 1962.
Machlup, Fritz. "If Matter Could Talk," in Morgenbesseret al. (eds.), pp.
286-305.
Mackie, J.L. "Ideological explanation," in StephenKorner (ed.), Explana-
/ion, New Haven: Yale University Press, 1975, pp. 185-197.
Maclntyre, Alasdair. "Is Understanding Religion Compatible with Be-
licving'?" in Wilson (ed.), pp. 62-77.
"A Mistake about Causality in Social Science," in P. Laslett and
W.G. Runciman (eds.), Philosophy, Politics and Society (second series).
Albada, Brun van. "sciences de la Nature et Societe," in Les Cahiers du
Oxford: Blackwell, 1962, pp. 48-lO.
c:omunistnede c:onseils3 (1969).
Mcleod, Malcolm D. "Oracles and Accusations among the Azande," in
Barden,Garret. "Method and Meaning," in Singerand Street(eds.),pp. 105-
Singer and Street (eds.), pp. 158-178.
t2 9.
Mandelbaurn, Maurice. The Anatom\ of Historical Knrnledge. Baltimore:
Barnes, Barry. Scientific Knowledge tutd Sociobgical Theory. London: Rout-
J o h n s Ho p ki n s. 1 9 7 7 .
ledge and Kegan Paul, 1974.
"Functionalism in Social Anthropology," in Morgenbesseret al.
Bernstein, Richard J. The Restructuring of Social and Political Theory- Phila-
(eds.), pp. 306-332.
delphia: University of Pennsylvania,1978.
Marx, Karl. Capital. Volume I. Harmondsworth: Penguin, 1976.
Bottomore, Tom. Sociology. A Guide to Problems and Literature. New York:
Capital. Volume II. Harmondsworth: Penguin, 1978.
Pa nth eo n.1 97 1.
Capital. Volume III. Harmondsworth: Penguin, 1981.
Brodbeck, May (ed.). Readings in the Philosophy of the Social Science.r.New
A Contribution to the Critique of Political Economy. New York:
York: Macmillan, 1968.
InternationalPublishers, 1970.
Clark, Stuart. "Inversion, Misrule, and the Meaning of Witchcraft." Past &
Grundrisse. Foundations of the Critique of Political Economy
Present 87 (May 1980), pp. 98 127.
(Rough Drajl). Harmondsworth: Penguin, 1973.
Cohen, G.A. Karl Marx's Theory of History. Princeton: Princeton University
"Results of the Immediate Processof Production," inMarx, Cap-
Press,1 97 8.
iral, Volume I, pp. 948-1084.
Evans-Pritchard,E.E. "The Comparative Method in Social Anthropology, " tn
Theories of Surplus Value. Part I. Moscow: Foreign Languages
E.E. Evans-Pritchard, The Position of Women in Pimitive Societies and
PublishingHouse, n.d.
Other Essaysin Social Anthropology. New York: Free Press, 1965, pp. l3-
Theoies of Surplus Value. Part II. Moscow: Progress Publishers,
36.
Press,
rf PrimitiveReligion.Oxford:OxfordUniversity
Theories r968.
- '-- andEngels,Friedrich.TheGermanldeology.Karl Marx and Fred-
r976.
Witchcraft,Oracles,and Magic amongtheAunde. Oxford: Oxford crick Engels,CollectedWrks,Yolume5. New York: InternationalPublish-
UniversityPress,1976. crs. 1 976.
Moscow: ForeignLanguagesPublishing
. SclcctetlCorrespondence.
Il orrsc.n. d.

132
I34 SOCIAL KN O W LEDG I

Morgenbesser, Sidney,et al. (cds.).Philostryilrv, tuulMtrlunl.Ilssirys

S<it'trt't,,
in Honorof ErnestNagel.New York: St. Martin's, l9(r9.
Nadel, S.F. The Foundationsof SocialAnthropology.Glencoc:Frcc Prcss.
1953.
Nagel, Ernest."On the Methodof Verstehen as the SoleMethodof Philoso-
phy," in Natanson(ed.),pp. 262-210.

1961.
The Structureof Science.New York: Harcourt,Brace& World, Index
Natanson,Maurice(ed.). Phibsophyof the SocialSciences. New York: Ran-
dom House.1963.
"A Study in Philosophyand the SocialSciences,"in Natanson
(ed.),pp. 211-285.
Pitt-Rivers,Julian."On the Word'Caste,"'in T.O. Beidelman(ed.), The
Tianslationof Culture. Essay.rto E.E. Evans-Pritchard,London: Tavistock
Publications, 1971,pp. 231-256.
Ravetz,I.R. ScientfficKnowledgettnd lts Soc'ialProblenr.r.Oxford: OUP,
191t .
Ruddick,Sara."ExtremeRelativism,"in SidneyHook (ed.),ktnguageand
Philosophy, New York: New York UniversityPress.1969,pp. 4l-7.
Ryan, Alan (ed.). The Phibsophy rtf Social Explanation.Oxfbrd: Oxford
UniversityPress,1973. A l b a d a ,B. V., l l 2 l Feyerabend, P.,50ff.
Sarana,Gopala.TheMethodologyof AnthropologicalComparisons. Tucson: 10, 75JJ.,
foundation/superstructure,
Universityof ArizonaPress,1975. B a r d e n ,G , 5 5 1 , 5 8 , 6 0 . 9 8 91
Schutz,Alfred. ThePhenomenologyof theSocialWorld.Evanston:Northwest- B a r n e s ,B ., 2 7 . 5 5 , 5 8 functionalism,61,'73f.
ern UniversityPress,1973. Bottornore,T., 4
Singer,AndreandBrianV. Street(eds.).ZandeTheme.r. Totowa,New Jersey: B r o d b e ck.M .,5 Gce r tz,C ., 7 3
Rowmanand Littlefield, 1972. Gcl l n e r ,E.,6 7 .f.
Suppe,Frederick.TheStructureof ScientificTheories.(2nded.) Urbana:Univ.
capital.v. 78, 82, 87, 89. 95Jfl
of Illinois Press,1977.
C l a r k , S .,7 l l . H a r r i s, M .,2 8 ,3 7
Thompson,E .P. ThePovertyof 7heoryand Other Essays. New York: Monthly
C o h e n ,G. A., 1 0 3 1, l l 0 Hayck, F. A.. t2Jl.. 65
ReviewPress,1978.
common scnsc/theory. 32J.,56J1., H e g e l .G. W. F., 7 - 5 ,l 0 l
Wallace,John. "TranslationTheoriesandthe Decipherment of LinearB, " rn
Theoryand Decisionll (1979),pp. I I l-140. 92f.,n1 H o l l i s. M ., 7 2 1
Weber,Max. The Methodologyof the Social Sciences.Glencoe:Free Press, 27f.,3Uf.
criteria.
1947. cross-cultural comparison, 18,34ff., Id e a l typ c, l 3 f.,'7 8
The Theoryof Socialand EconomicOrganization.New York: Free 69, 78 , 89 idcasand context,35. 4lf., 45, 59f.,
Press,1947. cul ture,1 4. , 24f . 37.
, 64, 81, 85 6 4 f.,'7 0
Wilson, Bryan R. (ed.). Rationaliry.Oxford: Blackwell, 1977. ideologicalexplanation,60ff., 96f.
Winch, Peter. The ldea of a Social Scienceand its Relation to Philosophy. D urkheimE.
. , 20, l12 incommensurability, 42. 501
London: Routledgeand KeganPaul, 1963. individualism,l3l
"Understanding a PrimitiveSociety,"in Wilson(ed.),pp. 78-11l. cconon r ics, 4, 33, 68, 76f . , 88, 93f . ,
971..r 00, r o5l, 108,I l0J. Koopman.T., 4
cnri c/ctic, 28, 31. 6l Ku h n , T. S., 5 0 . 5 2
, 5. 75, 90, l0l
IJrrl i ght cnr ncnt
Irvi rrrs E. 8. , vii, 36, 387f,
[' rit char d, l a b o r ,8 0 f,,8 3 f.,9 4
5tl l l a n g u a g e .l q . 2 . .1 3 .- 1 9 .5 1 . 5 5 . 6 5 /:
I3 6 INDEX
Machlu p,F., 3 2, 86
Maclntyre,A.,2 6, 66
Mackie, I. L .,60 ff.
Ma cle od , M. D., 62
Ma rx,K.,\J:f.,74 ff . , gof f .
"scientism", 8, I 15
conceptof abstraction,'l'UJ.,89
as anthropologist, Slff.
theory of society 10lfl
as socialisfin telle c t ual,1l6f
conceptof ideology,36, 60, 83
mental/physical,15, 22ff.
80,e2,e5
money,20,
N agcl. E., v. 7 , 1 7
naturallaw, 5, 43, lI2J'.
observarion,21fr'.,32J.
Pareto,V., 20, 3411. . 41J.. 83
philosop hy,38 ,54 ,l18
P iag et,J.,21 , 2 4
P itt-Rivers,J., 8 7, 89
production,'l'ry.,19Jf.
r at ionalit y'1 . l . 5 1 . 5 : 1 / . .l 0 '1
r c alit y .- 19.4 l l J . 5 1 1 . .5 4
r elat iv is ni.1 2 . 5 2 . t 1 .6. t l . l 1 3
Ric ar do,D. , 6 E . 7 7 . 9 0
r it ual, 35, 38 , 4 0 , 4 3 , 4 7
rules, 191, 33J., 31 Aboutthe Author
sociallaws, 5, 16, 78, 86, 901., l00l.
1 0 8 ,I t2
Sm it h,A. , 90 , 9 3
, es
lll,,fill, ,";,P;ilr,.68f
value'95' 10ffi
value-freedom, 6, lll, 1l4JI.
verstehen'1 l'7
'
Wallace,J., 2Jl., 69 Paul Mattick, Jr. studied philosophyat Haverford College and
W eber ,M . , ' 7 , 1 3 . f f . , - 1 0 / ' 7 8 ' 9 l ' l 1 5 receiveda Ph.D. from Harvard University.He has taughtphilos-
W i n c h ,P ., tg.l J' ..3U 1., 83,9U .
50.ff.. ophy and socialtheoryat a numberof institutions,and since 1981
95
has been on the faculty at BenningtonCollege. He is currently
n, 19.38. 54
Wi ttg e n s tciL..
also Associate ResearchScholar at the Center for the Social
Sciencesat Columbia University. He has publishedarticles and
reviews in a number of iournals; this is his first book.
tJt