Meaning and Comparative Concepts Author(s): Timothy McDaniel Source: Theory and Society, Vol. 6, No. 1 (Jul.

, 1978), pp. 93-117 Published by: Springer Stable URL: http://www.jstor.org/stable/656897 Accessed: 21/11/2010 03:47
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=springer. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Theory and Society.

http://www.jstor.org

93

MEANING AND COMPARATIVE CONCEPTS

TIMOTHY MCDANIEL

1. Introduction In recent years cross-culturalcomparison has increasingly come to be regarded as a general sociological methodology and not just a specialized sub-field. This changed perspectiveis partly a product of a greaterawareness in of the parochialnature of many generalizations sociology, which have had to be corrected when confronted with a broader range of cases. Thus, the nuclear family is no longer consideredto be a culturaluniversal.But it also corresponds to a greater emphasis on theory-testing in the hypotheticodeductive sense, an emphasiswhich carrieswith it the need to maximizethe range of variation of social phenomena.1 So closely, in fact, has the comparativemethod been identified with the search for generalpropositions or laws that other types of comparativesociology have been almost forgotten. Generalsocial science theories requiregeneralconcepts which can be related to observations across societies. Such a languagehas to abstract from the imprecision and cultural references of everyday usage in order to servethe scientist's theoretical purposes. In being generaland deductive, the concepts used in comparativeresearchcan avoidboth particularity ethnocentrism. and This does not imply that neologisms must replace ordinarywords, but that these words be given a clear, scientific definition. Thus, for comparative purposes the concept of a political party must be purged of some of its connotations in everyday American usage. Once this scientific language is created, propertiesof social phenomenafrom a wide rangeof societies can be comparedby expressingthem in terms of the observer'sabstractconcepts. But if cross-culturalconcepts must be abstract and general,it is equallyclear that social phenomena themselves are concrete and particular.To a large extent, this individualityis due to the different culturalmeaningsin terms of

Instituteof IbdustrialRelations,University California, Berkeley of

94

that actors relate to each and define their social world. Two practices which in fact be differentif the from observer'spoint of view are very similarmay case they must be. meanings are taken into account, which in this cultural methodology is to one Accordingly, of the centralproblemsof a comparative be reconciled with comhow show the relativity of cultural meanings can of the obserThere are two levels at which this can be done: that parison. the formerlevel themselves and that of the abstractconcepts. It is at vations empirical in middle-range the that problemof meaningis generallyconsidered language is usually The sociology. formulation of an abstract theoretical thus non-problematic,2 of to assumed be the prerogative the investigator,and concept is not truth but utility. In this view, the since criterion of a good human mind which have are concepts nothing but "free creations of the about experience".3The problem useful for the formationof theories proved observations equivalenceof meaningonly intrudesitself in relatingspecific of name of the conceptthese generalconcepts, a difficulty which, underthe to the equivalence relationship,has inspired a variety of studies about indicator including the difficulties of transstimuli and responses cross-culturally, of the challenges of estabComparativistshave freely acknowledged lation.4 formulatingthe equivalentindicatorsfor an abstractconcept, but by lishing it into a technical of issue meaning in this way they have transformed of the problem. In this sense, equivalenceis judged in terms methodological to cultural variations in language, often with no reference investigator's must be equivalent,but For meaning. example, Zelditch agreesthat meanings from culture: "the how the definition of meaning has shifted away notice is not the same as the of meaning French Gross National Product per Capita the United Kingdom".5 of meaning Gross National Product per Capita in of using have often simply ignored the question of the validity Sociologists actors themselves concepts to describesocial practicesto which the abstract alreadyassignedmeaning.6 have concepts that It is exactly this issue of the possibility of comparative with Peter Winch in The Idea of a Social Science, which opens preoccupies quotation from Lessing: thefollowing the same in themIt may indeed be true that moral actions are always however different the selves, however different may be the times and actions do not always societies in which they occur; but still, the same a differentname havethe same names, and it is unjust to give any action own times and amongst its own from that which it used to bear in its people.7 of this principle by Winch seeks to illuminate the philosophic grounding

95

inquiringinto the nature of the concept of the social. His a prioriconceptual study, based largely on ideas from the late Wittgenstein, challenges the freedomof the investigatorto conceptualizesocial life in any way he chooses, for social life is essentially made up of concepts. If he is not to distort the social phenomenon as it is constituted for the actors, the sociologist must incorporatetheir meanings.Since these meaningsvary with culturalcontext, the very project of a comparativesociology is called into question. No longer does the phenomenon of cultural force the investigatorto turn to a purer conceptual languagefree of ethnocentrism;rather,it inhibits such a strategy. This kind of argument, whose philosophical basis I will discuss at greater length shortly, has not received a very warm reception in sociology. Perhaps the issue is dismissedwith the argumentthat sociologicalpraxishas disproved philosophical theory, since comparative concepts do exist. This begs the question, for it is the validity, not existence, of such concepts that is in doubt. Or, such views have been dismissed as "phenomenological subjectivism" and discounted because of their disastrousconsequencesfor comparative sociology8 - though the way in which Winch's position could be describedas subjectivisticmust be sharplydistinguishedfrom the usual sense of that term. Finally, this argumentmay be relegatedto anotherphilosophical paradigm,which cannot on its own premisesbe refuted, but which is of doubtful relevanceto scientific sociology.9 Such views imply that argumentslike Winch's are not really worth considering because they undermine the whole enterpriseof comparison. My position in this paper is a different one. I take Winch'sdiscussion seriously, but I also show that he has unduly constricted the tasks of comparative sociology. At the same time, his analysisof the meaningof the concept of the social seems to point to some crucial gaps in the present-daypractice of comparison. These gaps are not due to a methodologicalfailing, but to the narrowingof perspectivethat this particular methodology entails. Althoughit is true that any methodology exacts some costs by excluding some kinds of issues, it is my feeling that the costs have been particularly high in the present case. 2. Winch'sIdea of the Social The title of Winch'sbook is really something of a misnomer, for it has little to do with the nature of social science. Rather, it inquiresinto the natureof the concept of the social. ThusWinchfollows a traditionin Britishanalytical philosophy known as ordinary language philosophy, which investigates the meaning and implications of the concepts that are fundamental to our

96 thought. These studies are not strictly empiricalbut a priori elucidations of the way concepts are used in practice. They aim to clarify the nature and tradition, presuppositionsof understanding. Accordingto the Wittgensteinian philosophicalanalysisrevealsthat all knowledge is groundedin the language that is given to us in social relations. Understanding alreadypresupposesthe criteriaof identity and differencewhich are implicit in speech. It follows that the epistemologicalsplit between concepts and reality, between the Cartesian subject and the world of matter, loses its force. Understandingis not a synthetic, psychological act of an individualbut an a priori ability which makes experience possible and antedates all questionsabout objects. Menare not isolated subjects who must somehow interpreta world of objects; they are social beings whose very nature it is to already understandthe world throughlanguage. But just as understandingpresupposes social relations, so social relations This is all implicit in the view that reality already presupposeunderstanding. appears to us .through language. Thus, the relations between sociology and philosophy are complex: a philosophicalinquiry is alwaysgroundedin social whose nature phillife, and social life is constituted through understanding, osophy elucidates. Philosophy seeks to reveal what kind of a reality it is that is defined by understanding. aim is not to dictate a specific methodology The for the social sciences but to "place limits on the acceptable forms of social scientific theory"10 showingwhat it is that these disciplinesstudy. by All perception and action, indeed, all experience, is, for Wittgenstein,rooted in social life. Thus, in a sense all action is social action. But Winchwants to reservethis concept for something more specific: social action is that which can be objectively identified in terms of sharedmeainings.The identification of social action pressupposesa common symbolic world because it is itself constituted by meaningsand intentions. If observerscannot assignintentions to action - as in the case of the behaviorof lunatics - it remainsat the level of movement in time and space. Thoughit has its genesisin social life and has consequences for other men, such behavior is not social action because it cannot be interpretedteleologically. This does not mean that social action is fully transparentto other men. Rather, what makes action social is that it transcends the particularity of psychological associations in order to be publicly identifiable. Although these two dimensions of an action -the privateand the public - never fully coincide even in social action, most often there is great overlap,because mental life depends upon the social categories of language. Another way of making the same point is to claim that all social action is

97

This rule-governed. phrasedoes not mean that rules determinesocial action, that the actor and the observermust sharecriteriaabout what is to count but a as certain kind of act for it to be meaningful.So, to use an example from the Winch, anarchist's behavior is rule-governedbecause there are shared for criteria deciding whether it fits this category. These criteriamay not be recognizedby the actor, but it is Winch'sclaim that they must be at explicitly if work actions are to have meaning.Similarly,observersmust in principlebe to able know which rule governsthe act, or the act is not meaningful.For we can only understandactions which we can describein terms of concepts, and such descriptionsimply sharedrules for their concrete application. Winch'sdiscussionof the natureof the social is, as he says, priorto Although any particular methodology, he does draw some general methodological for consequences the social sciences. If the concept of rule-followingis to have any meaning,the actor must be potentially capableof not following the rule,for he must be able to recognize what the rule is. It follows that all humanactions are decisions - even if not consciously made - for they the presuppose ability to do the contrary. This does not mean that they are is or arbitrary unpredictable,but that the touchstone of understanding not predictionbut the ability to interpret in terms of intentions. Conversely, is Understanding a neitherdoes the ability to predict entail understanding. sort of matter: it involves categorizingthe action in terms of the different actor's intentions and reasons, an ability which is given to us through language.Consequently, the explanation of action means elucidating the actor'sreasons. Even where prediction is possible, it is of a very different than in the naturalsciences. character of Itwould be a gravemisunderstanding Winchto view him as recommending a subjectivisticmethodology. The sociologist is not enjoined to empathize with the subjects of his inquiry in order to re-enacttheir beliefs. Criticizing He Winch denies that this is possible in historical research.11 is Collingwood, of the actors, but the not concerned with the specific mental experiences socially sharedcriteriaof meaningwhich permeate all social practice.In this of respect, his position is identical to Schutz's well-known re-interpretation verstehen: verstehenis, thus, primarilynot a method used by the social scientist, but the particularexperiential form in which commonsense thinking takes cognisance of the social cultural world. It has nothing to do with introspection; it is a result of processesof learningor acculturationin the same way as the commonsense experience of the so-called naturalworld. Verstehen is, moreover, by no means a private affair of the observerwhich cannot be controlledby the experiencesof other observers.12

98

Verstehen is not a scientific technique, but the prerequisiteof more specialized modes of knowledge such as science. And the social world whose meaningfulness language guarantees is, though permeated with ideas, an objective one, since it presupposesa community of rules. It is one of the major events of modern thought that men from such different traditionsas hermeneutics (the later Dilthey and Gadamer), phenomenology (Schutz, Ricoeur, and Heidegger), dialectic philosophy (Habermas), and English analytic philosophy in the ordinary language tradition have all converged upon this point. Yet the significanceof this cardinalinsightfor the methodology of the social sciences is still very much in dispute. I have alreadynoted that for Winchit makes explanation in the social sciences "incompatiblewith the kinds of AlisdairMaclntyreinterprets explanation offered in the naturalsciences."13 this to mean that Winchdenies the possibility of causalanalysisin the social sciences.14But Maclntyre objects that an analysis in terms of reasons and their setting in a sharedway of life must be supplementedby referenceto the conditions that shape the actor's actual choice among a number of possible reasonablealternatives.Further,he arguesthat an actor may offer the wrong reason for his act (a fact which Winchalso notes). Gellnerhas offered a much more broadly-based criticism, pointing to the conservativebiases inherent in ordinary language philosophy's linkage of meaning and use.15He reads this cannot be criticized doctrine to mean that contemporaryself-interpretations because they are enshrinedin social usage, part of a coherent way of life. In fact, notes Gellner, our concepts often serve as ideological masks for social practicesin need of criticism. Finally, Louch has accusedWinchof excessive intellectualism, claiming that for Winch all human actions are the result of conventions;'6sociology thereby becomes an essentially philosophicalenterprise.

Winch'sviews, althoughthey do point to issues These objectionsmisrepresent which are either ambiguously treated in his work or given very cursory attention. First, Winch never denies that there are external circumstances which shape our decisions: rather,he recognizesthese as partof the essential form of explanationsin terms of reasons,which is: "in view of such and such considerationsthis will be a reasonablething to do".17 If these "considerations" are to be given the name of causes in recognitionof their sometimes compelling character,Winchdoes not deny the possibility of causalexplanation. But it will be of a very different form from its equivalentin the natural sciences, a difference which KarlPopperalso incorporatesin his model of the or "logic of the situation."Hence, Winchprefersto speakof "circumstances" "occasions"ratherthan causes.The point of his book was not to offer advice

99

to sociologists "on how to go about discoveringthe occasionsfor war"but to "some remarksabout what it is that historiansor sociologists investimake gatethe occasions of."18 Winch may be rebuked for not developing a theory of action, which would clearly require a detailed disfull-fledged of cussion the relationshipbetween circumstancesand reasons. But this was after all, his project. In Schutz's terminology, Winchis concernedwith not, an"in-order-to"account of social action rather than a "because" interSuch a shift in emphasis would imply, notjust a study of the pretation. natureof action, but also a reflection upon the observation of action, As arguelater, this distinction is criticalfor an self-observation.19 I including of evaluation the relevance of Winch's social ontology to the practice of sociology. Gellner'scritique is simply wrong. Winch never contends that our only of standard truth is current usage, though he does argue that criteria of must come to us throughour language: validity we should not lose sight of the fact that the idea that men's ideas and by beliefsmust be checkable reference to somethingindependent- some reality - is an important one. To abandon it is to plunge straightinto an with all the paradoxes that involves... extreme Protagorean relativism, we Nevertheless could not in fact distinguishthe real from the unreal the withoutunderstanding way this distinction operatesin the language.20 is Louch'smisunderstanding perhapsthe most seriousfor my purposes,since it denies the very unity of concept and practice which is the cornerstoneof Winch's challengeto comparison.Winchmakesthe point very well himself: meaningfulbehavior is [not] simply a putting into effect of pre-existing reflectiveprinciples;such principlesarise in the course of conduct and are only intelligible in relation to the conduct out of which they arise. But equally, the nature of the conduct out of which they arise can only be as grasped an embodimentof those principles.21 There is not first the idea of a war, and then the actual war conducted according to this idea; rather, the war itself is action impregnated with concepts. To dividethe two, givingpriorityto one ratherthan the other, is to misinterpretWinch'spoint. Ideas are not the cause but the form of action, a position which accords remarkablywell with Schutz's discussions of how experiencein the social world is inherentlytypified. Partly because of criticismssuch as these, which I view as unfounded,Winch's

100

work has not received the attention it deserves from social scientists. As a "new idealism"(the phraseis Geliner's),his book has been seen as subversive of the very project of an empiricalsocial science. This is not correct. Rather, it asks sociologists to reflect upon our practices in order to see how they correspondto the nature of what we study. The purposeof such reflection is to make conscious the perspectiveswithin which a discipline approachesits subject matter, and, on the basis of this understanding,to suggest ways in which what has been missedmay be illuminated. It is in this light that it is useful to view Winch'schallengeto comparative sociology. In inquiringinto the meaningof the concept of the social, he has developedsome implicationsof the view that concepts are internal to social praccice,and not an arbitraryappendagewhich the scientist can remove at will. If the concepts vary between two different social contexts, this is no simple inconvenience that must be transcendedthrough theoretical abstraction. For if the concepts vary,the practicesalso vary,and we are in dangerof comparingtwo entirely different phenomena. Similarly, if the sociologist redefines the concepts in use in a society, "the events he is studying lose altogether their character as social events."22 In being stripped of their meaning,they are transformedinto something entirely different. Dumont's study of caste, to which I will returnlater, is a good illustrationof the above principles.23 According to Dumont, caste as a social institution in India consistsof certainpracticesof group formationmade intelligiblein terms of a set of religiousprinciples.To removethis ideologicalaspect by definingcaste asclasswithout mobility is to ignorewhat it essentiallyis. Theargumentis powerful,yet not decisive. Ina subtle way, Winchhas begged the central methodological question, which is the relevance of his social ontology for social science.24There is no simpleinferencefrom the natureof socialreality to the way it should bestudied, for social science is about social life rather than a description of it. But it is precisely the merit of Winch's bookto force attention upon this relationship,which is too often ignored. 3.Social Action as Object Winch'sdiscussion of the unity of concept and practice in social life is to applicable all perception.Winchrecognizedthis in his contention that the identification of objects is only possiblethroughconcepts given to us in very The theme is brought out even more clearly in phenomenology, language. whereperception is treated as a unity of matter and form. That is, every act perceptual is composed of two empiricallyinseparableaspects: the peritself and the concept of what is perceived.Husserltermsthe former, ceiving

101

the perception-as-occurrence, noesis; the noema is the abstract"what"of the perception, that aspect which makesit describableand permitsthe identification of sameness and difference. To use an example from Chapman,the noema is like a piano sonata, the noesis an actual performanceof it, "the actual playing being the materialcondition of the existence of the sonata in its full actuality, the sonata being in turn the formal condition of the rendering, the 'what' that is played."25The noema has no existence apart from action, yet it is identifiable as an "essence" underlyingany particular act. It is thus essentiallyconceptualin nature,but, like Winch'sconcepts, it is inseparablefrom the activity. The same abstract distinction applies to perceiving and perception,to speakingand what is spoken, and also to doing and deed. In pointing to these ideational elements in all experience, phenomenologists do not thereby commit themselvesto idealism. Chapman's description of Aristotle's Physics and Metaphysics is more appropriate:they are "reflectivesciences of the empirical."26 Given the inseparability of concepts and experience of objects, and the variabilityof concepts acrosscultures,one might be tempted to conclude that the study of objects must also take into account culturalmeanings.If objects were identical to our perceptionof them, this would indeed be the case, and there could be no cross-cultural science.27Yet I am unawareof any serious calls for a Nuer physics, as opposed to an account of Nuer perception of space, time, and objects. The reason for this is simple: we assume natural objects to be independentof our perceptionof them. It follows that they can be described in a special language which owes very little to everydayconcepts. For Winch, the special feature of social life is precisely this lack of independence of object from concept. While storms and thunder do not depend upon human perception, "it does not make sense to suppose that human beings might have been issuing commandsand obeying them before they came to form the concept of commandand obedience."28 Winchis right that lived social life alwaysinvolvesthe meaningsof the actors. Yet he is wrong when he implies that it can be reducedto the immediacyof interaction. His own analysisof meaningidentified an objective content given in language, so that the social meaning of an expression transcends the subjective associations of the speaker. This objective meaning allows the expressionto be identified outside of the context of any particular processof interaction. This dual aspect of linguisticexpression- as subjectiveintention and objective meaning- is in some ways parallelto the difference between the noesis as process and the noema as content of the event. The importance of this analytical separationis that it points to the independenceof the deed from its actual performance. Action, and therefore social action, has an

102

objective dimension which renders it irreducible to the intentions and meaningsof the actors, despite its dependence on these (justas the sonata is the same as any particular performance it.) of not Objectivityis thus not merely something that the scientific observercreates forhis own purposes,but it is an aspect of action itself. This canbe seen even moreclearly in some additional features of social action which transcendthe consciousprocessesof the subjects.29 First, actions have consequenceswhich may be beyond the control of and understanding the actors. This canonly occurif actions are separablefrom their experiential context and fixed as eventsof a certaintype for other then treat them as objects for men. Men can reflection reaction, and the historicalmeaningthey acquirethroughtheir and consequences may dwarf the importanceof their actual constitution in social interaction.This fact is part of the well-known difficulty of relating inactions to historical events, for history is largely made upof the dividual unintended consequences of intentional social action. In fact, both Popper andHayek view the explanation of unintended consequences rather than action itself (which can meaningful often be explainedin termsof reasons)as thecentral task of sociology. addition, since social actions become the can In of objects reflection for other men, they can interpretedin a variety of be ways. Hermeneuticshas even taught that the originalexperience cannot be as recaptured, the Romantics mistakenly held. Instead, men confront social phenomena, especially from other cultures and objectified epochs, from particularperspective which shapes their encounter. It is a for this reason historicallyimportantsocial actions or institutions that are, like great works of inexhaustiblein what they communicate. art, The philosophical implications of these ideas for the social sciences are crucial. injunction to treat social facts as things can seen as a be Durkheim's purely methodological postulate which has its justification in the partial of objectivity meaningful action. It canthus be divorced from a narrowly account of social life withhout sacrificing the possibility of positivist analysis. Further,if social life cannot be reduced to the processof scientific meaningful interaction, then there is something like that separationbetween perception object which allows natural science to transcend variable and cultural meanings. This same kind of separationwithin action has made a similar transcendencepossible in the social sciences, a possibility which has rise given to a number of objective approachesto the study of subjective phenomena.

103

Action 4. ObjectiveApproachesto Meaningful Action as noema, as objective pattern which can be abstracted from the actual occurrence, only enters into the actor's awarenesswhen he reflects about his action. This reflection gives him the possibility of a kind of knowledgebeyond the understandingwhich is presupposed in social life. When actor becomes an observer,he the can describe,evaluate, and causally what was before only an immediate analyze process,and he can do this from avariety of perspectives.This capacity allows him to distance himself from the actual performance - which in Anscombe's phrase was a "knowledge how"- in orderto know "what"he has done.30 Science also is a "knowingthat" ratherthan a "knowinghow," which is to that it observesratherthan acts. This meansthat say it establishesa relationto ship the contents of action which is not simply the actor's, and it does this to according what is theoretically and empiricallyrelevantin its own terms. Asa result, what is noticed in action by each discipline or theoretical in perspective the social sciences depends upon its particularframe of reference. But beyond this, social scientists have a choice of general tnodes of orientation the content of social life - what I to shall call the structural, and hermeneutic, descriptivemodes. By subordinatingthe contents of social life the theoretical purposes of the to scientist, the theoretical perspectives within of these modes can operatewith each truly comparativeconcepts. The StructuralRelation Structuralism is neutral with respect to contents. studies them only as the It occasion the abstract rules of for relationshipthat are its true interest. The Pythagorean School provides one of the clearest examples of an analysis in purely terms of form. They dissected entities into their principalparts in order reveal the patternof their to The characteristic the interrelationship. of that parts they usually analyzed was spatial dimension, and the relations the among parts primarilymeant the ratios of these dimensions. This set of ratios which characterizedthe whole was considered to be the essence or of logosthe entity.31 Structural linguistics,by treating languageas a system of signs detachedfrom their meanings, is in the same tradition. Here too the goal is to elucidate abstract relationshipsamong contentless entities in order to uncover structural patterns. But the best known example of a structuralmode of analysis in social sciences is surely the structuralanthropology,which took much of its inspirationLevi-Strauss' from linguistics. Levi-Strauss isnot concerned with contents of myths or social the institutions, but with the structural

104

patterns which underlie them. These structural patterns are always identifiable in terms of binaryoppositionsbecauseof the binarynatureof human thought. It follows that underneaththe diversity of contents lies a hidden isomorphism,which it is the task of the scientist to reveal. Like the signsof language,the contents of myths are of no interest in themselves; they are "merelyan example of what is possible.32 like -Levi-Strauss, Winch,is wiing to grant the great diversity of cultural content in social phenomena;only it makes no difference for comparison because structure is primary. Thus, restores to anthropology the universalitythat the concepts of L_vi-Strauss cultureand history threatento take away. There could be no clearer example of the elevation of theory over lived experience or the independenceof the observer from what he observes. HermeneuticApproaches The structuralist approach establishes its autonomy with respect to the contents of social life by abstracting from it. The hermeneutic strategy confronts lived social experience and demandsthat it be interpreted. actively Itassumes that the essential nature of the phenomenon is somehow hidden behind givennessof languageand meaning. The apparentmeaningof the the contents canobscure their reality in two very different ways. First, it can simply present difficulties of interpretationthat demand trainingand skill. example of this is the Protestant reformers' project of Biblical herAn where the truth of the Bible was assumedbut its contents were meneutics, felt be overlaidwith Catholic dogmatics. The truth of the text was indeed to but hidden, this hiddennesswas due to human misinterpretation ratherthan tothe nature of the Bible as revelation. The second type of hermeneutics views obscurity of the phenomenon as part of its very nature. Thus, the the phenomenon must be stripped of its appearances,demystified. Examplesof thefirst kind of hermeneuticsare Hegel's reinterpretationof appearancesin terms their essence as Reason and functionalism'sanalysisof institutionsin of terms functions. Criticalversionsof hermeneuticsinclude Freud'sstudy of of the defense mechanismsof consciousness and Marx'sanalysis of bourgeois institutions terms of the play of interests. These examples disclose a in fundamental aspect of hermeneutics:it re-names phenomenaby fitting them into world-viewor very generaltheoretical perspective.Only such a framea
work canprovide the basis for a systematic of re-interpretation appearances.

recent sociology both the functionalist and Marxist versions of herIn interpretation have been widely influential. In his "Towardsthe meneutic of Sociology the Mind"(first publishedin Englishin 1956 but written in the early 1930's), Mannheim was one of the first to call for a functional

105

interpretationof structured social processes,33 arguingthat it is only by its functionthat we canidentify what a phenomenonis. In this sense, the family is defined by its function", even though many causes - such as "child-rearing the desire companionship,sex, economic advantage - are responsiblefor for its formation.34This formulation leads Mannheim to posit two general sociologicaltasks: causal analysis, will trace the forces which create which andmodify structures,and interpretation,which has thegoal of revealingthe functions which sustaina concrete structure. Parsons' theory of action is a much more sophisticated attempt to classify actionin terms of a set of functional categories.He does not challenge the of meanings the actors as these are given in social life, but transposesthem into a set of comparative concepts developed in his general theoretical scheme. this, he is more systematic than In Mannheim,who did not seek to identifya universal of functional categories. Parsons thus set avoids the that Mannheim'sformulationseems difficulty to involve: if social structures only be identified by their functions, how do we can determinewhat these functions apart from locating them in are alreadygiven structures?Parsons' conceptual scheme, by establishing basic functions theoretically,provides the a method analyticallydefininginstitutions of terms of them. in Like Mannheim,the anthropologistWalter Goldschmidt develops no comprehensive scheme of functional concepts, but he avoids problem Mannheim's by grounding his functions in human needs as these are by social met institutions. Goldschmidt's argument for the need to develop functional for concepts comparisonis especially interestingin that he bases his case on the essential of non-comparability institutions. Like Winch's,his argumentis Institutions, he claims, are inherently incomparablebecause epistemological. they defined differently by are each culture. Goldschmidtrecountsthe futile attempts develop a comparativeconcept of to marriage,concluding that the taskimpossible,since is marriage family, clan and each of (or theother concepts of social strucdoes ture) not exist as somethingwhich is there if only we could it in a put propercentrifuge force out those extraneous and substances that contaminate cross-cultural its purity.3s Comparative concepts of culturallydefined entities therefore falsify reality. Yet Goldschmidt is just as critical of the holistic approachof social anthropology, the constants of human life across cultures. which fails to look Men always confront at similar must a set of problems which stem from a common and similar challenges from physical and psychobiology social

106 environments.These problemsgive rise to a number of needs which must be satisfiedif men are to subsist. Social organization one way of meeting these is needs, even though it also creates new ones, such as the socialization of the young and the constraintof self-interest.Functionsare the processesthrough which human groups meet these exigencies of survival,which include the perpetuationof adaptive social organizationitself. These functions can be performed a wide varietyof institutionalarrangements. by Since at least some functions are universal, concepts describingthem will necessarily be comparative, despite incommensurable institutions. The radical characterof this approachis immediatelyevident: Goldschmidt wants to shift anthropologists'efforts away from the study of culture. No longer will the field worker concern himself with a detailed account of as marriage interpretedby the social actors - the rules of endogamyand incest, reciprocity,and so on. Rather,usingtheoreticalassumptions exogamy, abouthuman nature, he will examine a varietyof functions often involvedin activitiesthat anthropologists intuitively classify as marriage. Once these activitieshave been appropriatelyre-named,they canbe studied as they are causallyrelated to needs. In this way the explanation of culturalidiogive way to an elaborationof scientific generalizations. syncraciescan writersbase their reconceptualization lived social life on the of Functionalist postulatethat action is essentially goal-oriented. Even though the conof creteness social life is sacrificed to an abstract conceptual scheme, functionalists claimto have done justice to the essence of action. Further,this instrumental characterof action is not somethinghidden to social actors;the scientific viewpoint merely elevates it to a central position. In contrast, the theorizing of such men as Habermas, Wellmer,and Apel found the dialectical of possibility a scientific analysis of social action on the duplicity of shared This requiresthat social action be hermeneutically re-interpreted meanings. notjust for scientific purposes but in order to help restore truth to social relations make men transparent tothemselves.Criticalsocial science, as a and of present-day institutions, is one of the tools for this historical critique project. Critical theorists36applaudordinarylanguageanalysiswhen it arguesagainst the positivist attempt to construct a universalscientific languageisomorphic with pre-given a objectivereality. Further,they do not deny that studies such as Winch's describe the nature of meaningful social action accurately, and that these accounts are incompatiblewith a naturalscience approachto social life. they insist that languagephilosophersinvertthe true orderof things: But in Ricoeur'sarrestingphrase, "languageis only thelocus for thearticulation

107 of an experience which supports it, and everything,consequently, does not For arrivein language,but only comes to language."37 this reason,they deny that the conscious subjectis awareof the real meaningof what is expressedin language.Even more, this real meaningcontradictswhat is immediatelygiven to the subject's awareness. This is first of all true for the individual's understandingof his own motivation: Habermasuses Freud to argue that subjectively intended meanings must be re-interpretedas symptoms arising from unconscious motives. Because the actor is not aware of their true nature, these symptoms acquire a compulsive force and act as causes.38 Similarly,on the level of collective action, social power distorts communication based on sharedmeanings.Whatis basic to the interaction- the power relation itself - is masked by the actors' interpretations.In concealing the reality behind language, the shared understandingsof social life become ideology. Marx gives a classic example of this kind of analysis in the EighteenthBrumaire: Upon the different forms of property, upon the social conditions of of existence, rises an entire superstructure distinct and peculiarlyformed sentiments, illusions, modes of thought and view of life. The entire class creates and forms them out of its material foundations and out of the correspondingsocial relations. The single individual, who derives them through tradition and upbringing,may imagine that they form the real motives and the startingpoint of his activity .. .39 It follows that, although men act accordingto reasons,these are not the real sources of their action. Men only have the illusion of choice, since unconscious motives and patterns of social relationship operate clandestinely and circumvent decisions. This gives them their quasi-causalcharacter.40 If hidden force is the rule in social life, the social scientist cannot restrict himself to the languageof his subjects:he must interpretbehaviorin termsof
imposed concepts and statistical laws.4l Therefore the very deformation of

social life creates the possibility for an abstract,comparative languagefor the of human behavior. This is something that neither Levi-Strauss explanation nor Parsons, nor indeed empirical sociologists, attempt to do: they all abstractfrom action ratherthan causallyexplain it. However,there is another side to Apel's model of explanation which radicallydistinguishesit from a purely nomologicalsociology: by revealingthese hidden forces the sociologist makes the actors conscious of them, thus depriving them of their power. The historical purpose of a causal social science, then, is to create the conditions of its own supersession.

108

Approach Descriptive Despite their many differences, both the structural and hermeneutic approachesoperate on a very abstract theoretical plane. This allows them to developconcepts which attempt to identify the essence of social life. Structuralists, functionalists, and neo-Marxistshave this trait in common. In general, comparative methodologists in the United States work with less sweepingtheories, ones which do not involve an overall interpretationof social life. But like the hermeneutic approach, middle-rangecomparativists aredirectly concerned with the contents of social action - not, however,to reinterpretit, but to describe and then explain its qualities.The particular qualities,or variables,which the sociologist studies are given to him by his theoreticalproblemsand assumptions.This procedureis explicit in Zelditch's rule that units are comparable"if and only if a) there exists a variableV common to each of them and b) the meaning of V is the same for all of them."42It is important to note that Zelditch does not distinguishbetween cross-cultural comparisonand any other type. This omission is not an oversight, but occurs because he feels there are no special features involved. Culturalvariationsmay make fulfulling the second criterion more troublesome, but as his example of the differencesbetween Frenchand BritishGNP demonstrates,such difficulties can arise from many different characteristics of the unit. The goal of middle-range comparative sociology is to causally interpret relationshipsamong variables.The emphasis on variablesis consistent with the neo-classical analysis of scientific explanation, which takes as the exThe differencebetween this of planandumcharacteristics objects or events.43 social action can be illustratedby two procedureand the study of meaningful differenttypes of the concept "class."On the one hand, "class"may referto collective action in which men understandthemselvesas part of a historical and moral tradition, and act in concert for sharedends. This concept of class implies all the difficulties for cross-culturalcomparison that Winch has indicated. "It was perhapsa unique formation, this Britishworking class of 1832", says Thompson.44On the other hand, middle-rangecomparativists define classes in terms of a set of qualities which stem from theoretical considerations.For example, Ossowski lists various characteristicsof social class, all of which can be present in varying degrees. As these traits are independent of each other, the decision as to what constitutes a class becomes arbitrary,since groups display different degrees of "classness":the judgment "must ultimately be reached by intuitive judgments... or by considering practical consequences...,,45 For Thompson, the self-interpretationof the actors forces the concept upon the investigator.

109

It is inconceivablethat the explanationof descriptionsof social action should not yield a very different kind of sociology from the comparative analysisof meaningful action. But both Winch and Goldschmidt have warned of the trapsin the latter kind of inquiry. Is there a logic of comparisonof conscious sociallife? How, does this kind of knowledgediffer from that of the objective I approaches discussedearlier?Whatare the canons of judgmentto be applied to such work?Theseare questions given little attention in treatises on the methodology of comparisonin sociology, although the successful empirical efforts of such men as Geertz, Bellah, and Dumont provide a compelling enough incentive. In this disjunction between the methodologist and the practitioner,sociology is re-experiencing the relationship between natural science and the philosophy of science in the heyday of the neopositivist vision. In the caseof natural science, however, actual scientific practice is forcingrevisions in the neopositivist view of scientific method and knowledge.46Similarly, the growing empirical attention in the social sciences to symbolismand meaning may compel us to think through the relationship betweenpractice and prescriptionmore thoroughly. But it would be untrue to say that there have been no efforts in this direction. Therehas been a long tradition reflection on these matters inboth social science and of 'philosophy, andsome recent writers have made major contributions. In the following section,I review some ideas of two such thinkers - the Germanphilosopher Hans-George Gadamerand theFrenchanthropologistLouis Dumont- about theshape of a comparative sociology of meaningfulsocial life. Social Life 5. The Comparison Meaningful of and Winch's Wittgenstein'slinking of meaning and practice has been interas preted entailing monodology of social forms of life, each form impenea trable the others. For Winch, as the final sections of his to essay "Undera standing PrimitiveSociety" show, the charge not true. Nevertheless,it is a is lesson of his work that the identification,not to speak of major comparison, ofinstitutions in different cultures is problematic.Culture,like languagein the Tower of Babel, excludes men from an immediate of understanding each other. As indicated previously,comparativesociologists I this circumvent barrierof meaning abstracting from lived social reality in order to interpret or by describe They are able to reveal common dimensions across cultural it. even boundaries, if these are not apparentto the social actors. The possibility of a strategy is given in the very nature of social life. But there is a such hidden assumption in this procedure which drasticallylimits the scope of comparative sociology: the view that only similar phenomena or commen-

110 suratequalitiesof phenomenacan be compared.47 can be seen most clearly As in Goldschmidt's book, the consequence is that meaningful social life is deemed to be incomparable, it is too richly idiosyncratic. for Although Gadamer'sTruth and Method does not directly addressitself to the social science it is profoundlyrelevantto the central problemsof comparative issue of what sort of understandingwe can have of other traditions.48 He refuses to reduce the foreignnessof these traditionsto common elements, for this involves the denial of their historicity. For the same reason, he rejectsa Romantic hermeneuticsof immediacy,in which the observerempathizeswith an historicaltradition or literary text in orderto transcendhis own position. For Gadamer,the observer'sfinite historical tradition is a given that cannot be transcended,except in a different sense of the word, which shortly will become clear. Correlatively,the objects of analysis have their own finite historical horizons which make them foreign to the outside observer.Their meaning cannot be re-enacted in any imaginative act. Comparisonwhich assimilatesthe foreign to what is familiar- either throughthe applicationof scientificconcepts or an illusoryimaginative identification - stays at the level of form, unable to incorporate contents.49 Ultimately, says Gadamer, it subjectsthe alien object to concepts which fail to do it justice. By refusingto analyze social phenomena in terms of their common elements, by insisting upon the particularityof culture and tradition, Gadamerhas returnedto the issuesraisedby Winchand Wittgenstein. However,he differs from these other two thinkers in systematicallyasking how understandingof alien cultural phenomena is possible, even given the boundaries history and culture. His answeris an elaborationof the position of thatman as a being is fundamentallyopen to other experience,that historical horizons are never completely shut off from each other. The observer's particular standpoint is not an impediment to understanding,but the prerequisitefor an experience of other traditions.Knowledge occurs when this horizonis expanded and enriched, not when it is imaginatively renouncedor eliminated. developingthis position, Gadamerdoes not claim to presenta In methodof understanding other traditions,but to describewhat alwaystakes place init, whether its object be a text, anotherperson, or a foreign culture. in Understanding his sense is not scientific, for it begins with neither a subject cleansed historicity and traditionnor an objective fieldgiven once-and-forof allto this subject. Its goal is not to establishgeneralizations. Rather,it is the between self and other which instructsthe observerabout a shared dialogue question.It follows that all understandingdepends upon something like a method. The knowledge gained in this experience, though noncomparative it scientific,is not thereby purely subjectiveor arbitrary; is the kind of truth that and philosophy provide.50 art

Ill Gadamerdescribesthe processof understanding a dialectic or question and as answer. In the case of understanding another culture(which he does not treat explicitly in Truth and Method) this meansthat the other cultureaddressesa question to the investigator. The specific question never exhausts the culture's potential for becoming relevant to the interpreter, but it does circumscribe any particularprocess of understanding. The question is not at first fully understood, for it is given an alien expression. But for the investigator to be responsiveto it, he must have some degreeof foreunderstanding, which stems from a common concern for the subjectmatterof the question. For the social scientist this might be some common issue of social life, such as the nature of authority, the relationof the individualand society, or the care and socialization of the young. The fact that both the investigator'ssociety and the alien culture must confront these issues is what ultimately makes understanding possible: it ensures that, despite cultural diversity, the investhe other culture are dealingwith the same problem.Only in this tigatorand way canthe other culturepresenta meaningfulquestion. The prerequisitefor understanding Gadamer's in sense is that the investigator relatethe other culture's"response"to the question to his own situation, and not merely treat it as culturallyrelative. This means that he takes the other culture's"statement" seriously as addressingan issue of common concern. Hermeneuticunderstanding only ariseswhen the other culture'sclaim to have something to say challenges the prejudicesand traditions of the inquirer. Expansion of perspective necessitates an essential openness to the other, which implies risks for the investigatorhimself. In genuinely referringthe other culture's"answer"to his own perspective,the interpreterbroadenshis own horizon, and he can then re-encounterthe other culture through the mediumof a slightly different question. When such a dialogueis fruitful and this, like a genuine conversation, is not a matter of will alone - the investigatoris able, not to recapture the other's experience, but to express dimensionsof it in terms transformedby the encounter. Gadamercalls this gradualprocess of understanding,through which the question is elaborated and the answers become progressively more transparent, the "fusing of horizons."5sOut of it develops a broader language able to express the expandedapproachto the dialogue'ssubjectmatter, the common issue which gaveforce to the originalquestion. Inthis view,"the hermeneuticalproblem is not one of the correct mastery of language, but of the proper underof standing that which takes place throughthe medium of language.52? Comparativeconcepts are thus specific to an encounter and elaborated around a subject matter. They are the result of inquiry rather than its prerequisite,and they embody knowledge as much as facilitate it. This

112

knowledge is thereforefinite in a way that scientific propositionsare not. For Gadamerthis accordswith the natureof the subjectmatter, since he wants to deny the distinction between the thing-in-itselfand its appearance.Aside from the philosophic success of this effort, the key point for our purposeis that Gadamer revealsa kind of knowledgewhose very heart is the comparison of meaningful phenomena. He does not claim that this is the sole task of history (or, by extension, social science), since the analyst may, for certain purposes,lack the respect for what a particular culturesays that is presumed in the hermeneutic process. Yet he believes that the true vocation of historicalstudy is only fulfilled when the historianis able to illuminatewhat is significantin anothertraditionby bringingit to bearupon his own. Evenif Gadameris right to call for a hermeneutichistory, the social sciences maybe in an entirely different position. Theirown historicaltraditionsseem to point away from this kind of understanding,which is admittedly not scientificin the usual sense of the word. But if analysessuch as Winch'shave apurpose, it is not to call into question scientific practice,but to show what ismissing.It is for the disciplinesthemselvesto decide whetherto take up the newtasks proposedfor them. Thata hermeneutic comparativesociology might be desirablefollows from Gadamer's analysis. That it is necessaryin order to do justice to sociology's subjectmatter canonly be demonstratedby empiricalwork. The Weberian with its heuristic use of ideal types, gives an importantglimpse of tradition, whatthis kind of sociology looks like, and it has continued to apply a corrective excessively narrowvisions of the discipline. It would be highly to to appropriate relate this type of work to Gadamer's analysis.However,I will discuss more recent contribution, Dumont's Homo Hierarchicus,which a explicitlychallenges comparativesociology as it has flourished in recent years. My reason for choosing this work is its remarkableempirical corto respondence Gadamer'stheoretical analysis of understanding.It is also a demonstrationof the importanceof what comparative powerful sociology has left largely undone. Gadamer,Dumont protests social science's treatment of very different Like social phenomenaas if they were the same. This kind of errorstems in large from the reduction of meaningfulaction to observablebehavior.Just as part much Winch, Dumont insists that ideas are intrinsic to social practices, as which therefore cannot be identified without referenceto beliefs and values. Dumont criticizes most empiricalresearchon caste becauseinvestigators also it (endogamy, lack of mobility, et define in terms of a set of characteristics without havingunderstoodit as a meaningfultotality. Whilethe latter task al)

113 also requiresthat the analyst transcendthe concepts of the social actors, this should be done by revealingthe basic principles behind them and not by them in generalterms from the beginning. perceiving These basic principlesare of course expressedin the conceptual languageof the investigator,for this is the only way that they canbe intelligibleto him, since he is also placed in an historical situation. Dumont therefore views his work as a study of "how man as member of a caste appearsto man as an individual(Westernman)."'3 This contrast sets up the dialectic of divergent which can lead to a greaterunderstanding self and other. As in perspectives of theory, the interpreterenlargeshis own vision by reflection on the Gadamer's foreign; this expanded horizon in turn engenders a more comprehensive interpretation.The basis of this dialectic, what keeps it from becoming sterile,is that all societies are built arounda set of commonthemes, to which their institutions are variant responses.54 Comparison uncovers these universal themes by showinghow they are expressedin idiosyncraticways by differentcultures. It is for this reasonthat caste canteach us about ourselves. Indian caste presentsthe Westernstudent with a hierarchical rankingof men andgroups in terms of religious criteriaof purity. The three main traits of thissystem are hierarchy,separation,and the divisionof labor(together with the interdependence that this implies). For Dumont, this is an adequate preliminary formulation,but the comparativecoordinatesof the caste system muststill be discoveredthrough "application"(Gadamer) - that is, through reflective self-reference.The systemic and ideological characterof Indianlife leads Dumont to refer it to Westernindividualism,with its emphasison man asan atom. But individualism,in turn, has some subordinateexpression in castesociety, an expressionwhich Dumont locates in the politico-economic domainand in religious sects. After a rich discussion of these themes, Dumont ready to formulate a genuinely comparativeconcept of caste: it is entails separation of status from power, with latter subordinateto the the former. Unless it fulfills this criterion, a system of closed status groups,such asAmerican race relations, must receive a different name. It is methodcrucial that Dumont was only able to arrive at this comparative ologically concept through the setting in mutual perspectiveof two different systems of
meaning.

Dumont regardsthis concept of caste superiorbecauseit unites ideological as and empirical dimensions, and because it displays the specificity of Indian institutions even while using universalconcepts. Status and power are dimensions social life in all societies, even though the relationshipbetween them of takes many forms. For Dumont, this is because they are linked to the

114

dialectic between individualand society, one of the universal omnipresent featuresof human life that makes the understandingof diverse cultures possible. sociology as a dialogueconcerninghuman thus portrayscomparative Dumont a possibilities, dialogue which requiresthat man beseen both in his plurality and universality.If it were notfor the former,there would be no alternatives to ponder, and thus no enlargementof cultural horizons. Without univerthere would be no common themes to reflect upon. Men would indeed sality, beclosed to each other. It follows that if comparativeconcepts are to be to adequate this kind of inquiry they must be able to do justice to both the and and the common. It is the lesson of Gadamer Dumont that such specific concepts only emerge through the process of understanding comparative itself. 6.ConcludingRemarks Inthis paper,I have taken a differenttack from the usual one. Becauseof the largely epistemologicalcharacter of the argument, which pointed to the validityof both objective and hermeneuticmodels of comparison,I have not ascribed superiorityto either approach.In this my discussiondiffers from any Radnitzky's(in some ways) similar efforts to reconcile the two "metaInstead of treating scientific"perspectivesof naturalismand hermeneutics.S5 of social analysiswhose naturalistic strategiesas permanentlyadequatemodes to validityis rooted in the objective aspects of social life, Radnitzky assigns model of them the subsidiary function of supplementingthe hermeneutic knowledge. Naturalism comes to the rescue when, because of a lack of transparency,motives and subterraneansocial forces act as causes. By these of their causalforce throughmakingactors conscious of them, stripping the naturalisticapproachpaves the way for its own supersession.In terms of my earlierdiscussion,Radnitzky'sanalysissuffersfrom a failureto appreciate the variety of possible modes of relationshipbetween the observerand social action. The restorationof truth to human action and relationswould hardly entail the invalidity of structuralist,functionalist, or causal studies of social the life. Radnitzky is thus guilty of underestimating broad scope of objective which he too readilyreducesto naturalism. analyticalstrategies, Yet Radnitzky's work has a value which transcendsthe limits of the discussion in this paper, for he asks not simply about the possibilityof different strategiesof knowledge,but about their purposesand relativeworth. Nothing I have said about the independenceof the objectiveand hermeneuticmodels should be taken to mean either that they must be isolated from each other or

115 that they are of equal value. However, in this paper I have followed the usual tactic in treatments of comparative methodology of centering on the issue of what we can, rather than on what we should, do. Attention to the latter issue involves more than just epistemology, or than methodology in the usual sense: it raises the question of the adequacy of a form of knowledge to its object. In judgments of this kind, an investigation, such as Winch's, of the nature of that which we study is indispensable. Comparative sociology, if it were to deprive itself of this criterion, would threaten to become, as Geertz wrote in a recent essay, a set of "logical dreams" and "academic bemusements" whose only virtue is "formal symmetry."56 Again, this is not to say that objective methodologies are necessarily inferior, but only that their appropriateness in a given context must be evaluated by extra-methodological criteria -criteria which can only grow out of the collective exercise of reason.

NOTES has and cross-cultural 1. The link between theory-testing comparison been recognized at least since Durkheim,but only fairly recently have investigators begun to test to theirmodelscross-culturally any greatextent. SocialInquiry(New 2. Ada, Przeworski and HenryTeune, The Logic of Comparative York:Wiley-Interscience, 1970), p. 94. 3. John Kemeny, cited in Robert Holt and John Richardson,"Competing Paradigms in Comparative Politics,"in Robert Holt and John Turner,eds., The Methodology Research(New York:Free Press,1970), p. 24. of Comparative 4. RobertMarsh, Brace,1967), ch. 8. Comparative Sociology (New York:Harcourt 5. MorrisZelditch, "Intelligible in Methods Comparisons," I. Vallier,ed., Comparative in Sociology (Berkeley:Universityof California Press,1971), p. 273. 6. This statementdoes not apply to writersin the phenomenological tradition(I use this term loosely). Cicourel,following Schutz, grants that the investigatormust make use of an abstracttheoreticallanguage,but this language must also take into in account the actor's concepts. See Cicourel,Methodand Measurement Sociology (New York:Free Press,1964). 7. Peter Winch, The Idea of a Social Science (London: Routledge and KeganPaul, 1958). 8. Neil Smelser, "Notes on the Methodologyof Comparative Analysisof Economic Activity,"SocialScienceInformation(April-June 1967), pp. 7-23. 9. Zelditch,op. cit., p. 277. 10. PeterWinch,"Mr.Louch'sIdeaof a SocialScience,"Inquiry7, 2 (1964), p. 205. 11. Winch,Idea,op. cit., p. 132. 12. Alfred Schutz, "Conceptand TheoryFormationin the SocialSciences,"in Maurice
Natanson, Philosophy of the Social Sciences (New York: Random House, 1963),

p. 239. 13. Winch,Idea,op. cit., p. 72. 14. Alisdair MacIntyre,Against the Self Images of the Age (London: Duckworth, 1971), ch. 19. 15. ErnestGellner,Causeand Meaningin the Social Sciences(London:Routledgeand KeganPaul,1973).

116 16. A. R. Louch, Explanationand HumanAction (Berkeley:Universityof California Press,1969), pp. 178-179. 17. Winch,Idea,op. cit., p. 81. 18. Winch,"Mr.Louch's," cit., p. 207. op. 19. Alfred Schutz, "Common-sense Scientific Interpretations HumanAction," of and in Natanson, cit., pp. 302-347. op. 20. PeterWinch,"Understanding Primitive a Society,"in BryanWilson,ed., Rationality (NewYork:Harper Row, 1970), pp. 81-82. and 21. Winch,Idea,op. cit., p. 63. 22. Ibid.,p. 108. of (Chicago:University ChicagoPress,1970). 23. LouisDumont,Homo Hierarchicus 24. Winchis moreambiguous what this relationship on should be than I haveindicated. At one point he merely notes that the theoreticalanalysisof socialpracticesmust presupposethe unreflectiveunderstandingof the participants before it can transcendit. Thus, the psychoanalyst is justified in using technical concepts. these must beintelligiblein principleto the patientwhose behaviorthey However, describe (Idea, op. cit.,89-90). But these passages resonatepoorly with the central tenorof his argument,including thepassagefrom Lessingwith which he chose to his preface own discussion.They serve to indicatehis uneasinessover the implications his doctrine. of 25. HarmonChapman,"Realismand in Natanson,Essaysin Phenomenology," Maurice (The Hague:MartinusNijhoff, Phenomenology 1966), p. 105. 26. Ibid., p. 99. 27. For example, George Steiner in his After Babel citesa number of psychological studies See which give evidencethat sensoryperceptionsare culture-bound. After Aspects of Language Translation (New York:Oxford,1975), p. 212. and Babel: Idea, op. cit., p. 125. 28. Winch, 29. Paul Ricoeur,'The Modelof the Text: Meaningful Action Considered a Text," as Social Research 38, 3 (1971), pp. 529-563. 30. Ibid.,p. 5 38. 31. Hans Jonas,The Phenomenonof Life (New York: Dell PublishingCo., 1966), 93. p. 32. EdmundLeach, Levi-Strauss (London:Fontana,1970), p. 42. 33. Mannheim's concept of a structureis not entirely clear this essay, but he seems in toequate it with patternedrelationshipof elements, whether a configuration of actions(capitalism). See (such as the baroque)or a system of interdependent works Karl Mannheim, Essays ontheSociologyof Culture(London:Routledgeand Kegan Paul,1956), 80-81. pp. 34. Ibid., p. 79. 35. Walter Goldschmidt, Comparative Functiolmlism (Berkeley: University of California Press,1966), pp. 25-26. 36. I referhere to theyounger generationof critical theorists, includingHabermas, Wellmer, Apel. and 37. Paul Ricoeur, Political and Social Essays (Athens, Ohio: Ohio UniversityPress, 1974), 262. p. 38. Jirgen Knowledgeand HumanInterests(Boston: Beacon Press,1971), Habermas, 256-257. pp. 39.KarlMarx,The EighteenthBrumaire LouisBonaparte (New York:International of Publishers, 1963), p. 47. 40. Aron makesa similarpoint in his work on the 1968 Parisrevolts: "in a periodof collective folly, one tries to uncover the causes rather than discuss the pseudointellectual content of the delirium." See Bernard Brown, Protest in Paris New (Morristown, Jersey:GeneralLearning Press,1974), p. 39. 41.Karl-OttoApel, Analytic Philosophy Language and the Geisteswissenschaften of Holland:D. Reidel, 1967), p. 56. (Dordrecht,

117 42. Zelditch,op. cit., p. 273. 43. Carl Hempel and Paul Oppenheim, "The CoveringLaw Analysis of Scientific Explanation,"in L. Krimerman,ed., Nature and Scope of Social Science (New York:Appleton-Century-Crofts, 1969), p. 80. 44. E. P. Thompson, The Makingof the English WorkingClass(New York: Vintage, 1963), p. 830. 45. StanislawOssowski, "DifferentConceptionsof Social Class,"in ReinhardBendix and S. M. Lipset, eds., Class, Status and Power, 2nd ed. (New York: Free Press, 1966), pp. 94-95. 46. GerardRadnitzky,Contemporary Schoolsof Metascience (Chicago:HenryRegnery Co., 1973). 47. One corrollary is that comparison requires the pre-existence of comparative concepts, for otherwisetherewouldbe no way to judge similarity.In this sense,the scientific language performsthe sametask acrossculturesas does everydaylanguage withina "formof life." 48. Gadameris interested mainly in the understanding texts and history, but his of discussionappliesjust as well to other cultures. 49. Hans-Georg Truthand Method(New York: Seabury,1975), p. 206. Gadamer, 50. Ibid., p. xii. 51. Interpretive are approaches often criticizedfor theirfailureto clarifytheirmethods of evaluation of evidence. It is of course an essential task of any discipline of perspectivewithin a disciplineto furnishcriterialfor this kind of judgment.These criteriaalways arise out of a communityof discourse,ratherthan being imposed from outside. And as much recent workin the philosophyof sciencesuggests,they are not canons of verificationin any strict sense, for the growth of science is, at In best, a processof successive approximations. this sense, physicsis differentfrom Gadamer's hermeneutics degreeratherthanin kind. For some helpfulsuggestion in on the kind of knowledgeand judgmentsabout it that should develop within an interpretivetradition, see Clifford Geertz, The Interpretationof Cultures(New York:BasicBooks, 1973), pp. 25-28. 52. Gadamer, cit., pp. 346-347. op. 53. Dumont,op. cit., p. 236. 54. Ibid., p. 340.

Theoryand Society 6 (1978) 93-117