1The exchange in the American Journal of Sociology [2] between the neoclassical economics-inspired sociologist James S. Coleman on the one hand, and the historian-sociologist William H. Sewell, Jr., influenced by Geertz’s cultural anthropology, on the other, is an opportunity to take note of the rise, or the return to prominence, of two influential currents in the American sociological field: the theory of “rational action” (Rational Action Theory, or RAT) and historical and cultural sociology. Beyond the men, two epistemological poles and two conceptions of social action and science are, in fact, in conflict.

Coleman and the offensive of homo economicus

2The RAT that Coleman calls for in his manifesto article does not, strictly speaking, constitute an original theoretical movement in the American social sciences, but rather an enlarged and modernized application of the “economic approach” to social phenomena, whose historical roots go back to the individualistic philosophy of Anglo-Saxon utilitarianism, [3] and whose founding postulate was stated by Gary Becker (1976, p. 14) as follows:

3

“All human behavior can be conceived as involving participants who maximize their utility from a stable set of preferences and who accumulate the optimal amount of information and other inputs in a variety of markets.”

4Under the various names of rational choice, [4] the problem of collective action, [5] the search for “microfoundations” [6] or methodological individualism, this approach is currently enjoying an unprecedented popularity in the United States, which can be explained by the conjunction of several factors.

5The first factor is the “microsociological” or “constructivist” revolt that put an end to the functionalist hegemony in the 1960s. The paradigmatic breakdown of American sociology that followed authorized a return to the actor, which, if it was mainly carried out in the “interpretative” or cognitive mode, with the symbolic interactionism of Blumer and Goffman, the ethnomethodology of Garfinkel and Cicourel, phenomenological sociology inspired by Schutz and conversation analysis of Sacks and Schegloff, [7] also includes an “instrumentalist” or rational variant represented by the behaviorism of George Caspar Homans and the exchange theory of Peter Blau. This shift in the center of gravity of the sociological field towards the subjectivity pole is combined today with a clear revival of interest in theory – even in “Grand Theory” – so as to create particularly favorable conditions for the propagation of RAT, in which it is not forbidden to discern a sort of return of the repressed utilitarian sociology that had gradually dissipated over the course of the 1970s. [8]

6The second factor is a full-scale attack on the sociological citadel by economists and their allies from within. Adam Przeworksi observes that :

7

“The social sciences are today being attacked by an offensive without precedent since the 1890s: a deliberate push to impose the monopoly of the economic method on any study of society. […] And this offensive has been largely victorious.” [9]

8Without going so far as to share this self-interested and somewhat hasty diagnosis of the outcome of the battle, one can only agree in noting the rapid proliferation of economic models in American social science [10]. This is the case in particular in the sociology of organizations, where the work of economists has been at the origin of important developments, such as the transaction cost approach or the theory of “principal-agent” relationships, [11] but also in the study of social movements, the family and the school, as in the scattered body of research and analytical techniques gathered under the label of network theory. [12] Admittedly, there are relatively few avowed disciples of RAT among sociologists – in contrast to political scientists – and the intrusion of the economic method into the field of sociology has not happened without triggering a vigorous counter-attack. Thus the neo-classical concept of the market is today, if not discredited, at least strongly contested by the work of Granovetter, Powell and White, and the rapid rise of economic sociology in recent years suggests that the sociological questioning of the notions of orthodox economics is only just beginning. [13]

9The fact remains that, reinforced by the social philosophy of “voluntaristic individualism” which, since the beginning, has permeated American sociology, and by the instrumental positivism which makes the successes – social if not scientific – of neo-classical economics the model of scientificity, [14] the scheme of rational action now underlies, even if implicitly, a good number of apparently atheoretical research sectors. In any case, the economist offensive is sufficiently threatening for some sociologists to be moved by it and to warn their colleagues of the dangers of giving in too easily to a fusion of sister disciplines under the aegis of the reductive concept of rational choice: sociology would have nothing to gain by trading its sense of the variety and complexity of reality for the “unrealistic neatness” of economic models. [15]

10Two examples of this epistemological arrogance of economists: At a symposium on Paradigms in the Social Sciences held in 1986 on the campus of the University of Chicago – a stronghold of RAT if ever there was one [16] – the Nobel Prize-winning economist James Buchanan did not hesitate to proclaim that (quoted by Hirsch et al., 1987, p. 317):

11

“Those who prefer to conduct research on class relations, states and organizations as such without taking the trouble to reduce the analysis to the behavior of the individuals who participate in them do not, in my opinion, deserve the title of social scientist, if the term is to have any meaning at all.”

12In a conference on sociological theory at the University of Chicago in 1983, his colleague Mancur Olson responded to Joseph Gusfield, who had just defended interpretive sociology (“humanistic science”) at the end of a paper on the social construction of the notion of a public problem such as drinking and driving, with these words (Olson, in Lindenberg, Coleman and Nowak, 1986, p. 202):

13

“I liked the talk and even found it enjoyable. When it was over, I wondered why I liked it so much […]. I finally realized why: it was like a beautiful sermon. Shouldn’t you then call it a theological approach to social science rather than a humanistic one?”

14On the side of sociologists, Michael Hechter’s book Principles of Group Solidaritv offers an archetypal example of this creeping economism that dares not say its name. Having, after a summary trial, condemned without appeal the “normativist, structuralist and functionalist” approaches (1987, p. 20-29), simplified at will for the occasion, this former follower of neo-Marxist theory (he was, at other times, one of the most prominent disciples of Immanuel Wallerstein at the University of Columbia), freshly converted to the RAT, proposes to replace them with a rational theory of group cohesion. This theory is articulated around three basic elements which immediately give an idea of the conceptual impoverishment produced by such a conversion: individuals “bearers of a set of discrete, unambiguous, and transitive preferences”; institutions, i.e. “sets of rules which effectively constrain individual behavior in various ways”; and “collective products” resulting from the combination of individual behaviors (p. 30-31). Question: why do individuals believe in or join groups? Answer: in order to be able to consume the exclusive and immanent goods that these alone are able to produce. At the end of a close examination of several cases, parliamentary groups, credit associations, internal labor markets of firms and communes (a very particular sample since it contains only formal elective groups. What about all the collectives where membership does not come from the logic of deliberate and conscious individual choice, families, clans, classes, nations, racial or ethnic communities, sexual groups or age groups?), the author arrives at this remarkable discovery full of utilitarian common sense: a group is all the more solidary the more its members depend on it and the more its capacities of formal control and sanction are extended.

15In conclusion, Hechter agrees that RAT suffers from some alarming shortcomings: it is unable to explain why one type of group emerges rather than another in order to quench the same unquenchable thirst for ‘immanent collective goods’ that endlessly torments maximizing individuals; it denies any role to collective identities and feelings; it proves incapable of shedding light on the dynamics of institutional development and the outbursts of collective action that occur in the apparent absence of any formal mechanism of control or retribution; and lastly, and above all, it remains totally silent on the causal variable which, supposedly, controls the whole model, namely the formation of individual preferences (p. 183-184). This does not, however, prevent him from asserting without further formality that “far from being immiscible, sociology and rational choice are mutually complementary… Together, [they] offer today the best hope of merging the individual and structural levels of analysis into a coherent whole capable of producing falsifiable empirical propositions” (p. 8 and p. 186).

16Finally, although rooted in a fundamentally conservative individualistic social philosophy, and even organically linked to a form of libertarian capitalism which would seem, at first sight, to make it antithetical to a critical social science, the theoretical claims of sociological RAT have been considerably reinforced on the left by the parallel emergence of Analytical Marxism (also known as game-theoretic Marxism or rational choice Marxism). This new school of thought, whose hard core consists of the philosophers Gerald Cohen and Jon Elster, the political scientists Adam Przeworski and Robert Van Parijs, the economist John Roemer and the (only) sociologist Erik Olin Wright (and to the periphery of which one can attach the philosophers Andrew Levine and Allen Wood, historian Robert Brenner and Norman Geras) aims to take up the problems of Marxism using the methodology of neo-classical economics. He does it on the basis of a philosophy of action – in fact a real social ontology related to that which founds Sartrian subjectivism [17] – which makes the structure of social relations the product of the composite and unintentional aggregation of the deliberate choices of individuals maximizing their interests during strategic interactions aiming for the efficient distribution of limited resources. [18]

17This heterogeneous and small (but extraordinarily active and cohesive, and therefore very visible) network that claims to be a hybrid between Marx and John Stuart Mill has forged the project of finally giving Marxism the logical and rational (to say the least: scientific in a sense that the Vienna Circle would not deny) foundations that, a century after the death of the author of Capital, it still lacks. [19] Its members distinguish themselves from their predecessors, notably the structuralists, of whom some are repentant defectors, by the fact that they apply without any qualms “the most advanced methods of analytic philosophy and ‘positivist’ social science” [20] to the traditional problems of Marxism – theory of history, exploitation, class conflicts, transition to socialism. That is essentially the theory of games, mathematical modeling and the concepts of equilibrium and market of neo-classical economics. One of its principal spokesmen does not hesitate to go so far as to set up analytical philosophy as a sine qua non of the scientific method. [21]

18This “marginalist historical materialism”, if we may be allowed to use an expression that borders on conceptual barbarism, thus agrees with the sociological RAT in making the rational choice of the individual the Archimedean point of social theory. Indeed, Analytical Marxism is in every respect the counterpart, in the sub-field of “professorial Marxism,” [22] of Coleman’s rational choice theory in the sub-field of orthodox sociology. Just as Colemanian RAT cannot be understood outside of its oppositional relationship to functionalism and its hypersocialized conception of action which, in the Parsonian scheme, makes the agent the passive servant of all-powerful consensual cultural norms (Wrong, 1961), so in the same way the rationalist and intentionalist hypersubjectivism of Analytical Marxism is defined in a reaction to the exacerbated anti-humanism of Althusserian Marxism for which the agent is reduced to the rank of support of reified and quasi-anthropomorphized structures, continuously interpellated by ideological apparatuses with infinite mystifying capacities.

19The RAT current is not yet sufficiently widespread, nor sufficiently homogeneous, to claim a large audience among sociologists as a theory, [23] and it is too early to predict whether it will continue to spread or, on the contrary, enter a phase of decline – as certain precursory signs lead us to suppose. [24]

20It is in this triple context – resurgence of microsociology and theory (including the micro-macro debate), invasion of economic models and invention of Analytical Marxism – that Coleman’s article is set. A true proclamation of a “theoretical state of emergency” in favor of utilitarian individualism, [25] it illustrates in a paradigmatic way this forceful introduction of the homo economicus in the heart of the American sociological field. However, the Chicago sociologist is not content with simply repeating the postulates of economic analysis: he strives to give the RAT a noble theoretical filiation by linking it to the Parsonsian work, even if it means a dubious reinterpretation of the history of sociological thought in the United States; he makes an original diagnosis of the origins of the growing divorce between theory and empiricism; finally, he suggests several ways of enriching and making the economic model more complex by introducing the missing notion of social structure in various forms.

21Coleman’s first objective is to give utilitarian individualism the titles of sociological nobility which, since Durkheim, it has sorely lacked by recreating an intellectual genealogy that links it to that totemic figure of theory that is, inevitably, Talcott Parsons. [26] His plea for RAT thus opens with a historical reminder of the disappointed promises of American sociological theory: if the scientific program announced in 1937 by Parsons in The Structure of Social Action did contain the premises of a “voluntarist” theory of action, [27] neither he, nor his epigones, nor his rivals were able to bring them to fruition. Unable to link the subjective aims of the individual and institutions, Parsons himself very quickly took refuge in the study of the social system and its preconditions. George Homans, for his part, went in the opposite direction and, completely losing sight of social structure and intentionality, came to reduce the sociology of action to a behaviorist analysis of exchange à la Skinner; Merton, finally, if he partially absolved functionalism of the original sin of teleology, did not succeed in giving social action those rational “microfoundations” which it lacked.

22From this theoretical deficiency was born a growing schism between theory and empiricism. For at the moment when the individual – or more precisely the notion of deliberate choice as a modus operandi of individual practices, which Coleman holds, as we shall see, to be consubstantial to a theory of action – disappears, more or less, from theory, empirical research makes it its new unit of observation. The revolution in techniques of observation and statistical analysis which took place in the immediate post-war period, and which saw sample surveys based on individuals abstracted from their social and cultural contexts supplant community studies (in the style of Lynds’ Middletown), led research in a direction diametrically opposed to that taken at the theoretical level by both the functionalists and their rivals of the so-called “conflict school.” This type of research has two shortcomings from the point of view of the sociological theory that Coleman advocates: although individualistic, surveying research uses a causal and behaviorist mode of explanation and not a strategic and intentional one; it is therefore incapable of tracing the individual to the “behavior” of the social system. In addition to this accelerated development of statistical techniques, reinforced by the whole demographic tradition which reduces the analysis of social causality to a simple manipulation of mathematical indices, an objective change in the social structure has aggravated this divergence between theory and empiricism. Under the effect of the exponential growth of markets, mass media, large companies and bureaucracies, the principle of structuring the social has shifted from the local community to the national society. This transformation of social relations has itself determined a mutation of the social demand for social science, hence the rise of market studies and policy research (cf. Coleman, 1978) which have accentuated the individualistic and causal bias (in the sense of statistical causality) of research and hindered the elaboration of the rational foundations of sociological theory.

23In order to remedy this growing divorce between an empiricism centred on an abstract individual devoid of intentions and a theory which operates exclusively at the level of a system artificially endowed with its own ends, Coleman proposes to restore the telos to the individual and to import into sociology the anthropological principles which underlie neo-classical microeconomics. [28] Once the ends that functionalists wrongly attribute to the social system have been repatriated to the level of the rational actor, sociological theory can be reduced to the elucidation of two puzzles: first, how do the intentional behaviors of actors combine to produce systemic consequences? Second, how do these economically goal-oriented actions in turn become shaped by the constraints resulting from the functioning of this system? [29] But by focusing the theoretical gaze on the combination of autonomous individual behaviors into collective outcomes, Coleman presupposes that the question of the nature and immanent logic of social behaviors has been resolved. In other words, and this is a major paradox, the unthinking adoption of the philosophy of utilitarian subjectivism prevents Coleman’s theory of action from posing as a problem the very thing that is supposed to constitute its object.

24It is striking to what extent Coleman takes the subjectivist philosophy of rational action, conceived as a reflexive sequence of conscious decisions by an economic actor who is economically and socially unconditioned, as a self-evident datum brutum that does not require the slightest proof. Thus, he opens his collection of articles Individual Interests and Collective Action (1986c, p. 1) with this statement held to be established sub specie aeternitatis:

25

“The rational action of individuals has a unique attraction for sociological theory. One can say that one has ‘explained’ a social institution or process if and only if one has accounted for it on the basis of the rational action of individuals. The very concept of rational action is a conception of action that is ‘understandable’ […], an action about which no further questions arise. Thus social theory can take the purposive goal-oriented individual as its starting point, although psychology may wish to explore this system in order to discover what makes his action coherent or ‘purposeful’.” [30]

26This statement can be compared with that of John Roemer, for whom the assumption of individual rationality is also in the realm of self-evidence: “Methodological individualism is a deductive method which attempts to derive historical observations from basic assumptions about individual behavior which are sufficiently fundamental to be taken as self-evident” (Roemer 1982b, p. 253). Or the philosophic anthropology underlying Jon Elster’s Making Sense of Marx, which Barry Hindess (1986, p. 442) noted in a recent review as “effectively introducing methodological individualism as a premise that requires no defense.”

27And it is only because he has accepted as fully proven (taken for granted) the utilitarian and individualistic philosophy that constitutes the anthropological unconscious of liberal economics, and assumes the question of the logic of social action to be resolved, that Coleman (1986b, p. 362-363) can assert that:

28

“The main theoretical challenge [faced by sociology] is to show how individual actions combine to produce a social product… There must be social institutions in reality (such as the market or electoral systems) which translate individual tastes and endowments into a set of prices and a distribution of goods or into a collective decision. What is needed in social theory is a conceptual apparatus to describe this translation.”

29

“The central intellectual problem of the social sciences […] is this: we understand and can model individual behavior, but we are rarely able to make an appropriate transition from that level to the behavior of the system formed by those same individuals.”

30Elsewhere he continues: “The most successful example of modeling this transition is the pure and perfect market model of neo-classical economic theory” (Coleman, 1987b, p. 157-58). Or (Coleman, 1986b, p. 364):

31

“I believe that the appropriate paradigm for sociology is one derived from Walrasian equilibrium theory, although it differs in part in that not all social goods are divisible, without external effects, and do not all obey the conservation principle; and in part because of the social structure that a Walrasian system ignores.”

32As a result (Coleman in Lindenberg et al., 1986, p. 124):

33

“It seems to me that sociology could take one aspect of microeconomic theory, namely the behavioral model of rational choice, and add to the market a variety of other structures through which social products are generated.”

34According to Coleman, the transition from the micro to the macro can take place, not through the market alone, as postulated in Gary Becker’s relatively crude economic models, but through three main types of mechanisms: markets, hierarchies or systems of authority relations or contractual relationships, and normative systems. He also mentioned the role of trust and communication networks. [31] All of them, however, are indebted to the same analysis in terms of individual interest. There is not a single norm that the Chicago sociologist does not try to explain as the result of the rational action of agents seeking to effectively regulate the behaviour of third parties whose ‘negative side effects’ they have to suffer. [32] The figure opposite illustrates the structure that any sociological theory must therefore adopt, the heart of the latter being formed, in Coleman’s eyes, by the type 3 relation, going back from the micro to the macro, the type 4 relations (macro-macro) being the epitome of fallacious reasoning in sociology. [33]

Basic structure of sociological theory according to Coleman

Basic structure of sociological theory according to Coleman

35James Coleman’s merit here is to go beyond the ultra-simplistic hypotheses of a Becker by recognizing the existence of micro-macro “translation” mechanisms other than the simple market. By clearly stating the need to link the “theory of action”, even if it is rational, to a theory of social structure, [34] Coleman also avoids some of the aporias that undermine Elster’s approach, for whom structure is reduced to a set of games (the prisoner’s dilemma, the battle of the sexes, the insurance game, the loan for a loan, etc.), that is to say, in the final analysis, a series of individual interactions planned under ex ante specified constraints. Finally, his diagnosis of the yawning chasm between theory and empirical research contains an implicit critique of the methodological fetishism of orthodox American sociology, which seems salutary to us. [35] But these contributions are only possible against a background of unanalyzed premises and a truncated anthropology whose prohibitive costs Sewell will hasten to expose.

Sewell and the “golden age” of historical sociology in America

36That it is a trained historian and not a “pure” sociologist [36] answering Coleman’s question is not surprising if we know that historical and comparative sociology has been expanding in the United States for the last fifteen years. Observers as different as the behaviorist George Caspar Homans (1986) and the Weberian Randall Collins agree in seeing in it one of the most promising recent advances in social science on the other side of the Atlantic, the latter not hesitating, in his assessment of current trends in American sociology, to call it a veritable “golden age.” [37]

37This resurgence of historical sociology has its origins in the discredit, both scientific and political, that the societal transformations of the post-war period and the socio-political conflicts of the 1960s have inflicted on the theories of modernization (the psychology of Inkeles, McLelland, and Lerner, or the structuralism of Marion Levy and Walt W. Rostow) and to the functionalist evolutionism of Parsons and Eisenstadt, which until then had reigned unchallenged in Anglo-American macrosociology. [38] Encouraged by the rehabilitation of the classical tradition of Marx, Weber and Tocqueville, a new generation of young researchers found in the historical and comparative method the means not only to neutralize, by making them visible, the ideological and ethnocentric biases of the discipline, but also to escape from the sterile opposition of the historical and comparative method.

38It also avoided the sterile opposition of “Grand Theory” and “abstract empiricism” denounced by C. Wright Mills (1977), and followed in the wake of the pioneering work of Barrington Moore, Reinhard Bendix, Charles Tilly, Gerhard Lenski and Immanuel Wallerstein.

39Since then, historical sociology has extended to an ever larger and more varied universe of objects: From ethnic relations to the emergence of welfare states and their social policies, [39] through the formation of the working classes, [40] revolutions and popular insurrections, [41] the professions, [42] the history of power or the world capitalist system, [43] and more recently, to culture, [44] there is no longer any part of social reality that is foreign to it, to the point that it is now asserting itself as one of the most intellectually dynamic sectors of American sociology (even if it remains socially less powerful than the RAT).

40[…]

41The popularity of the cultural approach in historical sociology is a function of the recent expansion of cultural sociology (marked by the creation in 1987 of a section of Sociology of Culture within the ASA), among others under the impulse of Foucault, Elias, Bourdieu, Habermas, and the work of the Centre for Contemporary Cultural Studies of the University of Birmingham, [45] and where one finds a connection similar to that between RAT and historical sociology, between institutionalist approaches which tend towards an instrumental view of action on the one hand and the sociology of culture on the other.

42This series of homothetic oppositions between rationality and norms, institution and agent, system and actor, structure and culture, impersonal cause and subjective meaning, is reproduced within historical macrosociology itself, within which, from this point of view, one can roughly distinguish two tendencies: one is structural and is partly linked to the current of American structuralism (associated with the names of Blau, Mayhew, Harrison White, Ron Burt, etc.), while the other is rather “culturalist,” and betrays the impact of Geertz [46] and of the symbolic anthropology born in Chicago in the 1960s, as well as the persistent influence of the hermeneutic and interactionist tradition.

43[…]

44It is this culturalist interpretation of the historical method that informs Sewell’s criticisms and counter-proposals in his response to Coleman. Against the abstractions of the individualist scheme of rational choice, Sewell outlines a theory of action attentive to the concrete spatial and temporal localization of the processes studied, centered on the dialectical interplay of meaningful actions and their structural contexts, and imbued with a concern for historical variation and specificity. Sewell underlines the simplifying and unilateral character of Coleman’s model and questions the distinction between the micro and macro levels of analysis that is at its core. His critique focuses on each of the three main types of relationships identified by Coleman, and especially on the exorbitant importance that Coleman gives to the micro-macro relation.

45Sewell argues first of all that the micro-micro relation (type 1 on the diagram above), which Coleman tends to present as self-evident, can neither be grasped nor even conceptualized without first giving oneself the macro-sociological framework which alone defines and determines it, and that therefore, in all rigor, it only makes sense as a direct and inescapable function of the totality of the type 2 and 3 relations from which Coleman extracts it by means of a theoretical coup de force. In other words, the relations between individual behaviors pass through the complete system of relevant social relations and are not amenable to an individualistic analysis. [47] In the same vein, Sewell notes that Coleman’s RAT is based on a Hobbesian conception of the individual as an autonomous entity outside of history and society, and assumes that “the interests, preferences, or goals of individuals can be deduced without difficulty from their social or economic position.” [48] Drawing on the work of Geertz, Therborn, Foucault, Bourdieu and Hochchild, he suggests, on the contrary, that it is not enough to posit ex cathedra the existence of individual interests, preferences and goals in order to elucidate the effects of the so-called “macro” level on the behavior of agents (a type 2 relation). For the individual and their values cannot be hypostasized and reduced to exogenous variables in this relationship. The social person is indeed constituted in and by the totality of the social and cultural links where the biological individual who is the “support” is inserted. Individuality is not ahistorical but results from precise sociohistorical processes. [49] And it is the history of these fields, such as it is specified through the filter of the individual and collective trajectory of the agents who invest in them, which defines the habitus as a generating system of preferences, that RAT declares, by methodological decree, to be exogenous to social action and structure. By omitting to ask the question of the social conditions of formation of these interests, preferences and goals, Coleman refrains from seeing the individual as a social and historical construction that is also subject to sociological analysis. From then on, he has no choice but to regress to the limited (and in many ways, it seems to us, narrowly Americano-centric) model of a homo economicus endlessly pursuing an abstract and unchanging interest.

46Finally, Sewell reminds us that the micro-macro transition (type 3 relationship), which according to Coleman constitutes what Lakatos would call the “hard core” of sociological theory, cannot be limited to the aggregation, whether by means of exchanges, hierarchies or normative constraints, of individual actions all uniformly oriented towards the maximization of an interest already constituted outside any historical and social context. In reality, agents never create a social system ex nihilo, and all theoretical edifices based on such Robinsonades, to speak as Marx, are doomed to produce conclusions as fictitious and unreal as their hypotheses. And Sewell aptly notes that the celebrated “Hobbesian problem of social order,” formulated by Parsons and by which American sociological theory has since bought itself a (false) philosophical conscience, is only a “pre-sociological or non-sociological” problem. [50]

47It is clear that, for Sewell, it is the very starting point of Coleman’s model that is problematic here: by postulating a presocial state of nature, the latter “reduces the problem of action in its dialectical reality – how structurally constituted actors act so that the combined effect of their actions transforms the very structures that constitute them – to the simpler, narrower, and far less interesting linear question of how, under varying conditions, the actions of individuals pursuing their own self-interest cumulate to produce various social consequences” (Sewell, 1987, p. 169). In place of this neo-Hobbesian problematic, Sewell outlines a sociological research program inspired by history and interpretive anthropology, anchored by the following assumptions:

481) The dispositions, practices and representations of agents are the product of specific cultural and historical conditions, not transhistorical data.

492) Every society contains within it conflicts and tensions leading to structural changes that generally differ from those deliberately pursued by the agents.

503) These systemic consequences, whether intentional or not, depend not only on the constellation of goals and resources of the parties involved, but also on their own temporal articulation.

514) Social agents are capable of reflexivity in that they tend to adjust their habits, desires, and worldviews to objective transformations.

52Sewell (1987, p. 171) closes his critique with a brief plea for a resolutely historical sociology:

53

“Instead of abandoning history for ‘as if’ stories, sociology would do better to strive for a more theoretically and empirically rigorous specification of the complex dialectical processes by which real social systems are transformed by historically dated and situated actors.”

54In his response to Sewell, Coleman (1987a) stands firm. [51] He does not want to see in Sewell’s criticisms anything more than the expression of a difference of “interest in knowledge,” as Habermas would say, between the sociologist, who would aim at establishing general relationships between variables, and the historian, who would seek to grasp phenomena or chains of particular events in a narrative mode. Taking refuge behind this perennial dichotomy between nomothetic and ideographic research Coleman merely reiterates what for him constitutes the two cardinal faults of all sociological theory: failing to descend to the level of individual actors and their (rational) conduct, and concentrating on the macro-micro articulation to the detriment of the opposite movement. It is true that to tackle the objections and questions raised by Sewell head-on would require him to question the implicit ontological and even metaphysical postulates of the rational choice scheme – in other words, to take the trouble to elaborate a real theory of social action instead of confining himself to a theory of the combination of rationally reconstructed individual behaviors.

Beyond the rationalist illusion of “mythological individualism”

55As we can see, the interest of this Coleman-Sewell debate goes far beyond the simple exchange between two authors, since it offers a paradigmatic confrontation between two epistemological poles whose force of attraction is growing today in the field of American social sciences and which can be quickly and simply characterized as follows: on the one hand, the individualist and rationalist pole, which conceives of the social order as the aggregation, simple or composite, of individual actions involving agents deliberately seeking to maximize their utility by the instrumental adjustment of available means to clearly given and ordered ends; and on the other hand, the historical and culturalist pole which endeavors to understand the logic of these same actions by piercing their subjective and contextual meaning, by discovering the logic of the constitution of the agents and their ends, and by tracing their reciprocal effects in time. On the one hand, there are formal models, most often mathematical, based on deliberately simplifying – sometimes grossly and unrealistically – postulates which form the basis of deductions of a predictive nature; on the other hand, there is an interpretative “reading” of reality as a ‘text’ which privileges induction and the comprehension of the meaning which the agent gives to their conduct. These are indeed two opposed conceptions of social action: homo rationalis vs. homo culturalis, or, in the language of Martin Hollis (1977), the Autonomous Man against the Plastic Man; hard science against soft science, Erklären against Verstehen, quantitative methods against qualitative procedures. And these two poles pull in opposite directions: while Coleman invites sociology to a rapprochement with economics, Sewell calls sociology to the side of anthropology and, in some ways, “anti-scientific” history.

56But did Sewell go far enough in his critique of the rationalist utilitarianism defended by Coleman? One is entitled to think not. Coleman appeals without further justification to the notion of rationality, a concept which is nowadays much debated within the camp of methodological rationalism. What kind of rationality does Coleman claim: Perfect or imperfect? Total (maximization) or selective (satisfaction)? Bayesian or non-Bayesian? Parametric or strategic? [52] So many questions to which he provides no answer. Nor does he answer the proven inability of RAT to explain the beliefs, preferences and cultural conventions that govern practices, [53] or the difficulty it has in not confusing preferences with reasons. [54]

57In order to circumscribe more precisely the field of validity of RAT, it is necessary and sufficient to recall the very thing that it must constantly repress in order to maintain its claims to paradigmatic hegemony, that is to say its social and economic conditions of possibility. The usefulness of the theory of rational choice depends on the degree to which its starting hypotheses are actually realized in practice. The most important of these assumptions is that the relevant forces are individuals with pre-existing interests or preferences that are coherent, hierarchical, transitive, unambiguous and not subject to intersubjective variation. From this postulate follows a second one, according to which individual behaviors are rational, i.e. efficiently oriented towards the satisfaction of the agent’s interests, which supposes that the agent has enough information to proceed to an intelligent choice (this premise of rationality can of course be diluted at will by resorting to the notion of “revealed preference,” which has the considerable advantage of rendering the demonstrations of the RAT entirely tautological). These postulates make the scheme of rational action particularly applicable in social contexts where agents are both highly individualized and clearly culturally defined, and where comparisons between alternatives are immediate because of inexpensive, complete information that can be confronted with univocal decision criteria. As noted by the anthropologist Jerome Davis (1973) in his critique of the economist theory of exchange, the situation that best corresponds to RAT is that of the consumer shopping in a supermarket. The question is whether it is acceptable to generalize the store paradigm into a “cafeteria conception of society” (Worsley, 1984, p. 246) or into a gigantic game of Monopoly.

58It is not enough, however, to inject a bit of culture or symbolism in order to escape the illusions of the RAT. We know that an analysis which proceeds, according to the Parsonsian model, by making culture the table of values and norms which fix the action’s objectives and which can be invoked to account, most often circularly, for almost any type of observed behavior, is just as unenlightening. To avoid this normativist trap, Ann Swidler (1986, p. 273) suggests infusing the analysis of culture with a strategic dimension by reconceptualizing it as “an array of symbols, stories, rituals, and worldviews that people can use in various configurations to solve different kinds of problems.” The causal analysis of behaviors then consists in showing how culture provides actors with the “prefabricated elements” from which they elaborate their strategies. Notwithstanding its intentions, such an approach does not get us out of the dead ends of the instrumentalist individualism of RAT, since it tends to reduce culture to a simple object, a means of action, rather than making it a constitutive or structuring element of practices. Although she recommends Bourdieu’s work, Swidler does not seem to perceive that the notion of cultural strategy that Bourdieu implements supposes neither conscious intentionality and calculating reason (Bourdieu, 1988), nor the deployment of individual strategies.

59There remains the thorny debate concerning the cross-cultural validity of the notion of rationality, very conveniently evaded by the Rats of all sides, but which also constitutes a major stumbling block for a historical sociology of culturalist inspiration. Pushed to their limit, the universalism of the RAT and cultural and historical particularism converge to make any social science impracticable, as revealed by the controversy provoked by the neo-Wittgensteinian thesis of Peter Winch in The Idea of a Social Science. [55] On the one hand, there is no doubt that the naive universalism of RAT is based on an ethnocentric projection of the analyst’s cultural presuppositions that denies them access to an adequate understanding of foreign cultures. On the other hand, the particularist argument pushed to its logical limits leads to a relativism that makes the notion of social knowledge highly problematic: if all knowledge is entirely dependent on its immediate context (“indexical” as ethnomethodologists say) and inseparable from the original “language game” in which it is formulated, then the very possibility of any cross-cultural understanding disappears. Some people, surfing on the post-modernist wave that has recently brought the nihilist critique of science back into fashion, have been quick to embrace this hyper-relativism and to deduce from its allegedly inescapable character the need to abandon the very project of a science of society. But as Charles Taylor (1985, chap. 4) has clearly seen, far from proceeding by static correspondence between concepts and linguistic references, inter-cultural understanding takes place through a process of mutual learning that transforms the two original conceptual registers.

60This opens the way for a reflexive social science capable of taking seriously the specificity of cultures and actors – including that of the analyst – without getting lost in it. In a similar way, recent advances in feminist theory have swept away the presupposition of sexual undifferentiation embedded in the ethnocentric universalism of RAT by exploding the idea of an actor without gender. [56] There are so many problems that it makes us wonder if there exists a methodological individualism that does not take as its starting point those “men in a state of fantastic isolation and rigidity” that Marx mocks (1967, p. 47-48); in short, that is, in the recent words of Michael Burawoy, something other than a “mythological individualism.” [57]

61But that is not all, because it would not be enough to historicize and sex up the notion of the individual and to release univocal and universal criteria of rationality to make up for the theoretical deficiencies of RAT. It is the model of science that it proposes that must be questioned in its turn. If we accept the utilitarian and individualistic premises of Coleman’s conception of action, we may ask ourselves whether economics offers the means to bridge the gap between theory and empiricism that he rightly deplores, given his disregard for empirical research. [58] The idea that economic science would progress, step by step, by dint of empirical tests of hypotheses logically deduced from the RAT, which underlies its representation as an ‘advanced’ science likely to serve as a model for sociology, a backward and immature science, is nothing more than a positivist illusion (Hollis and Nell, 1975). Moreover, it is paradoxical, to say the least, to invite sociologists to adopt neo-classical economic models at the very moment when their simplicity and unrealism make them the target of a radical challenge from within their mother discipline. [59] As Collins (1986a, p. 1352) notes:

62

“The rational choice model is an imperialist intrusion into sociology by economics which comes from the most classical and ‘backwards’ area of contemporary economics.”

63Universalizing the particular case of liberal economies with a high degree of objective rationalization, RAT, both in its sociological version and in its Marxist variant, projects the image of the capitalist market economy formed by neo-classical theory onto the whole of society. [60] In so doing, it refrains from discerning the multiple rationalities – and often irrationalities from the point of view of economics stricto sensu – that govern the various social fields. For (Bourdieu, 1987, p. 125; 1988):

64

“Far from being a sort of natural anthropological given, interest, in its historical specification, is an arbitrary institution. There is not one interest, but many interests, varying according to time and place, almost to infinity… The interest thus defined is the product of a determined category of social conditions: a historical construction, it can only be known by historical knowledge, ex post, empirically, and not deduced a priori from a transhistorical nature.”

65Thus, the notion of intentional strategy oriented only by the research of the interest of the “economic economy,” with all that it carries of deliberate decision, conscious calculation, rational reasoning, planned and reflected choice, and of “reasoning reason” (as opposed to practical reason), leads to projecting into reality the model which is supposed to give a reason for it and to attribute to the agents themselves the posture and the “rational” look of the analyst, transforming practice into an “imaginary activity of imaginary subjects” as Marx says in The German Ideology. RAT would not be so harmful to the development of a sociology of action if, under the guise of heuristics or theory, it did not carry in the background an atomistic ontology of the social which is at the same time a real scholarly denial of the social, of action and collective claims, of social movements and of the threatening political possibilities of which they are bearers. [61] By playing simultaneously on the registers of the model and of a derealized reality, by sliding insensitively from the analytical to the anthropological, the RAT lends to the agents the intellectualist relationship to practice which, as Pierre Bourdieu (1980, book I) underlines, owes its contemplative point of view on practice to the fact that it has to withdraw from it in order to take it as an object, and comes to confuse those decisive abstractions that are the rational reconstructions with reality itself.

66With so many presuppositions and blind spots the theory of rational action is not a solid or rigorous basis for a general theory of society, nor of the economy. This was clearly seen by Weber – whom certain Rats, notably in France, do not hesitate to call regularly and almost ritually to the aid of their cause [62] – when he insisted on the unbridgeable chasm that separates the practical logic of action from that of the model that the scientist builds for the purposes of analysis:

67

“For the purposes of a typological scientific analysis it is convenient to treat all irrational, affectually determined elements of behavior as factors of deviation from a conceptually pure type of rational action. […] Only in these respects and for these reasons of methodological convenience is the method of sociology ‘rationalistic’. It is naturally not legitimate to interpret this procedure as involving a rationalistic bias in sociology, but only as a methodological device. It certainly does not involve a belief in the actual predominance of rational elements in human life, for on the question of how far this predominance does or does not exist, nothing whatever has been said.” [63]

68And Weber (ibid., p. 7) warns us, with an astonishing sense of anticipation: “That there is, however, a danger of rationalistic interpretations where they are out of place cannot be denied.” It is to this theoretical sleight of hand, which consists in giving a theoretical model (which is, by the way, very simplistic) for a practical scheme, in formulating a generalized practice of the economy where there should be a generalized economy of practices, as much as to its obvious affinity with the dominant vision of the social order as a simple aggregation of freely consented individual choices and to its ability to exorcise the specter of collective action, that the theory of rational action owes its social success within the American academy.