The Sense of the Past and the Origins of Sociology
Author(s): Philip Abrams Source: Past & Present, No. 55 (May, 1972), pp. 18-32 Published by: Oxford University Press on behalf of The Past and Present Society Stable URL: http://www.jstor.org/stable/650221 . Accessed: 03/04/2014 12:44 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . Oxford University Press and The Past and Present Society are collaborating with JSTOR to digitize, preserve and extend access to Past &Present. http://www.jstor.org This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY * SOCIOLOGY IS AN ATTEMPT TO UNDERSTAND THE DIRECTION OF CHANGE. More emphatically, it is a science of social development. This is not just a matter of its early entanglement with evolutionism. If one asks what sort of project sociology was for Max Weber or is for Talcott Parsons, if one seeks the underlying significance of the mass of apparently disconnected empirical work, or the common concerns of, say, Theodor Geiger and Raymond Aron, one discovers a diverse but sustained and remarkably coherent effort, first to identify indus- trialism as a type of society in contra-distinction to a pre-industrial type or types, and second to tell industrial man where industrialization is going. Every so often this main commitment of sociology appears to go underground, as when Durkheim thought he had buried Comte and Spencer by constructing a non-historical science of social facts. But on each occasion what was taken to be a grave turns out to have been only a tunnel, as when Durkheim himself reverted to a rampant historicism in order to explain the social functions of religion. In 1937 we find Professor Parsons firmly repudiating evolutionary interests in the opening sentences of his manifesto for a new, non- historical sociology: "Who now reads Herbert Spencer?" But by 1967 it is apparent that Parsons himself has been busy reading Spencer and is devoting his energy to just the old Spencerian quest for evolutionary universals and stages of development.1 Certainly Comte and Spencer would have felt quite at home in the plenary sessions of the 1970 World Congress of Sociology,2 in which the effort to understand the future by extrapolating tendencies from relationships presumed to exist between abstract models of past and present was taken as seriously as it had ever been in their day. What the discipline is after, in other words, is not just explanations of social behaviour, but tendentious explanations of social behaviour: "a science to teach the laws of tendencies", as Buckle put it.3 * This paper is based on my paper to the 1970 Past and Present Conference on "The Sense of the Past and History". Since then some of the material it contains has been used in my inaugural lecture, Being and Becoming in Sociology (Durham, 1972). 1 T. Parsons, The Structure of Social Action (New York, 1937); Societies: Evolutionary and Comparative Perspectives (New Jersey, 1970). 2 VIIth World Congress of Sociology, Varna, 1970, Transactions (Inter- national Sociological Association, 1971). S T. H. Buckle, History of Civilisation in England, i (London, 1857), p. 27. This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY I9 Whether such an enterprise is, in principle, philosophically or empirically viable is a matter for debate. Personally I think it is, and that it is just this emphasis in sociology which gives the discipline its importance, or at least its seriousness. Still, it can be argued that the ways in which sociology has so far gone about the explanation of tendency, or transition, have been flawed by a radically unsound methodology. And it can be argued that much of this unsoundness is rooted in the manner in which sociologists conceive of the past. Some conception of the past is inescapable. Sociology proceeds in its most typical forms by way of the typing of structural systems - for example, industrialism, feudalism, legal-rational authority. But if structuralism of this kind is to explain anything it must be by advancing explanations in terms of the principles of structuring, or of what Piaget in a stronger phrase calls the transformation laws of structures.4 Now it is plainly not the case that all structuring is chronological structuring: this would not be so for linguistic or mathematical structures for example. But it is necessarily the case in the fields of history, sociology and anthropology, the social sciences for which the idea of action in time is the essential element in explanation.5 Analysis of the mechanics of historical transition is the proper basic activity of the practitioners of these sciences. The only reference the idea of structural transformation can have here is a reference to historical process. Far from detaching social analysis from chronology, structuralism in the social sciences entails histori- cally grounded explanation. I would agree with Gellner that the resonance and appeal of sociology in recent years springs from the impression the subject gives of dealing directly with the mechanics of the transition that rightly concerns us most - industrialization.6 But I am not as confident as he seems to be that sociology has been attacking this problem in any particularly useful way. What seems to have happened, rather, is that structural types have been put together in a generally impressionistic and historically casual manner 4J. Piaget, Structuralism (London, 1971): "Were it not for the idea of transformation structures would lose all explanatory import, since they would collapse into static forms" (p. 12). 5 To this extent I would agree with W. G. Runciman (Sociology in its Place, Cambridge, 1970) that there can be no serious distinction between history, sociology and anthropology. But by the same token I disagree with his further claim that all three disciplines can be reduced to some sort of psychology. It is just their central emphasis on historical structuring that makes them non- reducible. "Men make their own history, but they make it in spite of themselves" (Marx) - it is their effort to understand the "in spite of" that gives these disciplines their autonomy. 6 E. Gellner, Thought and Change (Chicago, 1964). This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions 20 PAST AND PRESENT NUMBER 55 - consider the way in which bureaucracy and anomie were identified as emergent properties of industrialism by Weber and Durkheim for example. And secondly, logically ordered contrasts between struc- tural types have been treated, quite naively for the most part, as though they effectively indicated chronologically ordered transitions. On this basis a sociological past has been worked up, a past which is linked to the present not by carefully observed and temporally located social interaction but by inferentially necessary connections between concepts. Discussions of the decline of community, of the traditional working class and of the problems of modernization in the context of contrasts between "developing" and "modern" social systems are among the better known contemporary examples of the application of this mode of thought. In each case a perspective on present social experience is gained by postulating a tendentious relationship between what is observed now and a structural type associated firmly but unspecifically with the "past". The function of the sociologist's past in other words has not been to provide a frame of reference for empirical studies of the mechanics of transition but instead to furnish a rationale for side-stepping such tedious historical chores and moving at once to the construction of predictive interpreta- tions of the present. We have the odd spectacle of a discipline which claims importance just because it takes the problem of the temporal transformation of structure as its central analytical issue, but which at the same time appears committed to a sense of the past which actually directs attention away from the need for analyses of structural transition as a temporally and culturally situated process. Parsons's influential and representative essay "The Institutional Framework of Economic Development" is perhaps our best recent example of both sides of this ambivalence in sociology.' Unlike Rostow, Hoselitz and a number of others8 who can be said to use the idea of stages of development in a fairly mechanical way in producing scenarios of development policy, Parsons displays a good deal of refinement and subtlety in applying his model of industrialization to the predicament of the underdeveloped countries. He allows for example that the actual present of these countries is importantly unlike the past of the European countries as a result of the intervening history of the latter. Nevertheless, when all his refinements and modifications are made, the 7 T. Parsons, "The Institutional Framework of Economic Development", in Structure and Process in Modern Society (Glencoe, Illinois, 1960). 8 Cf. W. W. Rostow, The Stages of Economic Growth (Cambridge, 1962); B. Hoselitz, Sociological Factors in Economic Development (Glencoe, Illinois, 1960). This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY 21 problem of development remains one of adjusting the history of the underdeveloped countries to a model of structural transformation abstracted from European and American experience. Although he sees that, as a result of the time lag in industrialization, political institutions will be relatively more important in the underdeveloped countries than they were in Europe, the trajectory of industrialization remains essentially the same; the point of departure (traditional society) and the destination (industrial society) are treated as conform- ing in all important respects to a common model. For such a procedure to make any sense at all it must be assumed, as it is, that the pasts of the developed and underdeveloped countries are basically similar and basically unproblematic. Robert Nisbet, criticizing what he calls the metaphor of develop- ment in Social Change and History,9 and Andre Gunder Frank, criticizing what he calls the ideal-typical index approach to the study of transition in "The Sociology of Development and the Under- development of Sociology",lo have exposed some of the more startling consequences of this state of affairs. In this paper I want to consider causes rather than consequences, however. How is it that sociology has remained so unregenerate in its commitment to a sense of the past which we have been told again and again contributes more to ignorance than to knowledge? The paper is not meant to provide yet another occasion for historians to feel superior at the expense of sociology. The attempt to understand the mechanics of transition involved in structural change seems to me unquestionably more important than the sort of thing that normally goes on in most Departments of History. We have, too, enough examples of success in this sort of enterprise to know that the work is not in principle futile: the best example, I suppose, is the first volume of Capital. So it becomes a question worth asking why sociologists have been so unsuccessful in striking a fruitful balance between the typing of structures and the empirical analysis of transition - why they have for the most part felt that the need "to order structural types and relate them sequentially is a first order of business" - and have in proportion neglected the business of using structural concepts to inform historical investigation.11 The ordering of structural types is a relevant heuristic setting for the analysis of change. It cannot 9Published Oxford, 1969. 10 A. G. Frank, Latin America: Underdevelopment or Revolution (New York, 1964). The essay cited is also published separately, New York, 1968. 11 T. Parsons, Societies, Evolutionary and Comparative Perspectives, p. III; and see Nisbet, op. cit., ch. 8. This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions 22 PAST AND PRESENT NUMBER 55 be a substitute for it. How did sociologists come to adopt an idea of history which so directly implied the opposite? We may start from John Burrow's observation that the social sciences were in the first instance a response to anarchy: "social anarchy as a fear, intelletcual anarchy as a fact".'2 More importantly perhaps the social and cultural confusion of the time was understood not as an effect of wickedness (as a comparable disorder had been understood in the seventeenth century), but as an effect of history. The sense of disorder was ubiquitous and acture. Its intensity was such that many felt unable to say, at even the most modest level of abstraction, what was going on. The predicament was well described by Lamartine in his account of wht it was like to live through the last months of the July Monarchy: These times are times of chaos; opinions are a scramble; parties are a jumble; the language of new ideas has not been created; nothing is more difficult than to give a good definition of oneself in religion, in philosophy, in politics. One feels, one knows, one lives, and at need one dies for one's cause, but one cannot name it. It is the problem of the time to classify things and men. The world has jumbled its catalogue.13 But the collapse of meaning had in addition a specifically historical content. Eric Hobsbawm has drawn attention to the propensity in all societies to use the past as a resource for either anticipating or prescribing the future.14 It was precisely the possibility of such thought that the pace and scope of change in the mid-nineteenth century seemed to undermine. The sense of the meaninglessness of the present was felt as a matter of the lack of relationship between present and past. The generation that gave birth to sociology was probably the first generation of human beings ever to have experienced within the span of their own lifetime socially induced social change of a totally transformative nature - change which could not be identified, explained and accommodated as a limited historical variation within the encompassing order of the past. One faced for the first time a situation in which the idea of historical action or accident - conquest, revolution or plague - could not begin (so it seemed) to account for the ways in which the present differed from the past. To act effectively in the present, a frame of reference which allowed one to identify the structure of one's situation, and so to anticipate the consequences of one's actions, was essential. But such a frame of reference could not be derived directly from the study of 12 J. Burrow, Evolution and Society (Cambridge, 1966), p. 93. 13 Cited by C. Geertz "Ideology as a Cultural System" in D. Apter (ed.), Ideology and Discontent (New York, 1964), P. 43. 14 E. J. Hobsbawm, "The Social Function of the Past", above pp. 3-17. This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY 23 the present - the world had jumbled its catalogue. Nor could it be derived naively from knowledge of the past because the nature of the historical connection of past and present had become obscure, the conventional categories of historical thought could not grasp it. Hitherto the past had provided the pattern for the present in quite straightforward ways. History had been an unproblematic matter of recording duration and succession. Neither duration nor succession had appeared to bring the nature of the principles of social organiza- tion into question. But that was just what happened in the mid- nineteenth century. One did not need to be very sophisticated to feel that the present when considered in relation to the past was deeply enigmatic. The merchants and landlords who joined together to form the Bristol Statistical Society in 1838 were driven to an interest in social research by motives not very different from those which were to inspire Durkheim or LePlay. In a simple state of society [they noted] a man may know tolerably well what his duties to the poor are . . . but what shall be said of that artificial and complicated state of things when a nation manufactures for half the world - and when the consequence unavoidably is the enormous distance between the labourer and his virtual and sub-divided employer?'5 The rapid and amazingly ramified extension of the division of labour was the beginning of the problem. But layer upon layer of complica- tion had been heaped upon it until all effective sense of historically anchored process was lost. Even Bagehot, the least flappable thinker of his generation, sensed the dilemma. "The greatest living contrast", he was moved to remark in 1861, "is between the old Eastern and customary civilizations and the new Western and changeable civilizations".16 What resources were available for making sense of the experience of living in a changeable civilization? Only knowledge of the past. Somehow that knowledge had to be used to yield up a new under- standing of what was happening in the present. G. H. Lewes expressed the problem very clearly. Like most of his contemporaries he found himself in "an age of universal anarchy of thought", an age '"anxious to reconstruct . . . but as yet impotent" - impotent because the anarchy was historically induced and historically incompre- hensible. "In this plight", he concluded, "we may hope for the future but can cling only to the past: that alone is secure, well- grounded. The past must form the basis of certainty and the 15 J. of the Statistical Soc., ii (1839). 16 W. Bagehot, Physics and Politics (London, 1872), P. 114. This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions 24 PAST AND PRESENT NUMBER 55 materials for speculation"."1 In turning to the past, then, the intention was somehow to transcend mere history. Here the emerging social sciences faced a fundamental strategic choice. Was the past to be understood as a structural system or as a field of history ? Because the most urgent issue was to identify the general organizing principles of industrial society and the general principles of change involved in industrialization, it was perhaps natural that the historical character of the past should in the first instance have been ignored, that the first response should have been a set of attempts to reify both the past as a structural type and history as a developmental process. What was not so natural, but nevertheless happened in almost every case, was that this intitial elaborate construction of ideal types did not lead social scientists back to substantive investigations of historical transition in particular settings but was allowed to stand as being in itself a theoretically and empirically adequate alternative to such investigations. Even Max Weber in The Protestant Ethic and the Spirit of Capitalism,18 the nearest thing to an example of good histori- cal sociology which the founding fathers of the discipline were to produce, was astoundingly casual about the detailed historical valida- tion of his argument. One of the things which makes it so difficult for students to answer the standard examination question which invites them to compare Weber's account of the development of capitalism with that of Karl Marx is that Marx is simply a much better a historian than Weber. Marx was, of course, always primarily interested in the mechanics of transition, the relational basis of industrialization. By comparison the construction of developmental types has a second-order, even a background, importance in his thought. Nevertheless it is strange that sociologists in general should not have been led as Marx was from the reification of the stages and processes of development to the sort of empirical historical sociology Marx himself achieved. We can hardly explain the failure by suggesting that the sociologists were work-shy. On the contrary the important nineteenth-century sociologists were at least as industrious as Marx. It is possible that Spencer accumulated more data than any other scholar has yet done. Nor were the early sociologists disinclined to handle historical data - Weber for one seems to have had an inexhaustible interest in such 17 G. H. Lewes, "The State of Historical Sciences in France", cited in Burrow, op. cit., p. 94. 18 M. Weber, The Protestant Ethic and the Spirit of Capitalism, first published in Archiv fur Socialwissonschaft und Social politik, vols. xx and xxI (1904-5), English translation by T. Parsons (New York, 1930). This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY 25 material. Generally it was through the reinterpretation of historical materials that they hoped to achieve an understanding of the meaning of the present. What is odd is that they remained committed to ways of using historical materials that were both ahistorical and historicist. It is this ahistorical historicism of sociology that needs to be explained. The explanation seems to have two main elements. First there was the intellectual ascendancy of evolutionism. Second one must recognize the apparent power of the analytical paradigm produced by the treatment of the past as a structural type. It did permit as, Marxism apart, nothing else did, a generalized account of the structure and tendency of industrialism. An exhaustive explanation would also have to consider the importance of some questions of academic convenience and convention. In establishing its own academic credentials, sociology had above all to differentiate itself from history. Since it, too, dealt in historical materials and problems, the differentiation virtually had to be in terms of sociology's special methodology. Once methodology became the hallmark of the discipline at this level it was surprisingly easy for it to prove an obstacle to the adoption of new ways of dealing with the problems as well. It is bizarre but not unrevealing that we should observe attempts to demonstrate that Stanley Elkins is "not really" a historian or that Barrington Moore Jr. is "not really" a sociologist.19 But this is by the way. As an empirical science of the laws of tendency, sociology sprang directly from the sense, pervasive and disturbing as it was, of a changeable civilization. Either changeability made civilization unpredictable - a prospect not even Herbert Spencer was sanguine enough to embrace - or it was scientifically ordered in ways which appropriate contemplation could reveal. Appropriate contemplation in turn was felt to involve three things: first, the discovery of a conceptual language capable of differentiating between present and past, of marking out the trajectory of change; then a general structural characterization of the present as distinct from the past; and finally the identification of the processes of change or growth in terms of which past, present and future were bound together. For each of these purposes it was not historical action but objectified historical 19 S. Elkins, Slavery (Chicago, 1959) and Barrington Moore Jr., The Social Origins of Dictatorship and Democracy (Benson Press, Boston, 1966) are among the better known recent studies to have created intra-disciplinary soul-searching by demonstrating the unavoidably inter-disciplinary nature of explanation in the social sciences. For current examples of pedantic boundary disputes of this kind, see The American Sociologist, vi (New York, 1971). This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions 26 PAST AND PRESENT NUMBER 55 process that was of interest. The idea that process could be ascertained only through careful observational studies of action occurred to very few people. In passing we might note that the problem of putting together a suitable language of concepts was itself an acute one. Few things are as evident in early nineteenth-century social analysis as the want of appropriate terms to specify the variations in social experience that observers wished to discuss. The vocabulary that served to describe traditional social relationships simply could not grip the experience of the present with any precision. Compare the vigour of the first part of this statement by Cobbett with the limpness of the end: "When master and man were the terms everyone was in his place and all were free; now in fact it is an affair of masters and slaves".20 Now of course it was not really an affair of masters and slaves. But Cobbett's repertoire of concepts simply could not get him any nearer. Nor was it sufficient to see the present simply as a negation of the past: Shelley's string of negatives - "sceptreless, uncircumscribed, unclassed, tribeless and nationless, exempt from awe, worship and degree"21 - was a good intuitive response to the situation but no basis for analysis. In the event the vocabulary problem was solved under the umbrella of the general attempt to characterize the present as a type of social order, and to infer from the supposed typological properties of types of social order supposed laws of tendency or principles of social development. The overriding necessity was to obtain an objective, abstract yardstick outside the flux of the present situation - the complicated and artificial state of things - to which the present situation could be referred and in terms of which it could thence be known. To this end the emerging social sciences seized hold of history in two ways. First in the form of a series of bold conceptual polarities, explicit antitheses between past and present which Nisbet has called the unit ideas of sociology. Second in the form of a set of ambitious descriptive theories of the stages of social development. The effect of both procedures was to turn history into an object.22 20 W. Cobbett, Political Register, lxxxvi (London, 1835) p. 767. 21 P. B. Shelley, "Prometheus Unbound", Act III, Scene iv, The Complete Poetical Works (Oxford, 1907). 22 R. Nisbet, The Sociological Tradition (New York, 1966). If one were disposed to accept the argument that the principal property of the culture of capitalism is a process of reification in which all secondary relationships tend increasingly to be perceived as relationships between things, one could then add to Engels's analysis of the way in which the real connectedness of man and (cont. on p. 27) This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY 27 The attempt to formulate laws of development as a matter of explicit historical process was of course a conspicuous failure. Its empirical difficulties were quickly apparent to most observers. Thus Henry Sidgwick observed in I885 how: With equal confidence history is represented as leading up, now to the naive and unqualified individualism of Spencer, now to the carefully guarded and regulated socialism of Schaeffle, now to Comte's dream of securing seven- roomed houses for all working men .... Guidance, truly, is here enough and to spare; but how shall the bewildered statesman select his guidance when his sociological doctors exhibit such portentous disagreement? Not surprisingly Sidgwick ended by begging his audience "to take no steps calculated to foster delusions of this kind".23 The more important epistemological difficulties of evolutionary sociology were no less effectively exposed, first by would-be evolutionists such as Hobhouse and Ginsberg, then definitively by Popper.24 Two years after Popper's first onslaught on sociological historicism Parsons proposed the repudiation of all interest in diachronic analysis and the reorientation of sociology around the synchronic investigation of systems of action in terms of formalized ahistorical properties.26 What actually happened at this point, however, was that, although the discrediting of the overt intellectual strategies of evolutionism was acknowledged, the infrastructure of evolutionism remained embedded in sociological thought. It was here that the conceptual polarities of sociology's unit ideas were important. Status and contract, community and association, organic and mechanical solidarity, traditional and legal-rational authority, the folk community and the urban community - all these double concepts were ways of trying to apprehend and identify the changes in the structural format of society associated with industrialization. More or less explicitly the changes indicated in the conceptual antitheses were treated as necessary concomitants of industrialization, an idea which surfaced from time to time (most recently in the work of Clark Kerr and his colleagues in the I96os) in the notion of the "logic of industrialism".26 There could be and was wide-ranging dispute as to the exact nature of (note 22 cont.) his history is "lost for fair" in the veils of fetishism spun by philosophers, political theorists and jurists the observation that the peculiar contribution of the sociologist to this process has been, as a final ironic transformation, to turn history itself into a thing. 13 British Association for the Advancement of Science, Proceedings (London, 1885). 24 L. T. Hobhouse, Social Development (Allen and Unwin, London, 1924) and Morals in Evolution (Macmillan, London, I901); M. Ginsberg, The Diversity of Morals (London, 1956); K. Popper, The Poverty of Historicism (London, 1957). 2r T. Parsons, The Structure of Social Action (Glencoe, Illinois, 1937). 26 C. Kerr, et al., Industrialism and Industrial Man (London, 1962). This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions 28 PAST AND PRESENT NUMBER 55 the logic of industrialism but the salience of the idea as a governing focus of thought remained strong. Not all of the early sociologists adopted the device of conceptual polarization in its fullest form. Often the polarity was merely implied in the assertion of some distinctive processual property of industrialization: the emergence of chronic anomie, urbanism as a way of life, bureaucratization, secularization, the isolation of the conjugal family. But the procedure is really the same. It is a matter of abbreviating history. It involved the observation of key structural differences in the constitution of the present as distinct from the past. But it did not necessarily involve any need to show how, historically, the differences had been effected. It was the observation of contrasted moments of development that mattered. Having characterized past and present as states of being in terms of some key properties, one could go on to infer laws of tendency by logical rather than historical procedures. Whatever the difficulties of the method, its sheer economy was among its principal attractions. Quite simply, there was no quicker method of producing a theoretical account of where society was going or of what were its significant structural components. It did matter, of course, to show that the past postulated by sociology - the world of the extended family, of community and corporation, of folk culture and traditionalism - had been really there in some concrete sense. But to see how this was done is to see still more clearly how profoundly unhistorical the whole enterprise really was. The point after all was not to know the past but to establish an idea of the past which could be used as a comparative base for the understanding of the present. Once the flood of ethnographic data became available and once it became clear that the Iroquois, the ancient Picts and the Irish in Manchester were, analytically, the same thing, the essential irrelevance of history in the construction of this past was revealed. This did not, of course, at all reduce the importance of calling it the past. That importance was irreducible. But it sprang from the sociologists' concern to achieve a theory of modernity, and if possible of modernization, not from any interest in the mechanics of historical transition. As J. F. McLennan put it in a general rubric for the social sciences with which most of his co- workers seem to have been thoroughly sympathetic: The first thing to be done is to inform ourselves of the facts relating to the least developed races ...their condition, as it may be observed today, is truly the most ancient condition of man. It is the lowest and simplest... and ... in the science of history old means not old in chronology but in struc- ture. That is most ancient which lies nearest the beginning of human progress considered as a development.27 27 T. F. McLennan, Studies in Ancient History, 2nd ser. (London, 1896), p. 16. This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY 29 There are few clearer statements of the central strategy of the social sciences - and few more indicative of their indifference to anything that could be called, strictly, the historical past. Long after McLennan's hopeful involvement with overt notions of progress and development had been abandoned, the conceptions implicit in the idea of systems being "old in structure" remained rooted in sociological method. Some consequences of this method are worth noting. It is not just that it directs attention away from the need for propositional theories about the organization of change in particular historical contexts; or that it permitted people like Bagehot to regard the working classes as "primitive" ;28 its economy, elegance and apparent effectiveness in differentiating past and present have encouraged a state of affairs in which a high proportion of sociological research is in fact research on myths which sociologists have invented. The sociology of the family provides some lovely examples of this process. Family sociology has until quite recently been dominated by the idea of the classical pre-industrial family, or, as W. J. Goode puts it, "a pretty picture of life down on Grandma's farm". With reference to this construct, assembled by means of McLennan's brand of structural history and the skilful extrapolation from it of ideal types, a whole series of quite detailed myths were formulated about what happens, and has to happen, to the family in the course of indus- trialization. Goode, who has been more involved than anyone else in the dismantling of this particular body of myth, now concludes that no determinate relationship can be established either way between family patterns and industrialization.29 This, however, is not so much a definitive finding as a statement that the ground is now clear for the sort of research that ought to have been done in the first place. Meanwhile an expensive research unit in Cambridge has devoted several years to proving the non-existence in pre-industrial England and elsewhere of a type of family which no-one familiar with the historical evidence ever said did exist.3" This sort of thing is the least of the costs of sociology's hidden ahistorical historicism. The higher costs are paid in the terms of reference embodied in whole strategies of sociological thought. A case in point would be the use 28 Bagehot, op. cit., pp. 82-5; cf. Nisbet, Social Change and History. 29 W. J. Goode, "Industrialisation and Family Change" in B. F. Hoselitz and W. E. Moore (eds.), Industrialisation and Soczety (New York, 1963), PP. 237-59. 30 T. P. R. Laslett, The World We Have Lost (London, 1965), pp. 8I-Io6. This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions 30 PAST AND PRESENT NUMBER 55 that has been made of Professor Parsons's influential proposal to analyse social action in terms of a scheme of pattern variables.31 The pattern variables appeared as an integral part of Parsons's manifesto for a new sociology twenty-five years ago, the attempt to reconstitute sociology as an analysis of the structure of social action dissociated from the study of tendency. In pursuit of this object Parsons proposed that orientations to action could be investigated schematically in terms of a limited number of pure types. He recommended that these types should be organized in four or five pairs of opposities. The four pairs of pattern variables (variable ways of patterning action) for which he found most use were identified as follows: particularism versus universalism; affectivity versus affective neutrality; ascription versus achievement; and diffuseness versus specificity. This set of variations is offered as encompassing, if not the full range of possible modes of action, at least such a large field that effectively all systems of action can be brought within the scope of sociological analysis. The merit claimed for the pattern variables as analytical tools in other words is precisely that they are independent of, they rise above, any particular historical context. They are quite simply value-free tools. Yet the use that has been made of them, in part by Parsons but more especially by some of his followers, makes this hard to believe. It turns out that they do have, again in the structural sense, a reference to history or at least to the difference between past and present, traditionalism and modernity, after all. Thus Sutton, Hoselitz and many others have identified the difference between modern and pre-modern social systems as a polarity of universalism, affective-neutrality, achievement orientation and functional specificity on the one hand and of particularism, affectivity, ascription and functional diffuseness on the other.32 Whether Parsons intended his polarities to serve the turn of socio- logical historicism in this way is not clear. His categories plainly are anchored in quite familiar contrasts between the presumed properties of industrialism and pre-industrialism, however, and the use that has been made of them is in this sense legitimate. They do serve as one more device enabling sociology to theorize about 31 T. Parsons, The Social System (Glencoe, Illinois, I95I). Parsons's con- structs are of course an explicit extension of Weber's distinction between the properties of "traditionality" and "rationality": M. Weber, The Theory of Social and Economic Organisation (New York, 1947). 32 Frank, Latin America: Underdevelopment or Revolution, discusses this procedure at some length. F. X. Sutton, "Social Theory and Comparative Politics" in H. Eckstein and D. Apter (eds.), Comparative Politics (Glencoe, Illinois, 1963). This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions THE SENSE OF THE PAST AND THE ORIGINS OF SOCIOLOGY 31 the course of development without reference to the mechanics of transition. Consider a final example. The literature of contemporary sociology is full of general characterizations of advanced industrial society - as mass society, the acquisitive society, the affluent society and most recently the chaotic society. Most particular research projects proceed under the intellectual auspices of one or other of these characterizations. None of the characterizations is the result of scholarly historical analysis. All of them depend, however, on the actuality of an assumed historical process. Daniel Bell's account of the theory of mass society provides a good example of what is involved: The revolutions in transport and communications have brought men into closer contact with each other and bound them in new ways; the division of labour has made them more interdependent; tremors in one part of society affect all others. Despite this greater interdependence, however, individuals have grown more estranged from one another. The old primary ties of family and local community have been shattered; ancient parochial faiths are questioned; few unifying beliefs or values have taken their place. Most important the critical standards of an educated elite no longer shape opinion or taste. As a result mores and morals are in constant flux, relations between individuals are tangential or compartmentalized rather than organic. At the same time greater mobility, spatial and social, intensifies concern over status. Instead of a fixed or known status symbolized by dress or title, each person assumes a multiplicity of roles and constantly has to prove himself in a succes- sion of new situations. Because of all this, the individual loses a coherent sense of self. His anxieties increase. There ensues a search for new faiths. The stage is set for the charismatic leader, the secular messiah, who by bestowing upon each person the semblance of necessary grace and of fulness of personality supplies a substitute for the older unifying belief that the mass society has destroyed.33 Whether or not this type of characterization, which is quite prevalent in sociology, is based on good history or not is not immediately relevant. The important feature of such thinking is that in it the characterization of historical process and the characterization of present structure are totally interdependent. Each pervades the other and the conception as a whole is inconceivable without both. All questions of how the various transformations entailed in the movement between structural types were effected are, however, firmly set aside. The point is not to focus investigation on the social organization of historical process but to set up a frame of reference for research on a thing called the social structure of the present. And yet structure is defined in terms which have meaning only in terms of conceptions of process. We are faced with the same paradox 33 D. Bell, The End of Ideology (Glencoe, Illinois, 1960). Professor Bell is not, of course, espousing the theory of mass society in this passage. His exposi- tion of it is nonetheless well-taken for that. This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions 32 PAST AND PRESENT NUMBER 55 as before: the identification of structural types, the formal differentia- tion of past and present, is effected with such elan and internal cogency that it ends up by apparently making unnecessary any further study of the intervening structuring through which the past presum- ably became the present. Yet, of course, it is only such work that will tell us whether our structural concepts make sense, let alone whether they explain anything. The academic and intellectual dissociation of history and sociology seems, then, to have had the effect of deterring both disciplines from attending seriously to the most important issues involved in the understanding of social transition. Many current accounts of the historian's past, requiring as they do a wholesale rejection of any form of structural analysis, strike me as no better suited than the normal version of the sociologist's past to deal with these issues. This is not the place to consider what changes of heart or shifts of emphasis would be needed to produce a more fruitful and sociological history. What I have tried to do is to show how one could begin to move towards a more penetrating historical sociology. The essential step is not to abandon the structural typing of past and present but rather to recognize that the function of structural types is not to allow us to by-pass history by inferring logically necessary tendencies, but on the contrary to direct attention to those kinds of historical inquiry which we should expect, theoretically, to explain phenomena of structural transformation. University of Durham Philip Abrams This content downloaded from 128.135.12.127 on Thu, 3 Apr 2014 12:44:56 PM All use subject to JSTOR Terms and Conditions