Concept formation in social science William Outhwaite Lecturer in Sociology University of Sussex Routledge & Kegan Paul London, Boston, Melbourne and Henley First published in 1983 by Routledge & Kegan Paul Ltd 39 Store Street, London WC1E 7DD, 9 Park Street, Boston, Mass. 02108, USA, 296 Beaconsfield Parade, Middle Park, Melbourne, 3206, Australia, and Broadway House, Newtown Road, Henley-on-Thames, Oxon RG9 1EN This edition published in the Taylor & Francis e-Library, 2010. To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk. © William Outhwaite 1983 No part of this book may be reproduced in any form without permission from the publisher, except for the quotation of brief passages in criticism Library of Congress Cataloging in Publication Data Outhwaite, William. Concept formation of social science. (International library of sociology) Bibliography: p. Includes index. 1. Social sciences—Methodology. 2. Concepts. I. Title. II. Series. H61.089 300′.1 82–7620 ISBN 0-7100-9195-8 AACR2 ISBN 0-203-83164-0 Master e-book ISBN Contents Acknowledgments iv Introduction 1 1 Concepts of science 4 2 Concepts in science 17 3 Constitution 49 4 Max Weber and concept formation in sociology 86 5 Concepts of society 97 Notes 112 Bibliography 153 Subject index 169 Name index 173 Acknowledgments I am extremely grateful to Tom Bottomore for his patient encouragement during the very long gestation of the thesis which forms the basis of this book, and to my two examiners, Roy Edgley and Anthony Giddens. A number of other people have also made invaluable comments on various parts of the manuscript. I should like to thank in particular Michèle Barrett, Roy Bhaskar, Sue Easton, Peter Halfpenny, Jorge Larrain, Gillian Rose and Norman Stockman. Introduction Since the early 1960s there has been a growing interest in examining the concepts used by scientists in general and social scientists in particular. This is part of a more general tendency towards critical self-reflection within science. Just as natural scientists come increasingly to worry about the moral and practical implications of research in nuclear physics or molecular biology, and social scientists worry about their possible complicity in the military-industrial complex or in the reproduction of bourgeois ideology, so their worries also focus, more precisely, on the validity of their own assumptions, theoretical frameworks, etc. These anxieties are perhaps more extreme in the social sciences, where we may wonder, in moments of despair, whether our concepts and theories give us any sort of grip at all on the world. We are told by outsiders that we are neither genuine scientists nor genuine humanistic scholars; that we not only use a hermetic and rebarbative jargon but that this jargon, unlike that of natural science, is also gratuitous because the propositions expressed in it are no more than glorified common sense. If questions of this sort have a particular poignancy for social scientists, there is by no means an absence of critical reflection on the natural sciences, as is proved by the immense interest in and development of the work of Thomas Kuhn. The social sciences have rarely generated paradigms with the sort of general authority that Kuhn’s theory of ‘normal science’ implies, but the notion that concepts, assumptions, measurement techniques and so on are linked together in a limited set of frameworks is only too familiar in the social sciences, where representatives of alternative traditions have characteristically ignored each other or talked past each other. Durkheim and Max Weber, those two contemporaries, are among the more striking examples. The endemic opposition between conflicting frameworks in the social sciences finds a particularly sharp expression in the differences between their terminology. Why is it, for example, that the term ‘society’, which plays such an important part in Durkheim’s sociology, is virtually absent from Max Weber’s work? (1) Why is it that the same term, ‘social structure’, is given a radically different sense by two different anthropologists: Radcliffe-Brown and Lévi-Strauss? For Radcliffe-Brown, ‘The (social) structure consists of the sum total of all the social relationships of all individuals at a given moment in time.’ (2) For Lévi-Strauss, by contrast: ‘The term “social structure” has nothing to do with empirical reality but with models which are built up after it.’ (3) The fact that differences such as these can be assimilated with apparent ease to traditional philosophical oppositions such as those between individualism and holism or between empiricism and conventionalism may lead the social scientist to see them as irresolvable and therefore to avoid any serious examination of them. Durkheim, for example, asserted robustly that ‘Sociology does not need to choose between the great hypotheses which divide metaphysicians,’ (4) and many sociologists have gone even further and argued, in a way which Durkheim would not have done, that these ‘definitional questions’ are arbitrary 2 Concept formation in social science and that the proof of the pudding lies in the predictive-explanatory power of the theories which are generated on the basis of some choice or other of initial assumptions. I shall argue in more detail later than this latter view is one of the more baneful influences of a mistaken philosophy of meaning and science, and that this attitude, at once permissive and dismissive, to alternative accounts of the nature of the social and of our knowledge of it is one cause at least of the intellectual and social disarray in which the social sciences find themselves. For if social scientists are inevitably pushed into taking serious notice of semantic aspects of their own practice, they are also compelled to adopt positions in the philosophy of meaning and science. They must, for example, support their choice of concepts and the ways in which they choose to specify them (by definition or otherwise), by metatheoretical reflections on the relation between scientific terminology and ordinary language, on the legitimacy or illegitimacy of stipulative definitions, and so on. More precisely, they can hardly avoid rejecting the more vulgar positivist positions on these issues—for example extreme empiricism or operationalism. (5) A more sophisticated view, which came to predominate within the positivist tradition itself, accepts that theoretical concepts cannot simply be reduced either to observational statements, in the manner of early logical empiricism, or to a set of operations. On this later view, it may be argued that the specification of individual concepts is relatively unimportant. What counts are the theoretical structures in which these concepts are combined. (6) But this view is itself open to an attack from within the philosophy of science which questions how far scientific theories are in fact deductively formalised and argues further that deductive formalisation as an ideal may be less appropriate to the sciences than had previously been assumed. (7) It can readily be seen that this argument, developed from within the second, ‘linguistic’ phase of analytic philosophy, will apply a fortiori to the social sciences, and indeed there is an important convergence between the analytic philosophy of the post-war period, developing out of logical positivism but increasingly opposed to it, and the earlier continental tradition of hermeneutics, (8) despite the latter’s concern with the human sciences to the exclusion of natural science. The hermeneutic tradition, which I have discussed in my previous book, (9) provides a much more adequate account of the place of language in the social sciences and, in particular, the way in which our access to social reality is mediated through language. There is, however, a problem within this tradition of establishing rational criteria for the choice of concepts and theoretical frameworks—a problem which it shares more generally with conventionalist approaches within the philosophy of science. Rationalist and realist philosophies of science, discussed in chapter 1 of this book, offer a possible solution to these difficulties, and I offer some arguments, there and in chapter 2, for the adoption by social scientists of a realist philosophy of science along the lines developed by Roy Bhaskar. (10) This approach also sheds a new light on the traditional opposition between positivistic and hermeneutic theories of society, the former stressing the essential similarity of the social and natural sciences, the latter asserting their radical distinctiveness. A realist philosophy of science makes it possible to reject positivist versions of naturalism, the thesis of the methodological unity of science, and to develop a new form of naturalism which does justice to the hermeneutic moment in the social or human sciences. (11) Introduction 3 Chapter 1 of this book presents the positivist, hermeneutic, rationalist and realist traditions in the philosophy of science and social theory. Chapter 2 develops the implications of these traditions for concept formation in the social sciences—in particular for the choice of concepts, the ways in which they can be defined or otherwise specified, and the relation between social scientific discourse and ordinary language. This chapter raises certain fundamental questions about the relationship between language, science and reality—in particular, the question of object-constitution to which chapter 3 is devoted. Here again the claim is made that only rationalism and realism furnish an adequate account of constitution, if we accept that critical theory has not provided sufficient backing for its own stronger account (and its claim that object-domains are constituted differentially by cognitive interests). Rationalist accounts of constitution are incomplete, in that they lack an adequate theory of society. Such a theory is needed to deal with the ontological sense in which the world is (partially) constituted by human practice; knowledge of this process is required for the reflexive reconstruction of the social base of acts of theoretical constitution, and hence the materialist grounding of theory in relation to past, present or future practice. Chapter 4 provides negative support for the realist programme by way of some critical remarks about Max Weber. These bear mainly on his metatheory, which was explicitly directed (in part at least) against something like the views advanced in this book. It is suggested that Weber’s arguments against these views are not conclusive, given certain antinomies in his own position. Chapter 5 examines positivistic conceptions of society, which either brush it aside or reify it as a substantive entity, and ‘transcendental idealist’ conceptions, loosely associated with hermeneutics, which reduce it to a quasi-transcendental postulate or at best a processual tendency. Finally, it is suggested that one particular attempt to develop a realist conception of society converges interestingly with other work in this area and furnishes an attractive research programme in sociological theory. (12) 1 Concepts of science Having sketched out, in general terms, the conceptual oppositions with which I am concerned, I shall now discuss in more detail three accounts of concept formation. These three (positivism, hermeneutics and various recent forms of rationalist and realist philosophy of science) represent clusters of views rather than clearly delimited and mutually exclusive positions. (1) I shall, in fact, argue that there is a close affinity between traditional Marxism and recent versions of realism and also that important elements of positivism are incorporated into verstehende social science, often in an unacknowledged form. However, I shall also suggest that the hermeneutic tradition encouraged a recognition of the importance of description, and a non-positivist account of explanation (2) which in turn have important affinities with realism. The following discussion ranges over both the philosophy of science and social theory, as these are conventionally defined. Two points should perhaps be made about this procedure. First, I do not believe that there is any clear-cut difference between the two. Although philosophy is primarily concerned with conceptual questions, and social theory is more concerned with empirical matters, this does not imply a sharp disjunction. Second, and related to the first point, I do not think it is possible to give any single account of the proper relationship between philosophy and social theory. The philosopher, or the scientist engaged in philosophical activity, is neither simply an ‘under-labourer’, nor a ‘master- scientist’, (3) nor an apologist for science. He or she may be all of these things, and more. Philosophy should be seen neither as an all-important foundation sine qua non of empirical science, nor as merely a scientific ideology. The concept of ideology is a helpful one here, but it must be used in this context without automatically pejorative connotations and in a way which admits that philosophy can make a genuine contribution to science. (4) 1 POSITIVISM The term ‘positivism’ is used in notoriously varied ways. (5) In subsequent chapters I shall use it in a rather broad sense, in which important aspects of the work of, for example, Weber and Schutz can be represented as positivistic. Here, however, I am concerned with the philosophical tradition which grew out of the Vienna Circle and flourished in the Anglo- Saxon countries of emigration. (6) This tradition is sufficiently well known not to require a general characterisation here, and the following discussion is heavily biased towards questions of language and of the logic of the social sciences. (7) The Vienna Circle’s attitude to language is best understood from the angle of its critique of ‘metaphysics’. This was the mainspring of the Circle’s simultaneous emphasis on and depreciation of language. (8) First, scientific theories were to be seen as systems of statements, to be analysed in the ‘formal mode of speech’; the philosophy of science was concerned with the syntax of the Concepts of science 5 language of science. (9) Second, this ‘linguistic turn’ (10) was simultaneously a rejection of much of the language of traditional philosophy, with its oppositions between realism, materialism, phenomenalism, idealism, and so forth. These no longer presented a choice between ontologies but merely between languages whose usefulness was in large part dependent on the particular context. Carnap reports that, while working on ‘Der Logische Aufbau der Welt’, (11) in discussions with friends he shifted freely between materialistic and idealistic, nominalistic and ‘Platonic’ kinds of language. ‘Only gradually, in the course of years, did I recognise clearly that my way of thinking was neutral with respect to the traditional controversies, e.g. realism vs. idealism, nominalism vs. Platonism (realism of universals), materialism vs. spiritualism, and so on.’ (12) Third, the distinction between good (scientific, logical) and bad (metaphysical) language was applied also to ordinary language. This contains unclear impressions which must be made more precise by being replaced by concepts. (13) At the same time, these must not be reified or rendered metaphysical. (14) The ‘search for a neutral system of formulae, for a symbolism freed from the slag of historical languages…for a total system of concepts’ (15) is primarily an attack on metaphysical errors within philosophy and empirical science, and only secondarily against the confusions of ordinary language. Neurath argued, for example, that the ‘purification’ of the concepts of the social sciences, although less advanced than in the case of physics, was also (16) less urgent perhaps. For it seems that even in the heyday of metaphysics and theology, the metaphysical strain was not particularly strong here; maybe this is because the concepts in this field, such as war and peace, import and export, are closer to direct perception than concepts like atom and ether. Fourth, it turns out that the ‘physical language’ proposed by Carnap as the reduction basis for the language of all science is itself reducible to ‘thing-language’, containing ‘those terms which we use on a prescientific level in our ordinary language’. (17) Underlying this programme of reductionism and linguistic purification was the Vienna Circle’s basic conception of an observation-language, grounded in observation, perception, experience and expressed in ‘protocol sentences’. Statements in the observation-language make up the base of our knowledge of the world; all other statements must be truth- functional compounds of these. Indeed, the meaning of a statement was identified with its truth-conditions, just as the meaning of a term was identified with its extension—the range of objects to which it applied. What logical positivism aspired to, in essence, was a harmonious fusion of the most advanced theoretical formulations in logic, mathematics and physics, and a relatively straightforward empiricism with a linguistic twist. (18) The origins of this marriage are well enough known; it will, however, illuminate the systematic questions to be discussed in the later chapters if we examine in some detail the strains which caused it to break up. The Vienna Circle began with a highly favourable view both of theory, which was conceived as a deductive system in the manner of mathematical logic, (19) and of empirical observation. (20) As the above quotations from Carnap indicate, it was at first believed that theoretical terms could be translated into observational terms without loss of meaning. Gradually, Carnap and others began to feel that this did not do justice to the specificity of what came to be called the theoretical language. In what Radnitzsky has called the ‘holistic