Durkheim's Epistemology: The Neglected Argument

by Anne Warfield Rawls
Durkheim's Epistemology: The Neglected Argument
Anne Warfield Rawls
The American Journal of Sociology
Start Page: 
End Page: 
Select license: 
Select License

Durkheim's Epistemology: The Neglected Argument1

Anne Warfield Rawls

Wayne State University

Durkheim's epistemology, the argument for the social origins of the categories of the understanding, is his most important and most ne- glected argument. This argument has been confused with his sociol- ogy of knowledge, and Durkheim's overall position has been misun- derstood as a consequence. The current popularity of a "cultural" or "ideological" interpretation of Durkheim is as much a misunder- standing of his position as the "functional" interpretation from which the current interpretations seek to rescue him. Durkheim articulated a sophisticated epistemology in the classical sense, a point that has been entirely missed.

Durkheim's epistemology, the argument for the social origins of the cate- gories of the understanding, is his most important and most neglected argument. The argument, which is articulated mainly in the central chap- ters of Les Formes e'le'mentaires de la vie religieuse: Le Systbme totdmique en Australie (1912), locates the origin of the fundamental categories of thought in the concrete empirical details of enacted practices. Unfortu- nately, this epistemological argument has been confused with Durkheim's sociology of knowledge, leading to its general misinterpretation as an ide- alist argument that beliefs and collective representations are the origins of the basic categories of thought. As a consequence, the epistemology proper has been generally neglected. The current popularity of a cultural

' This article has benefited from numerous comments and suggestions. In particular, I would like to thank Harold Garfinkel and Randy Collins for extensive comments and discussions. Norb Wiley, Ira Cohen, Anthony Giddens, Stephen Fuchs, Gary Shepherd, and several AJS reviewers also made helpful comments to which I have tried to respond. I thank Charles Lemert for renewing my interest in Durkheim and Peter K. Manning and David Maines for providing encouragement at critical points. Careful readings by Albert J. Meehan have also been extremely helpful. Lynetta Mosby deserves particular thanks for tireless work in the library locating obscure texts. I would also like to thank students in my Durkheim seminar-Recco Richardson, Wendy Evans, Tyronne Gibbs, Jerry Charbonneau, Abdi Kusow, Edward Mays, Gary David, and Carey Ford-for their interest in these issues. Finally, a special note of thanks to Bernard Elevitch with whom I read the British Empiricists and John Lavely with whom I read Kant. My work owes a great deal to their emphasis on the careful reading of texts. Direct correspondence to Anne Warfield Rawls, Department of Sociology, Wayne State University, Detroit, Michigan 48202.

O 1996 by The University of Chicago. All rights reserved. 0002-9602/97/10202-0004$01.50

430 AJS Volume 102 Number 2 (September 1996): 430-482

or social constructivist interpretation of Durkheim is as much a misunder-

standing of his position as the functional or positivist interpretation popu-

larized by Talcott Parsons from which the current interpretations seek

to rescue him. The conflicting interpretations of Durkheim's position as

functional and ideological both result from a failure to understand the

relationship between his various arguments, his sociology of knowledge,

and his epistemology.

Before Kant, epistemology struggled with a separation between thought and reality occasioned by essential differences between the two: thought, consisting of concepts that are general and continuous, and reality, con- sisting of flux. Since Kant, epistemology has had to deal with a separation between thought and reality created by human understanding: natural reality is always perceived in terms of human categories of thought and never in itself. In either case, human ways of perceiving and thinking add something to reality that was not there in the original. As a consequence, human knowledge seems not to stand in an empirically valid relationship with reality.

During the course of the 20th century, due to a growing consensus that an argument for empirical validity cannot be made, the classical form of the epistemological question, which required empirical validity, has been increasingly abandoned in favor of a social constructivist approach to knowledge, which embraces an element of indeterminacy. On this view, social consensus and socially accepted definitions of meaning are treated as the true measures defining the limits of validity. This can be seen in the growing influence of postmodern and pragmatist approaches within the social sciences and humanities.

According to Durkheim, this abandonment of classical criteria of valid- ity in favor of a consensus theory of truth, which had already been popu- larized by William James at the turn of the century, only appears to be necessary because the epistemological question has been cast, by both em- piricists and apriorists, in individual and naturalistic terms, How can indi- vidual perceptions of natural reality be valid knowledge of that reality? (See Cuvillier [I9551 for an extended discussion of the development and influence of pragmatism in France during this period.) Durkheim argued that the emphasis on the individual and individual perception of natural forces made the epistemological problem appear unsolvable. Durkheim sought to replace the individualist approach of traditional philosophy with an approach solidly embedded in enacted social practice.

Durkheim articulated his epistemological argument primarily in three works: The Elementary Forms ([I9121 1915),2 Primitive Classification

A new translation of The Elementary Forms (Durkheim [I9121 1995)by Karen Fields appeared while this article was already under review at AJS. Citations, therefore, are

(Durkheim and Mauss [I9011 1963), and Pragmatism and Sociology

([1913-141 1955). The essay on primitive classification worked out the

parameters for the origins of the category of classification but failed to

distinguish the social logic of the concept (the sociology of knowledge)

from its genesis in enacted practice (the epistemology). The lectures on

pragmatism worked out the classical epistemological problem in some de-

tail and critically evaluated the pragmatist, or social constructivist, solu-

tion to the problem. But Durkheim's own epistemology is not elaborated

in that work. It is only in his study of elementary forms of religion that

Durkheim's epistemology is presented in its entirety. There he criticizes the

epistemological schools of empiricism, apriorism, and pragmatism, which

were popular at the turn of the century, and presents his own socioempirical

epistemology as the only viable alternative in the central chapters.

The problem with an individualist approach to epistemology, according to Durkheim, is that, when knowledge is thought to begin with individual experience, certain problems arise: the things that persons experience change from day to day and from moment to moment. Nothing is ever exactly the same twice and the stream of experience (if persons have not already acquired general categories of thought) is constantly changing and undifferentiated (in a state of flux). Even if particular individual experi- ences are considered to be empirically valid, all attempts to generalize on the basis of them are invalid if what is added in the process was not an original property of the particular experiences. What does a general cate- gory represent that is different from the experience of particular objects and events? If general categories represent something not present in the separate particular experiences, something that is added by the mind to the collection of particulars, then general ideas have no empirical validity.

Attempts by empiricists to solve the problem through a detailed logical analysis of individual perception and the logical relation between objects in perception concluded that not only logical relations but all relations are properties added by the mind and not part of the perceived object in its own right. Thus, the empiricist attempt to establish a direct relation between perception and an underlying natural reality ended in skepticism. Durkheim (1 9 15, p. 2 7)criticized classical empiricism for failing to explain the possibility of deriving even the simplest general ideas directly from experience: "Classical empiricism results in irrationalism; perhaps it would even be fitting to designate it by this latter name." Apriorism, ac- cording to Durkheim, was no alternative as it resulted from accepting the empiricist dilemma and then treating the impossibility of generating the categories empirically as the basis for establishing their apriori status.

to the original translation by Joseph Ward Swain. I have also translated directly from the original French where noted.

Durkheim argued in his lectures on pragmatism that even James's

"radical empiricism" retains the problems of classical empiricism. By re-

placing the dualism of thought and reality with individual action as a

dynamic connection between the two in a context of utility, James had to

give up the possibility of truth and logic. The utility of pragmatic action

remained an individual utility, and James could not explain (without fall-

ing back on either realism or idealism) how the beliefs and practices asso-

ciated with individual utility can become general in an empirically valid

way. While rejecting both empiricism and apriorism, Durkheim also dis-

agreed with James, arguing that unless some common and valid concep-

tual basis shared by all persons can be established, the problem of ex-

plaining both individual knowledge and intersubjective communication

will remain unsolvable and truth and knowledge will remain indetermi-


Durkheim's socioempiricism focuses on a dynamic relation between group members as participants in ritual social processes and the social processes that their participation enacts. While natural forces could only be perceived as particulars, social forces, Durkheim argued, are inherently dynamic and continuous and, when perceived as such by persons assem- bled to enact practices, provide an empirical source for the categories. This way of addressing the gap between thought and reality replaces the individualist approach, which characterized both empiricism and aprior- ism, with an epistemology that is socially based. The resulting epistemol- ogy treats concrete social processes as natural processes whose function is to make general categories of thought available to their human partici- pants. In this way Durkheim hoped to overcome the epistemological di- lemmas resulting from individualism and to address the problem of mu- tual intelligibility. In rejecting the individual as a starting point, the way is opened for Durkheim to explain the origin of the necessary basic con- cepts in terms of concrete social processes, something that had never been tried before.

Unfortunately, this substitution of the social for the individual has been generally misinterpreted as a substitution of the ideal for the empirical. In attempting to develop a path between empiricism and apriorism, the argument goes, Durkheim was forced to place his knower outside the em- pirical world because he accepted the argument that valid categories of the understanding cannot be generated from within empiricism. Thus, it is argued, Durkheim is forced to create an ideal world of symbols, ideas (representations), and beliefs as an origin of the essential forms of human thought.

This is, however, a complete misinterpretation of Durkheim's epistemo- logical argument. He did not attempt to walk a middle path between em- piricism and apriorism by locating the origin of the categories in an ideal,

timeless, and ahistorical social position or by locating a knower outside the empirical. He quite explicitly emphasizes the importance of the histori- cal and the empirical. For Durkheim the middle path is a concrete and empirical social path composed of practices, not a system of symbols and beliefs. This concrete social emphasis placed individual persons among their fellows engaged in a mutual enactment of social practices, particu- larly religious ritual, instead of in an individualist natural context, before posing the epistemological question. This allowed Durkheim to distin- guish between individual sensations of natural forces, which posed prob- lems for empiricism, and the experience of moral force by persons engaged in enacting shared practices. He argued that the perceptions of moral force generated by shared enacted practice were internal, emotional, and imme- diate and therefore did not require the abstraction from a series of particu- lar perceptions, which had been a major stumbling block for the empiricist explanation of general ideas.3 For Durkheim, social practices are not ideal, and they do not consist primarily of ideas, representations, and beliefs. These are merely secondary phenomena. For Durkheim, society consists first and foremost of enacted practices that give rise to real social forces that participants in the assembled group experience jointly. Certain of these social forces give rise to essential ideas experienced in common, which Durkheim refers to as the categories of the ~nderstanding.~

The purpose of religion in human history is to provide the enacted practices necessary to generate these essential ideas.

The Elementary Forms presents a careful and thorough historical and comparative argument for the empirical origin of six basic ideas, or cate- gories of the understanding, identified by the philosophical debate as es- sential to epistemological validity (time, space, classification, force, causal-

'The emotional character of the perception of moral force is an important issue. By claiming a distinction between emotional perception and perceptions that come through the five senses, Durkheim reinforces his argument that perception of social forces is not subject to the same problems as perception of natural forces. Not only are social forces continuous in their own right, but they are perceived through a differ- ent faculty of mind, an emotional faculty. This point Durkheim shares in essential respects with Hume. Elsewhere (Rawls 19966) I have provided further discussion of this point.

The article "the" in the phrase "the categories of the understanding," is not a problem of grammar caused by ignorance of the difference between French and English. Locke used the phrase "the categories of the understanding" in 1690. As he wrote in English, there is no translation problem involved. The phrase as I cite it is Kant's ([I7811 1965) and is always translated into English with the article attached as it appeared in Locke. This is because the argument being made is not a relativistic one. It does not refer to just any categories of any understanding but rather to the categories of the understanding as a constant. Referring to the categories as "categories of understanding" suggests their indeterminacy. Reading the text this way contributes to its misinterpre- tation.

ity, and totality).$ Unfortunately, Durkheim's commitment to empirical detail results in long sections on totemism that represent essential parts of his epistemological argument but are so long and apparently focused on totemism per se that their epistemological significance has been missed. For example, Durkheim takes up animism in order to argue against an apriori explanation of totems. Similarly, he takes up the arguments of Frazer and Tylor in order to argue against a classical empiricist explana- tion. Durkheim's consideration of individual and sexual totems consti- tutes an argument that totems do not have an individual origin. Because Durkheim built his epistemology on details of the actual enactment of totemic rites, arguments that treat those details as having an individual origin contradict his position, and he must deal with them. Unfortunately, these long sections on totems have been treated as a consideration of to- tems in their own right, or as explorations of "conceptual systems," rather than elaborations of the empirical details of enacted practice as an essen- tial part of Durkheim's epistemology.

As a consequence, most scholars with an interest in the epistemology have concentrated their attention on the first and last chapters, where the epistemological argument is only sketched, while ignoring the epistemo- logical argument in the central chapters, and a curious misreading of the text has resulted. The first chapter is an earlier elucidation of the argu- ment, which appeared in 1909 as "Sociologie religieuse et thCorie de la

That Durkheim makes empirical arguments for the origin of only six categories (time, space, classification, force, cause, and totality) in The Elementary Forms has been missed. Commentators on the epistemology apparently do not grasp the exclusivity of the list, and it is represented in different ways by different commentators, often with an "etc." This may be due in part to references that Durkheim himself makes in several places to categories in the work of Aristotle and other philosophers. The first reference in TheElementary Forms (Durkheim 1915, pp. 21-22), e.g., is to Aristot- le's list of categories and includes "personality" and "number," which do not appear as categories in Durkheim's argument. "Number" appears again on p. 30 and "person- ality" again on pp. 31-32. These instances all appear in the introduction where Durk- heim is making reference to Aristotle's list of categories and to the general philosophi- cal problem with regard to categories. These are not the six categories that he argues for the empirical validity of in the body of the text. Durkheim also treats categories loosely in the conclusion. One result of this misunderstanding is that one of the most important of the six categories, classification, is generally treated as a survey of classi- fication practices and not as a category in its own right. One reviewer of this article commented, "The question that nags me most is why Durkheim's analysis of classifi- cation has promoted so much research while his analysis of the categories (space, totality, time, force, causality, etc.) has promoted so little." Ithink the question answers itself. Just as this reviewer does not recognize classification as one of the categories, the general sociological public has also not recognized Durkheim's studies of classifi- cation as having anything to do with his epistemology. Therefore, while the epistemol- ogy has been almost totally ignored, the part of it that focused on classification, be- cause it has been misinterpreted as a survey of symbolic systems, has received a great deal of attention.

connaissance," which leaves epistemological and sociology of knowledge

issues relatively, although not entirely, undistinguished (Lukes 1973, p.

408).6 The conclusion presents a similar difficulty. It takes up the sociology

of knowledge after the epistemological argument has been completed and

assumes an understanding of that argument. The epistemological argu-

ment itself appears only in the central chapters, which have been generally

ignored in this regard, leading scholars to infer the epistemology from the

introduction and the conclusion where it is only vaguely sketched. Thus,

the careful relationship drawn between specific practices and specific cate-

gories in the central chapters is missed, and Durkheim is interpreted in-

stead as having focused on symbolic belief systems.

Because the epistemological significance of the central chapters has been missed, the epistemological argument has never been recognized as such. Even the best treatments of Durkheim's theory of knowledge, David Bloor (1982) being notable in this regard, treat it as an argument about the sociology of knowledge and not an argument, in the classical sense, about the nature of mind and the origin of human reason. Yet, it is quite clear from the text that Durkheim articulated an epistemology in the clas- sical sense of the word. The sociology of knowledge is a distinct secondary argument that is meant to rest on this foundation. In his introduction to the English edition of Durkheim's lectures on pragmatism, John Allcock (1983, p, xl) wrote that "a rounded assessment of Durkheim's epistemology has yet to be undertaken." Allcock argues that discussions of the nature of Durkheim's social facts and his sociology of knowledge have been the primary focus of Durkheim scholarship and have been confused with his epi~temology.~

The resulting failure to understand Durkheim's distinction between moral (or social) force and empirical perception as a source for the categories of the understanding has led scholars to dismiss the episte- mology as untenable and focus instead on his sociology of knowledge. Even arguments that purport to address the epistemology proper, often citing Allcock's remarks, continue to treat the argument as an idealist account, entirely overlooking the empirical foundation of the argument in concrete practices (see, e.g., Godlove 1989).

As a consequence, Durkheim's epistemology has often been interpreted as consisting of the dictum that the categories of the understanding, for which he claims empirical validity, come from social representations or belief systems, clearly an impossible argument, which has led to the belief that Durkheim was ignorant of basic philosophical issues (Gehlke 1915;

'According to Lukes (1973, p. 582), an earlier section on sociology and philosophy
was also omitted from The Elementary Forms.
'Unfortunately, while Allcock's (1983) essay is one of the best in many regards, he
offers the lectures on pragmatism as the best source for the epistemology.

Dennes 1924; Parsons [I9371 1968; Lukes 1973; Mestrovic 1993; Stone and

Farberman 1967; LaCapra 1972). Durkheim has been accused of failing

to recognize the limitations of an empirical theory of general ideas, making

him appear to be a positivist, and of arguing that the categories of the

understanding have their source in other ideas, making him appear to be

an idealist. The idealist or social constructivist argument that ideas come

from other ideas is part of Durkheim's sociology of knowledge (for which

he does not claim empirical validity) not his epistemology. Because of this

misunderstanding, it has been assumed that Durkheim claimed an ideal

source for the categories in collective representations when in fact he

claimed for them a direct origin in the concrete experience of enacted


Durkheim argued that the categories of the understanding enter the minds of individual persons during enacted practice in such a way as to be empirically valid. He describes the process by which practices generate these ideas in detail. It is the socioempirical origin of the six categories in enacted practice that, according to Durkheim, allows his epistemology to overcome the duality of thought and reality. The categories correspond to the reality of social forces as generated during enacted practice. Durk- heim's argument is epistemological in the classical sense in that it explains the relationship between perceptions, ideas, and external reality in such a way that key thoughts and concepts can be shown to bear a valid or true relationship to an external reality, which in this case consists entirely of social forces. Durkheim intended his argument to replace the epistemo- logical arguments of Kant, Hume, James, and their followers, perma- nently transforming philosophical debate over the validity of knowledge.

Providing an empirical demonstration of the availability to participants in shared enacted ritual (religious) practices of the categories of the under- standing was a primary objective of The Elementary Foms. According to Lukes (1973, pp. 407-8, 459), the title Durkheim originally proposed for The Elementary Forms in a 1908 letter to Leon, "Les Formes 616- mentaires de la pensCe et de la pratique religieuse," suggests that the inter- est in religion served a long-standing and more important interest in epis- temology and moral^.^ It is only because he believed that religious

Durkheim's doctoral thesis (later published as The Division of labor)originally be- gan with an introduction outlining its implications for moral philosophy (included as an appendix to the George Simpson translation [Durkheim (1893) 19331). His Latin thesis ([I8921 1960) focused entirely on the relationship between the ethics of Montes- quieu and Rousseau and what Durkheim interpreted as their sociological approach to morality. Indeed Durkheim himself in a letter to Revue neo-scholastique, November 8, 1907 (see Durkheim [I8951 1982, p. 259), claims to have adopted a focus on religion in 1895 only after realizing that it provided the basis for his earlier attempt to ground an empirically valid ethics in the original version of The Division of Labor in Society.

practices are necessary in order to generate valid categories of thought that

Durkheim focused on religion. Durkheim adopted an order of argument in

The Elementary Forms that follows the order of Hume's Treatise of Hu-

man Nature ([I7391 1978). It is tailored to handle an epistemological argu-

ment, not a survey of religious beliefs. Durkheim was not interested pri-

marily in religious systems of belief. His interest was rather in the moral

forces, and the categories of thought corresponding to them, that were

generated by religious practices. Durkheim considered religious belief to

be an entirely secondary phenomena with no empirical validity, and he

dealt with religious belief in terms of the sociology of knowledge, not the


The argument that the work of the later Durkheim should be inter- preted as taking an idealist turn, popular from the beginning, achieved dogmatic status in the United States following Parsons's assertion in The Structure of Social Action (1968, pp. 460-65) that the later Durkheim had moved dangerously close to idealism and in Europe with the confusion by LCvy-Bruhl and others of Durkheim's sociology of knowledge with his epistemology. While the argument of The Elementary Forms has consis- tently been interpreted as idealist, Durkheim's discussion of social facts in The Rules of the Sociological Method ([I8951 1982) and in the earlier Division of Labor in Society ([I8931 1933), interpreted without benefit of his epistemology, is given a positivist interpretation, leading to the argu- ment that there are two Durkheims: a functionalist positivist and an ideal- ist (and more recently an idealist postmodernist or pragmatist) (Livy- Bruhl ([l910] 1966); LCvi-Strauss ([I9581 1963); Parsons 1968; Stone and Farberman 1967; LaCapra 1972; Lukes 1973; Alexander 1988; Lehmann 1990; Mestrovic 1993).

The fundamental error lies in failing to distinguish Durkheim's episte- mology from his sociology of knowledge. Durkheim's epistemology, best represented by the central chapters of The Elementary Forms, deals ex- plicitly with the question of whether the six categories of the understand- ing can be demonstrated to have empirical validity. He argues that the six categories do have empirical validity because they are perceived di- rectly as social or moral forces during the enactment of social (religious) practice.

Durkheim's sociology of knowledge, on the other hand, appearing mainly in the introduction and conclusion to The Elementary Forms and at the end of the lectures on pragmatism, consists roughly of the argument that language and culture consist of cosmologies of ideas, which arise as collective explanations for social forces, and that these ideas become forces in their own right whose logical relations can be studied. The cosmologies are responsive to structural changes but also have a certain inertia so that ideas first generated by the interpretation of social processes lose whatever

relation to social practices they might once have had. Therefore, they do

not bear any necessary relation to an underlying reality. Durkheim's soci-

ology of knowledge is concerned with the logic of ideas within such cos-


There is no contradiction between these two positions. They deal with different questions. The epistemology asks how, in the first place, persons develop a framework of six empirically valid basic ideas in common such that any ideas could be shared to a degree sufficient to allow for the devel- opment of a cosmology (Durkheim 1915, p. 30). The sociology of knowl- edge studies the relations between concepts within cosmologies once the categories have developed. Durkheim's epistemology posits a direct rela- tion between a limited set of categories and an underlying social reality. It creates only a small area of knowledge free of indeterminacy. With regard to the natural world, the world of science, and the bulk of social concepts, however, Durkheim's epistemology leaves intact all the prob- lems of indeterminacy that Wittgenstein ([I9531 1958), Quine ([I9661 1977), Kuhn (1962), Feyerabend (1975), and more recently Bloor (1976, 1982, 1983), Fuller (1989), and Latour and Woolgar (1979) have described. Durkheim argues that "only the necessities of action, and especially of collective action, can and must express themselves in categorical formu- lae" (Durkheim 1915, p. 412). Everything beyond the six categories falls within the province of the sociology of knowledge.

In laying out his epistemology, Durkheim focused on religious practices because he argued that religious practices in "simple" societies provided a sufficient degree of collectivity to support the formation of general cate- gories of thought. According to Durkheim, societies could only develop where shared enacted practices produced categories of the understanding in common, thus enabling members of the same group to communicate (Durkheim 1915, p. 30). His detailed argument concerning the origins of religion as well as the detailed descriptions of Australian totemism, which have been misinterpreted as cultural sociology, or sociology of religion, actually constitute a step-by-step analysis of the social generation of the categories of the understanding.

Durkheim articulated a consistent epistemology throughout his career that is much more sophisticated and serious than he is given credit for. This epistemology is evident in the early as well as in the later work: the coherence of his notion of social facts depending entirely on his epistemol- ogy and appearing to be positivist without it. The result is an argument that preserves a direct (although limited) relation between socially gener- ated categories of thought and social reality in general, explaining the development of the categories without invoking the problems of either apriorism or empiricism.

While Durkheim makes rather extensive arguments for the socioempiri- cal origins of six categories of the understanding, it is not possible to con- sider the arguments for all six categories in this article. The arguments for the socioempirical origins of force, causality, and classification are the most important. The argument for force, however, is essentially part of the argument for causality. Therefore, in the interests of brevity, this arti- cle will confine itself to a sketch of the arguments for the socioempirical origins of the two concepts of causality and classification, ultimately the two most important concepts from an epistemological standpoint, consid- ering force only as it relates to causality. Following this discussion, criti- cisms and misunderstandings of Durkheim's position will be taken up again in greater detail.


Durkheim's argument for the social origin of the category of causality, which occurs in book 3, section 3, is the centerpiece of The Elementary Forms. Causality is the key concept that any epistemology must explain, and the earlier sections of the book, particularly the extensive sections on totems, which establish the social basis for the other general categories of space, time, classification, and force, lay the groundwork for the socioem- pirical basis of the category of causality.

In spite of the importance of the argument for causality to the book as a whole, with other arguments leading up to it and building on it, most discussions of The Elementary Forms overlook it, focusing instead on the categories of time, space, and classification. There are three reasons for this. First, the general misunderstanding of Durkheim's treatment of clas- sification, time, and space as social systems of ideas instead of categories leads to an interest in the coherence of the "systems" Durkheim is taken to have described. Those categories that do not lend themselves to a sociol- ogy of knowledge comprised of "systems" of concepts are ignored. Second, this interest in cultural systems of ideas misinterprets Durkheim's argu- ment as an indeterminant social constructivist account of knowledge, thus obscuring his epistemological argument for the empirical validity of the category of causality. Third, there is a general lack of familiarity among sociologists with the importance of the concept of causality to epistemol- ogy and of epistemological issues in general.

According to Durkheim, force and causality are both relations between objects or persons that are known by their effects. Therefore, the problem of establishing the concept of force is similar to that of establishing the concept of causality. The sections preceding the argument for causality establish the socioempirical basis for the concept of force. However, cau- sality contains an additional element: necessity. Causality is a necessary relation between two objects or events. Durkheim describes the percep-

tion of causality as the perception of a special kind of force, which he

refers to as necessary force or efficient force (Durkheim 1915, pp. 406,

410-11, 488).

Durkheim argues that, in order for causality to have a socioempirical basis, it must be the case that there are social rituals in which necessary force is an integral enacted feature (Durkheim 1915, pp. 408-13). As with the other categories, necessary force must be immediately available to per- ception as an enacted feature of the ritual in any single case. It must not be a generalization drawn from a series of perceptions of social rituals, in which case it would still be subject to the objection that general concepts cannot be validly derived from a collection of particular sensations.

Durkheim focuses his argument for the socioempirical origin of the con- cept of force on the totemic rites of sacrifice and oblation because, he ar- gues, these are the rites where moral force is reproduced (1915, pp. 318- 92). These arguments relating different sorts of totemic rites to force and causality are quite specific. He argues that the moral force of the commu- nity and of the totemic symbol are both reproduced via the rituals of sacri- fice (1915, pp. 381-85). The moral force of deities is also reproduced in and through rituals of oblation in so far as deities are representations of social and not supernatural forces (1915, pp. 385-92). Imitative rites, on the other hand, enact the principle of necessary force, or causality according to Durkheim, and, therefore, these rites are the focus of his argu- ments for causality (1915, pp. 393-414). Whereas rituals of sacrifice and oblation achieve a deepening respect and moral sense that is perceived as force, imitative rites invoke and produce a direct relation between action and effects, actually creating and recreating the group that enacts them, thereby achieving the concept of necessary force or causality as an inher- ent enacted quality.

Lukes (1973, p. 448) argues that Durkheim provides no evidence in The Elementary Forms for the causal relation claimed between particular so- cial practices and particular "conceptual systems." The issue is obscured by Lukes's introduction of the notion of conceptual systems into the equa- tion, a sociology of knowledge interpretation of Durkheim's argument. Durkheim never attempted to argue that particular social forms created specific conceptual systems in his sociology of knowledge. However, Durkheim's epistemological argument consists of the much stronger claim that there is a direct relation between specific sorts of social practices and particular categories of the understanding. Imitative rites are the origin of the concept of causality. Properly understood it becomes clear that the entire discussion of totems consists of evidence of just the sort Lukes claims Durkheim has not offered.

Durkheim argues that participants directly perceive necessary force as part of their experience of certain rituals because, as a social relation, necessary force is an enacted feature of those rituals. Durkheim's (1915, pp. 393-414) detailed discussion of imitative rites serves the purpose of establishing the occurrence of rituals that succeed only by enacting the principle of causality.

Causality is an essential concept in philosophy and played a pivotal role in the debate between empiricism and apriorism. Without causality, philosophers on both sides of the debate agreed, science would have no foundation. Indeed, knowledge itself, which Hume argued consists of in- ferences from effects to causes, would be impossible without the concept of causality. If Durkheim secured a basis for the concept of causality in di- rect experience of the social, he achieved something of great importance that has serious implications for current methodological and theoretical debates.


Because Durkheim argues that he has found a social origin for the cate- gories of the understanding, he is often taken to be a Kantian or neo- Kantian. However, both empiricists and apriorists discussed the problem of epistemology in terms of categories and/or general ideas. The difference between the two was in their explanation of the origins of the categories, not whether categories played an important part in their epistemological argument.

Durkheim makes it quite clear that he does not consider Kant's aprior- ism to contain criteria of validity at all and several times refers to the argument as "no explanation" at all (1915, p. 27). He writes: "To reply that it [the idea of classification] was given to us a priori is not to reply at all" (1915, p. 172). He goes on to say that this "lazy man's solution is . . .the death of analysis" (1915, p. 172). Either one explains the existence of the categories, or concludes, as Hume did, that they cannot be explained and, therefore, have no validity. Durkheim clearly seeks an empirical ex- planation of the origins of the categories, although by empirical origins he means origins in shared enacted practices. Durkheim's socioempirical approach satisfies Hume's empirical criterion of validity, not Kant's.'

Durkheim is generally interpreted as a Kantian or neo-Kantian. When his criticisms refer to empiricism, he is often taken to be referring to Locke. However, Locke was quite happy with empiricism as he had presented it and did not make the skeptical arguments that Durkheim cites. These arguments are more characteristic of Hume and James. The argument that Durkheim might not have been familiar with empiri- cism because he studied with neo-Kantians overlooks the fact that Durkheim took a degree in philosophy from the Ecole Normale Superieure. A survey of German univer- sities from 1862 to 1890 (Kohnke 1991, p. 249) shows that the rise in neo-Kantianism brought with it an increased interest in and availability of courses in empiricism and rationalism as well. During this period, according to Kohnke (1991, p. 249) "the share of courses devoted to Plato, Aristotle and Hegel continuously declined, while those

While he set himself to accomplish Hume's task, however, Durkheim

did not intend to employ a modified and watered down empiricist ap-

proach. On the one hand, he intended his argument to fulfill all of Hume's

criteria, while, on the other hand, he felt that shifting the foundations of

the argument from the individual to enacted practices, resulted in an en

tirely new solution to the problem. Therefore, Durkheim should not be

considered an empiricist but rather a socioempiricist. It follows that the

problems elaborated by Hume must be considered carefully before the

subtlety of Durkheim's argument can be appreciated and its effectiveness


Hume argued that all knowledge could be divided into two sorts: mathe- matical knowledge and moral/empirical knowledge (Hume [17 7 71 1975, p. 25). While mathematical knowledge was for Hume a matter of pure reason (because numbers are not real but are logical constructions in the mind), empirical or moral knowledge depended on the concepts of cause and effect, contiguity and resemblance (Hume 1975, p. 26). "By means of this relation [cause and effect] alone we can go beyond the evidence of our memory and senses. . . .All our reasonings concerning fact are of the same nature. And here is it constantly supposed that there is a connection between the present fact and that which is inferred from it" (Hume 1975, pp. 26-27).

Without causal relations between perceptions, they would remain par- ticular and unconnected according to Hume. It is the concept of causality that introduces order into the flux of experience. Every inference, every connection between a present fact and a prior condition, requires the idea of causality. It is causality that allows persons to pull particular ideas out of the flux and connect them to others. Therefore, it seems obvious that persons make use of such a concept. The question for Hume was not whether the concept exists, as it obviously must, but where it comes from and whether or not it has empirical validity (Hume 1975, p. 27). Hume argued that knowledge of cause and effect cannot be attained from experi- ence of a single instance. That would require that objects contain and display their causes, which Hume (1975, p. 27) asserts is not possible: "No object ever discovers, by the qualities which appear to the senses, either the causes which produced it, or the effects which will arise from it." In addition, according to Hume, "every effect is a distinct event from its cause. It could not, therefore, be discovered in the cause" (1975, p. 30).

devoted to Spinoza, Descartes, Leibniz, Locke and Hume, as well as to Kant, showed a continuous increase.'' Durkheim's neo-Kantian training increases the chances that he would have been familiar with Hume, whose arguments are necessary for under- standing Kant. Durkheim uses Hume's arguments with great sophistication. It is only Durkheim's sociology of knowledge that has neo-Kantian overtones and even that argument should not be interpreted separately from Durkheim's epistemology, within the context of which it changes in essential respects.

Generalization over a series of particular experiences, on the other hand, only adds to those particulars the perceiver's feelings about the series as a series, feelings which are not an empirical experience of cause and effect. "For as this idea [causality] arises from a number of similar instances, and not from any single instance, it must arise from that circum- stance, in which the number of instances differ from every individual in- stance. But this customary connexion or transition of the imagination is the only circumstance in which they differ. In every other particular they are alike" (Hume 1975, p. 78).

Hume argues that all that is added to single instances by their collection in a series is the feeling persons get about the series. There is nothing empirically different about the series that is more than the sum of its parts. Therefore, Hume concluded that causality is merely the result of feelings that lead to custom and habit. There are, for Hume (1975, p. 78), no empir- ically valid experiences corresponding to causality. "Every idea is copied from some preceding impression or sentiment; and where we cannot find any impression, we may be certain that there is no idea. In all single in- stances of the operation of bodies or minds, there is nothing that produces any impression, nor consequently can suggest any idea, of power or neces- sary conne~ion."'~

Repetition can never produce knowledge of causality. It can only produce the feeling or belief that similar things will keep hap- pening. This belief becomes customary and habitual, and, according to Hume, the feeling or belief itself is the origin of the concept of causality: "But, when many uniform instances appear, and the same object is al- ways followed by the same event; we then begin to entertain the notion of cause and connexion. We then feel a new sentiment or impression . . . and this sentiment is the origin of that idea which we seek for" (1975, p. 78). Thus, causality originates in a feeling, not in an empirical experience. It, therefore, has no empirical validity and persons are mistaken when, following custom and habit, they believe the transfer of this feeling to external relations has any validity: "And as we feel a customary connexion between the ideas, we transfer that feeling to the objects; as nothing is more usual than to apply to external bodies every internal sensation, which they occasion" (1975, p. 78n). Hume demonstrated that causality cannot be explained on the basis of individual experience of natural ob-

'O In this quote, Hume uses both the 18th-century meaning of "idea" or impression to denote an exact copy of a sense impression or sentiment and the more modern sense of an idea as a thought. If an idea is a copy of a sense impression, it has empirical validity. If it is a copy of a sentiment, as in the following quote, it is not empirically valid. While the passage is somewhat ambiguous due to Hume's failure to use the word "idea" consistently throughout, he does not mean that people do not have and use the concept of causality. He means that there is no valid idea or sense impression corresponding to the concept being used.

jects because causality is not a sensible quality of natural objects and their

relations that can be experienced by individuals in a single case. If the

disposition to believe in causality is an instinct, or a conclusion drawn

after perceiving a series of natural objects or events, or a matter of custom,

then it is something added by the mind to perception, not part of the

original perception, and therefore not empirically valid.

The significance of Hume's project is that it fails, or rather, that Hume shows that an individually based empiricism could not explain the empiri- cal origin of the essential general categories. A modern social constructiv- ist account of this problem accepts Hume's conclusion that knowledge of causality has no empirical validity. Social constructivists then focus on an account of the social determination of custom and habit that provides for more consistency between individual ideas than Hume's conclusion would seem to allow for. Although it could be argued that Locke's ([I6901 1959) "common acceptation" is already an early version of a social con- structivist argument (Rawls 197 7).


While Durkheim accepts Hume's argument that individual perceptions of natural events cannot give rise to a valid idea of causality, he does not accept Hume's conclusion that causality can have no valid empirical ori- gin. For Durkheim, social practices are empirical, are perceived, and can furnish an empirically valid origin for the concept of causality. When Durk- heim argues that the category of causality is not merely a generalization based on a series of perceptions of the social but, rather, can be immediately perceived as a general phenomenon in the single instance, he is striking at the heart of the epistemological problem. When he argues that the empirical (perceivable) effects of ritual action are more than the sum of the individual parts (i.e., that shared ritual has meaning and social or moral force, which the sum of its parts taken alone as particulars does not), he is again making a significant improvement in the empiricist posi- tion." For Durkheim the categories of the understanding and the social relations that produce them are not something added by the mind to a series of perceptions of natural objects and events but, rather, inherent characteristics of enacted ritual practices and thereby immediately avail-

"This is what Durkheim means by insisting that "society" is more than the sum of its parts. He is arguing a solution to the problem of general vs. particular, a point that has been missed. The "parts" are usually interpreted as individual persons, not individual perceptions, and the conclusion generally drawn is that if a group of persons is more than the sum of the individuals in the group then Durkheim must be articulat- ing a group-mind theory. He is not. It is rather an epistemological argument about the possibility of perceiving general ideas or categories of the understanding in enacted practices, which enact social forces, which are more than the sum of their parts.

able in perception, in each particular instance, to those participants who enact the ritual.

Essentially Durkheim picked up where Hume left off but with a sig- nificant twist. While Durkheim accepts the impossibility of an individual- ist empiricism, which Hume demonstrated, he argues that the solution can be found in a socioempirically based empiricism. By substituting en- acted social practice and perceptions of the moral forces generated in and through those practices by the individual participants enacting them for individual perceptions of natural phenomena, Durkheim avoids Hume's skeptical conclusion. Enacted practices do display their causes in their effects. However, as with Hume, the categories still have no valid applica- tion to natural science.

The general idea of causality is revealed by Durkheim as having been immediately available to perception all along on certain sorts of social occasions. It is this substitution of participation in enacted social practice for individual perception that not only provides the key to Durkheim's solution to Hume's dilemma but also, consistently through the years, in spite of the misunderstandings, comes to define what is distinctive about sociology as a discipline. The primary unit of analysis for sociology has not been the individual but various posited levels of social organization and empirically observable social practice.


Durkheim argues that the category of causality has its origins in those imitative rituals through which the totemic group is reproduced and through which the force of the group is achieved, enacted, and experi- enced. Imitative rituals, Durkheim argues, have as their central principle that "like produces like," that is, that, by imitating something, the thing imitated can be caused to reproduce. He argues that these imitative rites are causally efficacious in a singular way; the rites constitute, enact, and reproduce the very forces that they represent and therefore give persons who participate in them the experience of necessary force. In other words, the rites are what they represent: the relationship is one of equivalence and creation, not representation or correspondence (Durkheim 1915, pp. 393-414).

One of the problems Durkheim faces in making this argument is that his own interpretation of imitative rites, which is critical to his argument, differs sharply from the interpretation current in his day that imitative rites are based on a crude misapplication of the idea of causality. For instance, according to Durkheim (1915, p. 399), the empiricist perspective taken by Frazer (1887, 1889, 1910) and Tylor ([I8741 1973, 1899) treats the point of the ritual as the reproduction of the actual animal species

that the totem represents. Given this perspective, it appears that the mem-

bers of the totem have made a mistake because the totemic symbol cannot

cause the actual reproduction of biological animals. If the rite is inter-

preted as having successfully reproduced the totemic species only if it

causes the biological animal to reproduce, then the belief in the causal

efficacy of the ceremony seems so gross an error that it is hard to under-

stand. Hume7s argument that causality is beyond the capacity of the hu-

man mind and that customary and habitual ideas are being crudely ap-

plied to natural relations in the name of causality seems to have been


In order to establish his own position, Durkheim undertakes an elabo- rate and detailed consideration of the empiricist interpretation of totems. This discussion has been mistaken for a survey of totemic practices and a social constructivist account of the logic of totemic symbols, instead of being recognized as an essential part of his epistemological argument. Durkheim agrees with the empiricists that people are able to form crude causal connections between actions or events and effects based on faulty generalizations from experience. However, more than that would be needed, he argues, to explain these beliefs in empiricist terms. Taken strictly in classical empiricist terms, these rituals not only look like a false application of crude principles of contiguity and resemblance to natural events, as Frazer and Tylor argued, they make primitive peoples appear to be altogether devoid of reason. Durkheim argues that no one would believe in the literal causal efficacy of imitative rituals unless there was more to the ritual than the crude perception of custom, habit, similarity, and contiguity. Persons must get reinforcement for these beliefs in terms of some real causal efficacy of the moral forces involved because there is no literal causal efficacy with regard to natural events. They then misun- derstand the origin of this causal efficacy and transpose it onto natural or supernatural events or agencies, thus, generating religious beliefs.

If the true efficacy of the rite is, as Durkheim says, not the reproduction of the biological species but rather the reproduction of the moral energy of the group, then the belief in the rite is explained. The rite really is causally efficacious. If the rite reproduces the moral unity of the totemic group, then it does cause the moral reproduction of the totemic species (1915, p. 400). It is the effect on the men themselves, their feelings of well- being and moral unity, that give them the feeling the rite has succeeded, not any presumed magical action of the rite on natural or animal objects (1915, p. 402). Durkheim makes an important distinction between causal- ity as a relation between natural events that must be inferred from a series of incidents and causality as a moral force evident in any single case of enacted practice.

Durkheim argues that hidden within the principle of similarity, or "like produces like," unrecognized by the empiricists, is a principle involving creation or causality. This hidden principle, he argues, has two sorts of results: (1) to reproduce the moral community and (2) to reproduce the totemic species (1915, pp. 400-401). Durkheim argues that imitative rites concern the reproduction of something entirely new (1915, p. 399), which, therefore, cannot be explained on the basis of similarity and contiguity (the traditional empiricist explanation).

Because, for Durkheim, logic is itself the result of social practices, the question of whether or not the rite is actually causally efficacious must be considered an empirical matter of fact, not a matter of logic. Therefore, Durkheim argues that the possibility that the principle of causality is gen- erated in and through imitative rituals needs to be established through careful empirical examination. Instead of considering the principle in its general and abstract form, he writes, "let us connect it with the system of ideas and sentiments" that the "rites put into practice" (Durkheim 1915,

p. 400).

In service of this argument, Durkheim engages in an extensive discus- sion of the actual enactment of imitative rites. This discussion, like the other discussions of totems, is generally treated as a sociology of religion, and the relationship to the epistemology is missed. The purpose of this discussion, however, is to establish the concept of causality and demon- strate its relationship to particular enacted practices, not to explore the details of imitative rituals in their own right.

In imitative rituals, members of the totem imitate the animals or objects that the totem represents. Through the ritual, members of the totem not only represent and reinforce the totem, but they "make it" and "remake it" as a moral community. "Since they are emus or kangaroos, they comport themselves like the animals of the same name. By this means, they mutu- ally show one another that they are all members of the same moral com- munity and they become conscious of the kinship uniting them. The rite does not limit itself to expressing this kinship; it makes it or remakes it" (Durkheim 1915, p. 400).

Because it is the totemic idea that binds them together and that they all have in their minds, it is quite natural and necessary that there be representations of the totem accompanying the ritual. But that should not mislead us into thinking that the idea that these symbols can cause reproduction comes from their simple resemblance to totemic animals or that causal ideas are produced by the symbolism, an idealist interpreta- tion. The idea that totemic rites are causally efficacious results, according to Durkheim, from the fact that what the symbol represents is the actual fact of the members of the totem constituting and reproducing themselves as a moral community. They cause their own unity and identity through the enactment of the ritual.

By enacting their belief in the causal efficacy of the totemic symbol- displaying the totemic symbol to one another on their bodies and acting like the totemic animal during the ritual-members of the totem not only reaffirm but actually "remake" the kinship group in and through the ritual.

It also exercises a profound influence over the souls of the worshipers who take part in it. They take away with them a feeling of well-being whose causes they cannot clearly see, but which is well founded. They feel that the ceremony is good for them; and, as a matter of fact, they reforge their moral nature in it. How could this sort of well-being fail to give them a feeling that the rite has succeeded, that it has been what it set out to be and that it has attained the ends at which it was aimed? As the only end which was consciously sought was the reproduction of the totemic species, this seems to be assured by the means employed, the efficacy of which is thus proven. (Durkheim 1915, p. 402)

The feeling of participants that the rite has been successful is not merely personal. Durkheim argues that it is a general feeling shared among the participants that has a general source in the mutual enactment of practice. Through the ritual, the participants and members of the totem become stronger in their totemic feelings for one another. Their participation in the totem is strengthened. Therefore the totemic species is reproduced in them.

Because the community only exists in so far as the totems it represents are enacted, thus creating and sustaining shared ideas and beliefs, the rituals that make and reinforce the feeling of the causal efficacy of the totem also quite literally make and remake the moral community. Mem- bers of the totem who participate in the ritual have been reproduced as members of a moral community. It is the resulting feeling of well-being and moral unity that makes participants believe the rite has succeeded (Durkheim 1915, p. 402) not the biological reproduction of the totemic species. The causal relation is a social force available in direct perception, not an inference concerning the relation of the ritual to a natural event (e.g., the reproduction of the species).

If persons enact rituals to achieve mutuality, this is something of which they can be immediately aware. The rituals are their own effect. "It is because they serve to remake individuals and groups morally that they are believed to have a power over things" (Durkheim 1915, p. 414). Durkheim writes, "a full conception of the causal relation is implied in the power thus attributed to the like to produce the like" (1915, p. 406). Thus, the causal relation in imitative rites is immediately displayed in its effects, and the problem of inference from effects to causes that posed such a stumbling block for classical empiricism is avoided.

The causal relation involved in imitative ritual is a very different kind of causal relation from a physical/natural causal relation, and Durkheim carefully contrasts the two; the physical gestures involved in the ritual have no natural efficacy; their efficacy is purely social (1915, p. 402). If Durkheim's argument hinged on relations between natural events or be- tween social and natural events, the causal relation would depend on in- ference and thus be subject to his own critique of empiricism. Durkheim argues that the specific gestures that are enacted in the ritual are irrelevant and could be replaced by others. They have causal efficacy, not because of magical or scientific properties with which they act on objects or ani- mals, but rather because, as enacted representations of the shared totemic symbol of the group, they create moral forces that in turn create feelings of moral unity in its members, strengthening the group and hence the totem. "The true justification of religious practices does not lie in the ap- parent ends which they pursue, but rather in the invisible action which they exercise over the mind and in the way in which they effect our mental status" (Durkheim 1915, p. 403).

The causal relation that Durkheim speaks of is a moral one that he says operates on our "internal" states. Durkheim's discussion of "feelings" and "internal states" as a basis for knowledge, which has been disparaged as bad crowd psychology, in fact constitutes an argument of some conse- quence. Feelings of well-being and moral force are one of the results of totemic rituals that give rise to the category of causality. The causal effi- cacy of the ritual is perceived directly as a feeling of moral unity. Durk- heim argues that, because these feelings are manifest internally, they are immediately available and, therefore, can be known directly rather than indirectly as with external objects. Knowledge of these internal states is, therefore, better and more valid than knowledge of external states of af- fairs.

Hume (1975, p. 78) had also called causality an internal feeling resulting from a person's feeling about a collection of particulars in a series. How- ever, for Hume such feelings were necessarily individual. Durkheim (1915,

p. 408) acknowledges that purely personal and individual feelings are in- communicable. However, he argues that the feelings of moral force and well-being generated by totemic rituals constitute a perception of general social forces shared by members of the group (see Rawls 19966). They are the product of cooperation with others in the enactment of totemic rituals: "Being the work of all, they are not the possession of anybody in particu- lar" (Durkheim 1915, p. 408). Causality is actually a quality of the social practice itself, or, to paraphrase Hume, the effect does contain and display its cause. All participants will feel (i.e., perceive) it thus, and all will get the same idea or category. As in Suicide ([I8971 195 I), Durkheim argues that these internal or "psychic" states have a common shared socioempir- ical origin in social facts that explain and cause those internal psychic

states. They are collective forces even though internal and psychical "so-

cial forces: they are a part of our internal life" (Durkheim 1915, p. 408).

It is important to note that not all collective feelings are general in this

special sense, only those corresponding to the six categories, which Durk-

heim calls moral forces.

The moral forces generated by enacted practice are internalized as cate-

gories. They have an external social source in totemic rituals whose pri-

mary purpose is to generate these feelings, so as to provide a shared basis

for reason. The function of religion, according to Durkheim (1915, pp.

465-66) is to produce this logical basis for understanding. The moral

forces do not have their source in a "mystical collective conscience" or

group mind operating via some sort of constraint, as the theory has often

been interpreted (Nisbet 1974; Parsons 1968), but are, rather, the practical

result of concrete social practices.'' The feeling of moral unity is the pur-

pose of the ritual. It is not in the mind and not added by the mind. The

feeling is the substance of the ritual and therefore external, general, and

shared, not internal and individual.

Thefeeling is a perception of the efficacy of the ritual. Just as particular impressions of color and shape are perceptions of external states of affairs, feelings of the creation of moral unity are also perceptions of external social states of affairs that are quite "real" both in substance and effect. In this case, however, Durkheim treats feelings resulting from social ritu- als as having greater empirical validity than perceptions of external ob- jects because they are immediately perceptions of general social or moral forces, which act on persons internally, whereas perceptions of external objects remain particular and still require the problematic operation of inference to become general.

This argument is of some significance to understanding Durkheim's claim that socially derived categories are more valid than the categories of natural science and logic, which appears to many critics to be illogical, unfounded, and based on a rather fuzzy psychology of collective conscience. The argument is really quite logically tight. Whereas natural forces cannot be perceived directly, social forces can, and the perception of them can therefore be validly shared with others. The shared perception of moral forces in and through enacted practice is a better sort of knowl- edge than the individual perception of particular instances of natural forces that natural science and natural philosophy must depend on.13

See discussion of constraint in n. 23 below. l3 See Vico ([I7441 1948) for an earlier, also much neglected, argument that knowledge of the social has potentially greater validity than knowledge of natural events. How- ever, while Durkheim focuses on the moral force of collective action as the deciding factor between the two, Vico focuses on immediate knowledge of intentions, an obvi- ously problematic notion.

Durkheim is left with a degree of relativity as the categories may vary to some degree between groups, and in any case the argument only provides empirical validity for six categories-thus, his need to elaborate a sociol- ogy of knowledge. But, according to Durkheim, the purpose of the catego- ries of the understanding was not to secure absolute truth or transcenden- tal validity across groups. The categories did not evolve to fulfill philosophical purposes.14 The categories and the social practices that pro- duce them came into being to fulfill the social need for shared categories of the understanding among members of the same group. He argued that without shared categories, intelligibility and social cooperation would be impossible: "If men did not agree upon these essential ideas at every mo- ment . . . all contact between their minds would be impossible, and with that, all life together" (1915, p. 30). This is an early argument that the needs of intelligibility place demands on the development of forms of so- cial practice. There is no reason why the categories cannot exhibit a cer- tain degree of variation from place to place and still fulfill that social need. However, because they have their basis in experience, translation between groups should be possible even where variation does occur.


Classification is an even more basic concept with regard to epistemology than causality: the problem of identity and difference going back at least to early Greek thought. Unfortunately, Durkheim's argument with regard to classification, while it has received more attention than the argument for causality, has been understood primarily as a social constructivist ac- count of the social origins of particular systems of classification, and its importance with regard to an argument for the empirical validity of classi- fication as a category of the understanding has been missed."

l4 Durkheim's argument that the categories of the understanding do not originate for philosophical purposes should not be taken to mean that Durkheim's own argument is not a critique of philosophy (see, e.g., Hilbert's [1992, pp. 79-80] misinterpretation of this point). Durkheim intended his epistemology to make traditional philosophy obsolete. However, the categories of the understanding that society produces may not entirely satisfy a conventional philosophical interest in epistemology because that is not their purpose. Their purpose is to make intelligibility and social cooperation pos- sible.

One AJS reviewer said that my argument with regard to the philosophical impor- tance of Durkheim's idea of "classification" was wrong because, among other things, in the original French, Durkheim used the word "genre," not classification. The re- viewer said that attention to the original French edition would show that Durkheim (1912, p. 205) used the word "class" only once. It is sufficient to point out in response that in the section between pp. 200-22 (cited by the reviewer), which is chap. 3 of book 2 in the original French edition, Durkheim uses the French terms "classifica- tion," "classe," "classee," and "classes" (with or without accent marks) no fewer than 29 times in the text and notes. While the terms "class" and "genre" appear some-

The neglect of this argument may be due to the fact that Durkheim

dealt with the issue of classification on three levels, which he did not dis-

tinguish clearly. First, he argued that the ability to perceive similarity

and difference in crude terms that do not have empirical validity is a

capacity that humans share with animals (1915, p. 170). This parallels his

discussion of a similar capacity with regard to causality (1915, p. 410).

Second, he explored the development of social systems of classifications

or "cosmologies" patterned on divisions in social relations. This is a pri-

mary focus of Durkheim's sociology of knowledge, and classifications at

this level are not empirically valid. Third, he argued for the development

of the category of classification through the direct perception of moral

force in the enactment of those practices that create the binary relations

of sacred and profane, totem and nontotem. This is his epistemological

argument for the empirical validity of the category of classification (1915,

pp. 174-75).

There is a great deal of misunderstanding regarding Durkheim's treat- ment of the category of classification, and it has generally been interpreted only in terms of the second of the three levels as a sociology of knowledge concerned with documenting systems of classification corresponding to particular cultural groups. This misunderstanding treats the category of classification as a group of collective representations that develop over the course of a series of perceptions of cultural events. Because it is viewed in this way, Durkheim appears to fall into the old empiricist dilemma of trying to abstract empirically valid general ideas from a series of particu- lars.

The criticism that Durkheim's arguments for time, space, and classifi- cation are circular, which is cited with such regularity (Dennes 1924; Schaub 1920) is based on this misunderstanding. Durkheim is interpreted as arguing that repeated perceptions of social instantiations of space and time give rise to the categories of space and time through a process of abstraction. Because a series of particulars cannot give rise to a valid general idea, Durkheim is accused of philosophical nai'vetk. In addition, critics point out that persons would need already to have ideas of space and time in order to perceive the instantiations of space and time, which are supposed to give rise to those ideas as instantiations of space and time.

In making this argument, the critics confuse the first and second levels of Durkheim's argument with the argument for the empirical validity of the categories of the understanding. The argument that animals can per- ceive crude relations of space, time, and difference or similarity (first level) and the argument that the periodicity of social rituals and social divisions

times to be used interchangeably, Durkheim certainly meant to signify the idea of classification as a form of thought, not just as an idea representing a kind of thing.

can be turned by the crude capacity to discern time, space, and class into classifications based on social forms (second level) are not the basis for Durkheim's argument for the socioempirical validity of the category of classification (third level). While Durkheim (1915) allows that persons have the capacity to form invalid general ideas before the development of the categories, these invalid forms are not the basis of the categories. A valid category of classification cannot evolve from ideas that have no validity.

An empirically valid category of classification only develops, according to Durkheim, when practices that enact moral relationships produce feel- ings of moral force. Only after the development and enactment of the first moral division producing moral force is any complex level of social organization possible (1915, pp. 174-75). Durkheim explicitly rejected the argument that valid ideas could be abstracted from an accumulation of particular perceptions, a position that he identified with classical empiri- cism. A cosmology with no empirical validity could theoretically be ab- stracted from an accumulation of perceptions. (1915, pp. 171-72). But, even then, according to Durkheim, a society complex enough to generate such a cosmology would require prior development of empirically valid categories, which Durkheim considered a prerequisite for any complex social development. From Durkheim's perspective, the critics' own social constructivist argument is circular because it assumes that complex societ- ies are operating on the basis of fictive shared realities. Durkheim argued that societies cannot develop unless they can first solve the problem of mutual intelligibility through the generation of valid categories.

Because he has been interpreted as elaborating systems of classifica- tions, a sociology of knowledge argument, and not as arguing for the em- pirical validity of the category of classification, Durkheim has been heavily criticized both for the inaccuracy of the data on which he based his argument, accuracy being in a sense the whole point for a sociology of knowledge, and for not pursuing a consistently social constructivist position. The third level of argument has been generally overlooked and the epistemological implications that were the purpose of making the ar- gument in the first place continue to be dismissed with regularity. Thus, it is generally held that Durkheim's argument does not explain the origin of an empirically valid category but only the development of particular uses of concepts loosely labeled as classifications in various societies; a sociology of conceptual systems or sociology of knowledge.

Certainly Durkheim does elaborately document systems of classifica- tion. However, in both the earlier work on primitive classification (Durk- heim and Mauss 1963) and in The Elementary Forms (1915) the empirical detail is not aimed at elucidating various social systems of classification per se but rather toward evaluating whether such classification systems

might have provided in and through their enactment for the development

of this category. The relation of the discussion of various classification

systems to the category of classification is the same as that between the

various discussions of totems and the categories of force and causality.

Because an empirically valid category cannot emerge over a series of

perceptions, the category of classification must develop from a moral divi-

sion enacted in social practice, which is experienced as a moral force in

a single instance and not as a natural division perceived over a series of

particulars. For Durkheim the first moral division to be enacted becomes

the first binary social division from which all other classifications follow.16

Totemic practices, according to Durkheim, enact this moral division in human social relations such that persons can immediately perceive it in enacted practice. Although totems are a more complex form of social divi- sion than the first division between sacred and profane, they are the closest to the original of those practices that are still enacted. From one original division of social relations into two, totem and nontotem, the category of classification and subsequent divisions could develop.


In order for totems, or clans, to be able to provide a social origin for the category of classification, it must be the case that totems themselves are social and not individual in origin. If totems had an individual origin, Durkheim thought that his argument for the generation of the concept of classification would be circular and subject once again to the empiricist dilemma. "If [individual totemism] was the primitive fact, we must say that religion is born in the consciousness of the individual, that before all else, it answers to individual aspirations and that its collective form is merely secondary" (Durkheim 1915, p. 200). Consequently, Durkheim en- tertains an extended discussion of the origins of totemism, which has two different objectives:jifirst, to demonstrate that individual experience is not the source of totems and, therefore, that an individualistic approach to understanding totems and religion in general is invalid; and, second, that the concept of classification is based on totems as enacted practices and therefore is valid (in Durkheim's special sense) and not circular in the empiricist sense. When these two purposes are not both understood, Durk- heim's elaborate argument for the social origins of totems seems once again to be a social constructivist account of religious cosmology instead

l6 This was the point of departure for LCvi-Strauss (1963). However, he missed the origin of this binary opposition in shared enacted practice and located it instead in shared belief. This led him to rest his argument on a universe of belief structures, a position from which we are still struggling to disentangle ourselves theoretically.

of a proof that the enacted practices do not themselves have an individual

origin so that he can argue for the empirical validity of the categories that

they generate.

In addressing thefirst point, Durkheim developed a sociology of knowl- edge of classification systems, the point of which was to show their origin in social as opposed to individual experience. Durkheim argued that as social groups developed an internal organization, they divided the parts of nature and society into classifications, called totems, which became part of the internal organization of the group. All things, animal, vegetable, and mineral, were assigned a totem: "The unity of these first logical sys- tems merely reproduces the unity of the society" (1915, p. 170). Religious or totemic order does not reflect the individual perception of natural order. Rather, the classification of natural phenomena is based on familiar divi- sions in organized social life. "In the Mount Gambier tribe . . . there are ten clans; consequently the entire world is divided into ten classes. . . . when brought together, these ten families of things make up a complete and systematic representation of the world" (1915, p. 179).

It is the confusion between the divisions of social forms and the divi- sions of nature in the initial totemic classifications, according to Durk- heim, that first led to the application of this idea to nature. For instance, all things belonging to the same totem (animal, vegetable, and mineral) are classified together by virtue of being related to a common ancestor, an idea clearly derived from social relations as it could have no possible basis in nature. "In all probability, we would never have thought of unit- ing the beings of the universe into homogeneous groups, called classes, if we had not had the example of human societies before our eyes, if we had not even commenced by making things themselves members of men's society, and also if human groups and logical groups had not been con- fused at first" (Durkheim 1915, pp. 172-73).

The equation of natural and social principles of order is a confusion because the classifications reproduce social relations and impose them on nature where they do not belong. But it is also a useful confusion. The social construction of classifications provides a way of categorizing things in nature that is shared to a degree. As Durkheim (1915, p. 169) writes: "It is because men were organized that they were able to organize things."

Without classifications based on social divisions, persons might not have found the crude relation of resemblance to be a significant one. "Men would never have thought of arranging their knowledge in this way if they had not known beforehand what a hierarchy was. But neither the spectacle of physical nature nor the mechanism of mental associations could furnish them with this knowledge, the hierarchy is exclusively a social affair. It is only in society that there are superiors, inferiors, and equals. . . . It is society that has furnished the outlines which logical

thought has filled in" (Durkheim 191.5, p. 173).

The ability to perceive resemblance and difference between natural

entities is basic, and even animals perceive things as being of different

types. However, totemic classifications introduce the idea of hierarchy,

which is a social or moral organization not existing in nature. Thus, all

classifications involving the notion of hierarchy must have a social origin.

In addressing the second point, Durkheim makes the entirely distinct argument that the category of classification as generated by enacted prac- tices is empirically valid. The category of classification itself is not con- structed on the basis of the many and varied classifications that appear in any given society but rather on the basis of feelings of moral force that the enactment of these social divisions in and through totemic rites give to participants. The enactment of the totemic rites by members of the totemic group enacts "an internal bond [which] attaches them to the group in which they are placed" (1915, p. 174). This provides the original experi- ence of a binary division between totem and nontotem, sacred and pro- fane: "A bond of mystic sympathy unites each individual to those [totemic] beings, whether living or not, which are associated with him" (1915,

p. 174).

Durkheim argues that the experience of social division as moral force becomes the basis for the empirically valid logical construct of classifica- tion providing the mind with a way of thinking that is not tied to the conditions of its conception: "This organization, which at first may have appeared to us as purely logical, is at the same time moral" (1915, p. 175).17 The basis for the logical relation between beings in the same classification is their moral relationship: the beings in a totem are "really" the same because the same moral force binds them together.

According to Durkheim, it is the focus on individual perception and the application of the category of classification to natural objects, events, and forces, with regard to which it has no validity, that has made the problem of empirical validity appear to be unsolvable both in philosophy and sociology. This sort of individualist empiricism can only explain the development of what Durkheim calls "generic ideas" in the individual organism. Generic images and "logical symbols" like classification are not the same thing. According to Durkheim (1915, p. 172), animals can make generic representations: "An animal is able to form generic images though ignorant of the art of thinking in classes and species." Only beings who enact moral force develop categories.

I' The feeling of moral force with regard to classification only provides a socioempirical basis for the validity of the idea of classification itself. It does not guarantee the validity of any particular classifications, which are often extremely arbitrary.



Durkheim's distinction between the basic ability of individuals to notice

similarity and difference that has no empirical validity (first level of the

classification argument) and empirically valid categories of thought (third

level of the classification argument) has been interpreted by his critics as

a dualist position. The charge of dualism, however, like the charge of

circularity, is based on misunderstanding. Dualism is a name for various

arguments concerning the gap between thought and reality or mind and

body. Generally the argument is that the nature of mind and body, or

spirit and matter, are so different as to be incompatible. For instance, in

Cartesian dualism the rational mind can only have knowledge of what is

rational in the universe, not physical matter (unless physical matter is

conceived as purely rational, an idealist position). In sociology the term

dualism is often used to denote the gap between individual and society.

Durkheim's alleged dualism is largely a result of misunderstanding his attempt to distinguish the crude animal ability to perceive contiguity, dif- ference, and similarity from the valid category of classification combined with misreadings of an earlier article, "Individual and Collective Repre- sentations," in which Durkheim ([I8981 1953) argued against what he characterized as James's psychological reductionism. Interpretations of this article consistently confuse Durkheim's own position with the indi- vidualist position he attributed to William James (Rawls 1996~).

Durkheim allows for a basic empiricism within individual perception that is not valid knowledge but is nevertheless in touch with reality. "It is not our intention to deny that the individual intellect has of itself the power of perceiving resemblances between the different objects of which it is conscious. Quite on the contrary, it is clear that even the most primi- tive and simple classifications presuppose this faculty" (Durkheim 1915,

p. 170).

While this argument is generally considered evidence of Durkheim's Cartesian dualism (Lukes 1973), in fact, Durkheim is arguing that the biological individual is not capable of the exercise of reason-not a Cartesian position at all. The Cartesian argument is that the logic of rea- son is, first of all, a given and, secondly, completely incompatible with the organization (or lack thereof) of the "corruptible" world of nature that approximates but always falls short of the pure light of reason. According to Durkheim, the biological individual only possesses a basic ability to sort perceptions according to resemblance and contiguity, which goes no further than Hume's custom and habit and certainly does not produce valid knowledge.

In addition, the argument that persons possess the innate ability to form generic representations ought to satisfy those who have criticized Durk-

heim's argument as circular since Durkheim clearly states his belief that

nonvalid ideas of space, time, class, and causality precede the develop-

ment of valid categories of thought. Yet, critics who charge Durkheim

with dualism and therefore must be familiar with this argument also

charge him with circularity.

In contrast to the general interpretation of his position as dualist and

neo-Kantian, Durkheim in fact makes the highly original argument that,

prior to participating in ritual practices, persons have only the basic abili-

ties of animals to make generic representations, which are similar to the

faculties that the empiricists argued for, whereas, after experiencing the

enactment of ritual social practices, they have empirically valid categories

of the understanding. Durkheim's argument that categories of thought

have an empirical origin is completely incompatible with both the Kantian

and the Cartesian positions that have been attributed to it.

While Durkheim argues that natural relations cannot be the origin of the categories of the understanding, arguing instead that they have their origin in social relations, he also maintains that social relations are natural relations of a special sort and therefore believes that there is no deep in- compatibility between reason and the empirical world of nature or be- tween individual and society: "If experience were completely separated from all that is rational, reason could not operate upon it; in the same way, if the psychic nature of the individual were absolutely opposed to the social life, society would be impossible" (Durkheim 19 15, pp. 28-29). Durkheim does acknowledge a gap between reason and perceptions of natural forces and argues that categories could not arise from perceptions of natural events. However, he also argues that the categories are not random with regard to natural order. There are two reasons that he gives for this. (1) The social order, which gave rise to the categories, is a natural order itself and as such obeys natural laws. Therefore, socially generated categories are in some sense natural categories (this would not be true for cosmologies which are ideal). (2) While there are limits to the certainty of knowledge that sense impressions can give rise to, they do allow for a sorting of things by resemblance, which, while not empirically valid, places limits on the social re-creation of nature.

If categories of the understanding that develop from social processes are added to the natural ability to perceive resemblance, Durkheim (1915,

p. 486) feels that there is then no reason that sorting by resemblance could not have some limited empirical validity with regard to natural phenom- ena. The validity would not be equal to that for social phenomena but not completely out of touch with natural reality either.

For instance, Durkheim (1915, p. 170) argues that the division of things into "opposites" appears in so many societies because a high degree of contrast leads to a high degree of natural visibility and "intuitions" and

"feelings of affinity or of repulsion" for the contrasts. He says there "is a

certain intuition of the resemblances and differences presented by things"

and that this has played an important part L'in the genesis of these classifi-

cations [of opposites]" (1915, p. 170). However, the division by resem-

blance of natural things is not valid knowledge. Resemblance and contigu-

ity can only give rise to an opinion or habit of thinking in a particular

way. The categories can only develop when the social being joins in assem-

bly with its fellows to witnessably enact the moral divisions of their


Critics have generally interpreted Durkheim's discussions of individual perceptions and the distinction between the social self and the biological organism as evidence that Durkheim is a positivist, a dualist, a realist, or a Cartesian rationalist. However, the distinction is no more dualist than George Herbert Mead's I versus me distinction to which it bears striking parallels. Both Mead and Durkheim allow rudimentary abilities and an active principle that precedes the development of a social self. Durkheim's point is that only the social being has empirically valid categories of the understanding. The biological individual has only rudimentary abilities that could never supply the framework for knowledge or mutual intelligi- bility: "The feeling of resemblances is one thing and the idea of class an- other. The class is the external framework of which objects perceived to be similar form, in part, the contents. Now the contents cannot furnish the frame into which they fit. They are made up of vague and fluctuating images, due to the superimposition and partial fusion of a determined number of individual images . . . the framework, on the contrary, is a definite form, with fixed outlines, but which may be applied to an undeter- mined number of things, perceived or not, actual or possible" (Durkheim 1915, pp. 1'71-72).

A realist position would posit the natural validity of the categories while a rationalist would hold that they are perceivable by any individual ratio- nal mind. Durkheim denies both of these positions. Durkheim argues that feelings of affinity and repulsion by themselves cannot lead to the develop- ment of the category of classification. They are only a feeling about natural relations between particulars, they cannot give rise to categories that are by definition general. The positivists and realists treat these feelings of affinity and repulsion as the only possible origins of empirically valid knowledge. Durkheim's is, rather, an antipositivist, antirealist position because the natural abilities that he posits have no empirical validity and

l8 "Witnessably" and "witnessable," as used here, refer to a quality of an object or event, not whether an object or event is witnessed. Moral divisions, for instance, are not generally witnessable, nor are feelings or concepts. They are rendered witnessable through their enactment in social practice.

cannot give rise to categories of the understanding but only to feelings of

resemblance as they do for Hume.

On the other hand, while Durkheim insists on a socially based rational-

ism, he is not a Cartesian rationalist as Lukes (1973) asserts. A Cartesian

rationalist would posit a thoroughly rational universe to begin with and

argue that knowledge of that world must be derived from a few basic

principles of reason. Durkheim argued that principles of reason all have

their origin in the moral forces experienced while participating in certain

enacted practices and that knowledge derived in this way has only limited

validity when applied to the natural world.

An a priori, or Kantian, explanation of the development of the catego- ries of the understanding has in common with rationalism the assumption that the categories are purely logical in character. In contrast, Durkheim (1915, p. 175) argues that the organization of the world into totems that "appeared to us as purely logical, is at the same time moral." In addition, a Kantian position assumes that the categories are innate and precede all understanding, whereas Durkheim allows that both crude abilities to perceive contiguity and resemblance and crude social behaviors precede the development of the categories. Thus the categories are empirical his- torical necessities not logical necessities. Logic itself must follow the devel- opment of the categories. Therefore, arguments for the status of various ideas as categories should be made empirically and historically not logi- cally, according to Durkheim. It is as enacted practices that totems lead to the development of the category of classification, not as elementary systems of classifications. The category of classification, which the enact- ment of totemic rites generates, then becomes a basis for logic, not the other way round, and, while Durkheim argues that the concept of classifi- cation probably needed to come before the other categories, he is careful to make a historical and not a logical argument for its priority as he feels that logic itself is a social product and cannot be introduced into the argu- ment at this point (1915, p. 169).19

l9 For Durkheim (1912, p. 2 11)' logic is something that remains to be socioempirically established, which is why he does not make a logical but rather an empirical argument for the development of logical categories of thought: "C'est la sociCtC qui a fourni le canevas sur lequel a travaille la pensCe logique." I use the original French here because both available translations obscure Durkheim's sense that logic is a social creation. Swain translates it as "It is society that has furnished the outlines which logical thought has filled in" (Durkheim 1915, p. 173), while Fields translates it as "Society furnished the canvas on which logical thought was worked'' (Durkheim 1995, p. 149). I take Durkheim's meaning to be that logic is a social creation, painting, or labor, worked within the constraints of a social canvas. He certainly does not mean to convey a preexisting entity-logic-that works within a social framework. See also Durk- heim's (1955) criticism of James for making use of logic in his argument while at the same time arguing that logic varies from moment to moment.

The argument that logic is the result of a social process beginning with the simplest totemic rites is not Cartesian or neo-Kantian. This is an en- tirely new argument regarding the social origins of logic. The categories, according to Durkheim, serve a social purpose in giving rise to a shared logic, and society depends on this. Society needs shared enacted practices to produce categories of the understanding without which cooperation and communication would not be possible (Durkheim 1915, pp. 30, 465, 482, 487). The primary purpose of religion is to provide the enacted prac- tices necessary to produce the categories (1915, p. 467). All societies will have the categories, according to Durkheim, because, if they cannot pro- duce the categories, they cannot exist as societies. Therefore, wherever societies do exist, they will have generated these categories: "Society could not abandon the categories to the free choice of the individual without abandoning itself. If it is to live there is not merely need of a satisfactory moral conformity, but also there is a minimum of logical conformity be- yond which it cannot safely go" (1915, p. 30).

Religious cosmology is a way of representing moral force, but religious practice is a way of creating it. This is the central distinction that Durk- heim makes between the sociology of knowledge and the epistemology: there are practices (social facts), which create categories, and there are cosmologies (collective ideas), which explain and represent them. The cat- egories are empirically valid. The cosmologies are not.

Locating the basis for science and philosophy in aboriginal social and religious practice leads Durkheim to a respect for what he refers to as "primitive intelligence," which is rare for the period in which he wrote. Durkheim (1915, pp. 15, 166, 203, 219, 361-62, 476, 487) argues that "primitive" peoples are not lacking in logic; they have their own logic. It is only by assuming a universal standard of reason and logic that they are found to be deficient. This deficiency is a social construction for Durk- heim, not reality. LCvy-Bruhl had difficulty following Durkheim on this point. He refused to give up the primacy of logic, which led him to argue that "primitive" intelligence was deficient in essential respects. Even the relativist notion of truth advanced by the pragmatists finds deficiencies in primitive religion, judging them by the principle of utility without an appreciation of the social and epistemological importance of what primi- tive rituals achieve.


That Durkheim intended The Elementary Forms to forever change the face of philosophy by offering an alternative, nonphilosophical (socio-logi- cal), solution to the problems posed by the debate between the empiricists

and the apriorists is clear: "Thus renovated, the theory of knowledge

seems destined to unite the opposing advantages of the two rival theories,

without incurring their inconveniences. It keeps all the essential principles

of the apriorists; but at the same time it is inspired by that positive spirit

which the empiricists have striven to satisfy" (Durkheim 1915, p. 32).

Durkheim expected the new science of sociology to become the center

of both scientific and philosophical enterprise. Sociology would replace

philosophy as the proper arena for epistemological debate. Yet, clearly this

has not happened. Sociologists continue to rely on epistemologies based on

individualism, and philosophers continue to criticize the epistemological

basis of sociology (Winch 1956; Rorty 1979; Turner 1994).

From the beginning, the confusion of Durkheim's epistemology with his sociology of knowledge has led to the dismissal of the epistemology as a very bad argument. Durkheim made claims of empirical validity with regard to his epistemology that appear absurd when applied to his sociol- ogy of knowledge. In addition, not recognizing the epistemological sig- nificance of the extended discussion of totems leaves critics with the impression that the epistemological argument is scattered and brief. Gol- denweiser (1915, p. 733)' one of the earliest American critics, wrote: "While the author's remarks on that subject are not extensive nor systematic . . . the author's attempt to derive all mental categories from specific phases of social life which have become conceptualized, is so obviously artificial and one-sided that one finds it hard to take his view seriously." It was Goldenweiser's opinion that Durkheim's epistemology failed because it presupposed a complex social and conceptual system that was not available in all societies. Like many early critics, Goldenweiser's prejudice with re- gard to the sophistication of primitive cultures combined with his misinter- pretation of Durkheim's epistemological argument as it concerned conceptual systems kept him from appreciating Durkheim's point.20

ZoGoldenweiser(1915) argues that Eskimos and other primitive groups are lacking in even the rudimentary social classifications necessary to support Durkheim's argu- ment. In part this is due to a misunderstanding of the role that totems play in Durk- heim's argument. But it also has clear overtones of a general presumption of the superiority of Western forms of thought. Similarly, Dennes (1924, p. 52) makes the blatantly racist argument that Negroes and Eskimos in the United States prove Durk- heim's theory wrong: 'Tt is only members of closely similar races that can be trans- ferred with such results as Durkheim describes. A group of Eskimo or Negro infants introduced into an English community would not develop the same qualities of mind and sustain the same relations to social customs and institutions that a like number of German or French children would, as the experience of the United States of America has shown." There was a general agreement among the early commentators that Durkheim had gone too far in attributing a basic equality to all human beings. This L'fault" is very much to Durkheim's credit. Durkheim insisted, in contrast to

A few years later, Dennes published a much cited critique in which he pronounced Durkheim's argument circular and confused:

Durkheim's theory of the origin of the categories depends upon his ambigu- ous conception of mind. If he takes mind in the Kantian sense, the sense usual in epistemology, as the subject's system of cognitive faculties, it is ridiculous to say that the categories of the mind are in any sense transfer- ences from social organization. The category of quantity would have to exist and to operate in order that an individual mind should ever recognize the one, the many, and the whole, of the divisions of his social group. . . . If, on the other hand, Durkheim means by mind a mere aggregation of repre- sentations or ideas, there is sense in supposing that the first ideas of time may have been of the periodicity of primitive religious rites, the first ideas of quantity, the division of the tribe, etc. But the supposition is then of merely historical importance, so far as it is of any importance at all. It has no direct bearing upon either the epistemological or the psychological study of the nature or status of the categories of the mind. (Dennes 1924, p. 39)

Dennes argued that Durkheim's epistemology was only valid as a theory of contents, or sociology of knowledge, not as a theory of the form of the mind. He pronounced the epistemology a failure (Dennes 1924, p. 53). This analysis by Dennes has been extremely influential. It was a major influence on Parsons's reading of Durkheim and is still regularly cited by commentators on Durkheim's epistemology and/or sociology of knowl- edge.

It is difficult to reconcile the careful argument that Durkheim actually made with the interpretation it has generally received. One gets the im- pression that, following an initial round of misinterpretation between 19 10 and 1925, the epistemological argument was never reevaluated (1996~). The overwhelming consensus is that Durkheim's epistemology is a com- plete failure. The view consistently presented is of a Durkheim who either never realized that his later sociology of knowledge contradicted his early empirical studies (Parsons 1968), or of a man who died too soon to fix this problem (Stone and Farberman 1967; Hughes [I9581 1977; Parsons 1968). There is no recognition that both the sociology of knowledge and the ear- lier empirical studies rest on an epistemology that supersedes both. Schol- ars are extremely critical when discussing their disappointment with what they see as the unresolved contradictions in Durkheim's work, which they attribute to philosophical naYvet6. In book after book, article after article, critics complain that Durkheim has not fulfilled the expectations of good scholarship.

Critics often express the feeling that they have been let down, that they

LCvy-Bruhl, that "primitives" are not less intelligent than other human beings but rather that their forms of thought are a product of their religious practices. The preva- lent belief in racial inequality obviously prevented an appreciation of Durkheim's insistence on a basic human sameness in capacity of mind.

expected more. Dominic LaCapra (1972, p. 287) even goes so far as to

suggest that Durkheim's argument is "like the conglomerate body of some

incredible Frankenstein monster." Given the broadly acknowledged im-

portance of Durkheim's work, the frustrated commentaries on his episte-

mology stand out in bold relief.

Parsons, who obviously respected Durkheim's work, defending him against several prevalent misreadings and building structural functional- ism on his interpretation of it, nevertheless expressed exasperation and even embarrassment over Durkheim's epistemology: "His epistemology has brought the basis of human reason itself into the same relativistic circle, so as to make the previous relativism itself relative, since the rela- tivism of social types is itself a product of a system of categories which are valid only for the particular social type. This is a doctrine which may be called "social solipsism." It involves all the skeptical consequences which are so well known in the case of individual solipsism. It is in short, a reductio ad absurdum" (Parsons 1968, p. 447). These criticisms by Par- sons of Durkheim's epistemology as circular and solipsistic, themselves drawn from Dennes (1924) and Gehlke (1915), have been extremely influ- ential.

LaCapra reflects the general trend in failing to distinguish either be- tween Durkheim's sociology of knowledge and his epistemology or be- tween categories and collective representations, arguing that "collective representations," in Durkheim's usage of the term, seemed to cover the gamut from shared verbal behavior based on deeply rooted beliefs, through elaborate "ideologies," to more or less sophisticated theoretical reflections." (LaCapra 1972, pp. 265-66). LaCapra (1972, p. 281) con- cludes: "Unfortunately, in confiding in something as palpably ineffective as social metaphysic, Durkheim dissipated both his massive intelligence and his genuine spiritual intensity. From him one might have expected a more convincing attempt to forge a synthesis between uncoordinated elements of the modern experience and the heritage of devalued symbolic forms."

It is also common to judge the epistemology on the basis of texts in which it does not appear. The early critics Gehlke (1915), Schaub (1920), and Dennes (1924) based their criticisms primarily on an article published 14 years prior to The Elementary Forms, "Individual and Collective Rep- resentations" (Durkheim 1953), which was a critique of James, not a pre- sentation of Durkheim's epistemology (Rawls 1996~). One of the only books devoted to Durkheim's epistemology, Paul Hirst's (1975) Durkheim, Bernard and Epistemology, presents an analysis of Durkheim's epis- temology based on The Rules of the Sociological Method, in which Durk- heim does not make an epistemological argument. Predictably, Hirst (1975, p. 5) pronounces Durkheim's position untenable: "The epistemol-

ogy developed in The Rules of the Sociological Method is an impossible

one." Allcock (1983), one of the few to defend Durkheim's epistemology,

argues that the practice of deducing the epistemology from the earlier

empirical works instead of from the later work is prevalent. Unfortu-

nately, Allcock himself focuses on the lectures on pragmatism as the best

source for Durkheim's epistemology instead of on The Elementary Forms.

This creates further problems as the epistemology is not presented in these


Lewis Coser (1971, p. 140) echoes the general theme, arguing that while "in the light of later critical discussion of the thesis it can be said that Durkheim failed to establish the social origins of the categories of thought," it is important to see that he made an important contribution to the sociology of knowledge. The consistent message is that, even though Durkheim failed miserably in establishing an epistemology, his empirical work and his sociology of knowledge can be saved.

In his introduction to The Rules, Lukes (1982) interprets Durkheim's epistemology as a focus on symbols without a corresponding interest in their hermeneutic dimension. This equation of the sociology of knowledge with the epistemology is particularly unfortunate because it leads Lukes to argue that Durkheim overlooks the dimension of meaning or hermeneu- tics. In fact, Durkheim's epistemology takes intelligibility as its central problematic. His discussion of meaning does not become hermeneutic, however, because he believes he has avoided the problems of indetermi- nacy and interpretation at the epistemological level by establishing empir- ically valid categories of thought. If the categories do not require interpre- tation, there is no hermeneutic. Lukes (1982, p. 15) misses this point, interpreting Durkheim's approach as "dictated by an obsession with an 'absolute' conception of objectivity, [which] can only be a sterile prescrip- tion for the human sciences."

Even when critics are friendly to what is termed Durkheim's "theory of knowledge," they have still interpreted his position in idealist social constructionist terms consistent only with his sociology of knowledge and not his epistemology. Their positive assessment is only due to the fact that they are themselves social constructivists or pragmatists who accept the indeterminacy of knowledge. Stone and Farberman (1967) are among the earliest such supporters of the allegedly idealist pragmatist Durkheim in the United States. While claiming to make a radical departure from previ- ous Durkheim scholarship, Stone and Farberman (1967, pp. 163-64) fol- low the general trend in confusing the epistemology and sociology of knowledge, leading them to the belief that Durkheim "was confounded and groped for a new epistemology and, implicitly, a new ontology." They argue that his alleged idealism was turning Durkheim toward pragma- tism.

More recently this argument that Durkheim's alleged idealism is his

more important contribution has become quite popular. It has been widely

advocated by Alexander (1988) and Collins (1988) with regard to cultural

studies, by Stjepan Mestrovic (1988, 1993) and Jennifer Lehmann (1990)

in connection with poststructural idealism, by Godlove (1989) in religious

studies, and by David Bloor (1982) and others in the sociology of science

who interpret Durkheim's sociology of knowledge as a social constructiv-

ist precursor to modern studies of scientific practice.

Hilbert's (1992) interpretation of both Durkheim and Harold Garfinkel as social constructivists continues this recent trend toward embracing the idealist interpretation of Durkheim's epistemology. According to Hilbert (1992, p. 78), it is "not at all hard to understand why Durkheim would draw this final equivalence [between society and ideas]." For Hilbert, Durkheim's epistemology involves a set of socially defined ideas that con- strain experience. Practices, on Hilbert's interpretation, are no more than "artful" ways of constructing lines of social action. There is no real pattern in practices, only an apparent pattern: "we see in members' artful manage- ment of indexical expressions a collective method of sustaining the folklore of relatively fixed underlying patterns" (Hilbert 1992, p. 5 1; see also pp. 56, 66, 75, 91). The fiction of a stable reality, on this view, depends on constraint, and constraint, according to Hilbert, is provided by ideal cate- gories. The categories are "made up" by society (Hilbert 1992, p. 78). There is no empirical reality corresponding to them, and no reality can be experi- enced without them. Therefore, truth is entirely relative to any particular society's ideal categories. Hilbert appears to be unconcerned by the lack of any independent criteria of truth that his interpretation of Durkheim implies.21 This idealist interpretation stands in direct contradiction to Durkheim's own epistemological argument, which was motivated by the need to specify truth criteria.

While Collins has contributed to modern Durkheim scholarship through his interactional interpretation of Durkheim's theory of practice and supports the relevance of Durkheim's sociology of knowledge to con- temporary social theory, he shares the idealist interpretation of Durk- heim's epistemology. Collins (1988, p. 108) maintains that The Elementary Forms has been generally overlooked by Durkheim scholars: "What is left out?" he writes. "Generally speaking, what I would consider the more valuable parts of Durkheim. His most important book, The Elementary Forms of the Religious Life . . . seems to be regarded as a specialized matter for sociologists of religion, though some references (usually very critical) are made to its social reductionist theory of knowledge." Like

Z' He is equally unconcerned with the implications of his interpretation of Garfinkel as an idealist social constructionist: an interpretation that is equally far from the mark.

Stone and Farberman, Collins considers the epistemology to be "social

reductionist," finding the value of Durkheim's work mainly in its rele-

vance for cultural studies.

Even Godlove, who directly addresses Durkheim's argument concern- ing the origin of the categories of the understanding in The Elementary Forms, confuses the sociology of knowledge with the epistemology. He argues that Durkheim's "idealist" epistemology is more sophisticated than the critics have given him credit for and that Durkheim had good reasons for trying to "trace the modal structure of the categories to an ideal object outside the world" (Godlove 1986, p. 385; emphasis in original). According to Godlove (1986, p. 390), Durkheim "turned to idealism as a way of overcoming what he perceived as the empiricist-apriorist dead- lock." Godlove alleges that, in the face of the empiricist critique, Durk- heim concluded that only a nonempirical (i.e., ideal) origin for the catego- ries was possible.

Durkheim's critique of empiricism was much more sophisticated than Godlove realizes and his epistemology completely empirical in that he considered social practices to be concrete witnessable phenomena and not ideal. While emphasizing the importance of distinguishing between the sociology of knowledge and the epistemology, Godlove nevertheless con- fuses the two. He has interpreted the epistemology in idealist terms, as a substitution of the ideal for the empirical, which are appropriate only for the sociology of knowledge. He concludes that "although Durkheim was keenly aware of the epistemological problem produced by his sociological theory, all his attempted solutions failed badly" (Godlove 1986, p. 400). In a footnote to this citation he says "the details are too well documented to bear repeating" (Godlove 1986, p. 400, 11.28).


So much disappointment with regard to a scholar who has led the disci- pline to expect so much and whose work has sustained interest over the course of a century requires an explanation. In this case the explanation involves a complex history of misinterpretations and reactions to those misinterpretations in both the United States and Europe. The critics from the very first got off on the wrong track, interpreting Durkheim as an idealist and confusing the epistemology and the sociology of knowledge. From there, Durkheim scholarship has consisted of debates between the various misinterpretations, none of which question the initial assessment of his position as idealist.

While the European and American interpretations of Durkheim's work have for most of this century appeared to be entirely at odds with one another-the Europeans interpreting Durkheim in idealist and social con-

structivist terms, while, in the United States and Britain, a positivist func-

tionalist interpretation based on his earlier work developed-the very first

scholars in both Europe and the United States interpreted Durkheim in

terms of his later work, particularly The Elementary Forms, which they

considered to be idealist or social constructivist (Dennes 1924; Schaub

1920; Gehlke 1915; LCvy-Bruhl 1922). The divergence between the Ameri-

can and European views did not appear until the 1930s when the positivist

interpretation appeared in the United States. The original interpretation

of Durkheim's work as idealist continued unchallenged in Europe, and

Durkheim's influence on continental anthropology, linguistics, and sociol-

ogy was largely in terms of his alleged idealism.

In the United States, however, Durkheim scholarship took an entirely different trajectory as the pragmatists and other social thinkers who had interpreted Durkheim primarily as an idealist in the teens and twenties fell out of favor (Rawls 1996~). In the 1930s, due particularly to the efforts of Talcott Parsons to rescue Durkheim from the earlier idealist interpreta- tion of his work, a new interpretation of Durkheim as a positivist function- alist became popular. Parsons (1968, p. viii; emphasis in the original) wrote: "In 1924-25 I spent a year as a research student in sociology at the London School of Economics. . . . Durkheim was of course known both in England and America, but discussions were overwhelmingly de- rogatory; he was regarded as the apostle of the 'unsound group mind' theory."

Although Parsons heavily criticizes Durkheim's epistemology and soci- ology of knowledge as idealist, it is not Parsons who introduced the inter- pretation of Durkheim as an idealist. Rather, in his attempt to rescue Durkheim from the already popular idealist interpretation, Parsons cre- ated a positivist interpretation of his work. On this view, the work of the "early" Durkheim was interpreted in positivist functionalist terms and sharply differentiated from his "later" work, which continued to be in- terpreted as an idealist sociology of knowledge. This interpretation of Durkheim's work as comprising an earlier positivist and a later idealist period, commonly referred to as the two-Durkheim hypothesis, effectively banished the later work from serious consideration by scholars in the United States until recently.

From the 1930s until the 1980s, Europeans espoused one "side" of Durkheim's argument while Americans championed the other. Europeans interpreted Durkheim as an idealist, while Americans emphasized what they argued was a positivist functionalist side of Durkheim's argument. Finally, in the 1980s, with the decline of Parsons's influence, allegiance switched from the "early" Durkheim, which Parsons had preferred, to the "later" Durkheim, which Parsons had ridiculed. This "rediscovery" in the United States of the "later" Durkheim led to an increased interest

in Durkheim's sociology of knowledge, or cultural sociology, and in Euro-

pean social theory, which had in important respects been built on an ideal-

ist interpretation of Durkheim's work. However, it did not challenge ei-

ther Parsons's two-Durkheim hypothesis or the original interpretation of

Durkheim as an idealist.

The original interpretation of The Elementary Forms as idealist is at the heart of the problem. It is only because of this first misinterpretation that Parsons invented the two-Durkheim hypothesis in order to avoid the consequences of accepting an idealist interpretation of Durkheim's work. The contemporary rediscovery of a cultural Durkheim merely returns scholarship to the original misunderstanding of his work. Because of this interpretation, Durkheim's work was initially rendered unacceptable to sociologists in the United States and Britain (except to a small group of pragmatist sociologists at the University of Chicago). It was not until Par- sons distinguished the "early" from the "later" work, getting rid of the sociology of knowledge and epistemological arguments and creating a pos- itivist functionalist interpretation, that Durkheim really became popular in the United States.

In the United States, references to the two-Durkheim argument gener- ally cite Parsons's analysis in The Structure of Social Action (1968). Parsons speaks of an Anglo-American prejudice against idealism in this re- gard: "The individualistic bias of the main Anglo-American tradition of positivistic thought has had the effect that any theory which presumed to question it has almost automatically been branded as 'idealistic' and hence condemned as 'metaphysical.' This has happened in striking fash- ion in the case of Durkheim, with the result that he is still predominantly known as the theorist of the 'unsound' and 'metaphysical' group-mind concept" (Parsons 1968, p. 307).

In spite of his insight into the prejudicial nature of the rejection of Durkheim's allegedly idealistic epistemology, however, Parsons does not challenge the idealist interpretation. Rather, he suggests that the last phase of Durkheim's work is only idealist because it remained incomplete due to Durkheim's premature death (Parsons 1968, p. 304). The implica- tion is that Durkheim would never have left the work in this idealist con- dition. The explanation is not convincing, however, as five years elapsed between the publication of The Elementary Forms and Durkheim's death, during which he continued to reiterate his allegedly idealist epistemologi- cal p~sition.'~

Unfortunately, in order to save the rest of Durkheim's oeuvre from the prejudice against idealism, Parsons sacrifices The Elementary Forms.

22 See, e.g., the lectures on pragmatism (Durkheim 1955) and "Dualism of Human Nature and Its Social Conditions" (Durkheim [I9131 1960).

Parsons (1968, p. 445) argued that Durkheim had "in escaping from the toils of positivism overshot the mark and gone clean over into idealism." To protect sociology from this idealist influence, Parsons urged a focus on what he saw as the functional positivist Durkheim of the earlier empirical studies. "At its final stage, Durkheim's sociology stood at this parting of the ways. Both paths represent escapes from positivism, but in terms of the tendency of sociological thought analyzed in the present study [Par- sons's study], especially in the following section, the idealistic phase must be regarded as an aberration, a blind alley" (Parsons 1968, p. 447). Thus, The Elementary Forms did not have much impact on the development of American social theory. Parsons's attempt to separate earlier and later phases of Durkheim's argument heavily influenced subsequent readings of The Elementary Forms. According to LaCapra (1972, p. 5), "In his mon- umental The Structures of Social Action, Talcott Parsons presented Durk- heim's thought as caught in an unresolved tension between early positiv- ism and latter day idealism as it tortuously worked its way toward convergence with other voluntaristic theories of action."

At least since Parsons, references to "two Durkheims" have appeared with regularity in commentary on Durkheim (Parsons 1968; Stone and Farberman 1967; Hughes 1977; Coser 1971; Lukes 1973; LaCapra 1972; Fenton 1984; Alexander 1988). Alexander (1988, p. 10-11) writes that "while some interpreters still do not accede (e.g., Giddens 1977 and Trau- gott 1978)' there is an increasing agreement today that in the later period Durkheim's sociology underwent a decisive shift."23

23 Giddens is almost alone in having completely rejected Parsons's formulation of Durkheim's later work as making a radical departure from the early work. Giddens (1971, pp. 65-82, 105-6, 114) challenged the two-Durkheim hypothesis as early as 1971, carefully documenting the continuity between arguments made by Durkheim in the early review of Schaffle and arguments that appear in the later work. As Gid- dens notes, the similarity between this early review and the conclusion to The Elemen- tary Forms is particularly striking. In addition Giddens (1971, pp. 87-89) argues that the division into an earlier and a later Durkheim on the basis of a changing notion of constraint is also a misconception. Other commentators who claim to reject the two-Durkheim hypothesis in fact make an equally problematic distinction between two forms of constraint. Even though Nisbet is referred to in the literature as a sup- porter of the two-Durkheim hypothesis (e.g., Fenton [1984, p. 321 refers to Nisbet's two-Durkheim theory), he best fits into the group distinguishing two forms of con- straint. Nisbet (1974, p. 59) clearly criticizes the two-Durkheim position: "There are those who regard this work [The Elementary Forms] as marking the passage of Durk- heim's mind to a more idealistic, or at any rate less positivistic, stage of thought. I am obliged to confess that I cannot find this." However, in The Sociological Tradition, Nisbet ([I9661 1993) argues that Durkheim underwent a transition in the writing of The Division of Labor. That transition, according to Nisbet, had to do with his realiza- tion that solidarity and collective representations needed to play a role in modern society also via constraint. It was not a transition to idealism. Nisbet (1993, p. 87) again writes: "For too long students of Durkheim have persisted in placing these

H. Stuart Hughes assumes the two-Durkheim thesis in Consciousness and Society. Hughes gives his version of the transition, which closely fol- lows Parsons's text: "In Durkheim's case it [the confrontation with reli- gion] marked the beginning of a slow transition to a new and deeper un- derstanding of his subject matter. But this transition was completed too late for him to carry out in thoroughgoing fashion the revision of his previ- ous researches that he had planned" (Hughes 1977, p. 284). The influence of Parsons on Hughes is clear in this and in the following passage, which Hughes (1977, p. 285) takes directly from Parsons: "By the same token, Durkheim was led in effect to the assertion that society existed 'only in the minds of individuals. . . . In escaping from the toils of positivism' he 'overshot the mark' and went 'clean over to idealism. . . . An ironical conclusion for one who had first tried to define society in terms of 'things'." Hughes argues that this contradiction was never resolved by Durkheim:

But in a more philosophical sense, his teaching gave no clear lead. To the very end, a central contradiction remained. On the one hand, there was the positivist vocabulary, the striving for a system of "imperative rules" in which critics have seen the last traces of the Comtian metaphysic. And with it there went a hankering after simple and unilateral explanations. At the same time, there was the sweeping spiritual definition of social reality, whose implications were obviously idealist, permissive, and pluralistic. These two aspects of his doctrine Durkheim never brought into any conclu- sive synthesis. (Hughes 1977, p. 286; emphasis added)

Parsons had warned against following the path laid out by the later Durkheim and focused his sociology on what he saw as the earlier func- tional and "positivist" Durkheim. This created an emphasis in the United

works in separate intellectual categories, as though they marked discontinuous phases of his life's labors." Coser also makes the constraint distinction. In his discussion of the earlier and later Durkheim, e.g., Coser (1971, pp. 129, 132, 136) argues that there were several changes: from external constraint to the internalization of constraint as moral; from one opinion about the need for collective conscience to another; and from externalized law to internalized rules. Parsons himself (1973) came to accept the con- straint argument in later life. However, Durkheim objected to this interpretation of his work. He considered the interpretation of his position as concerned with external constraint to be one "from which more than one misunderstanding has resulted" (Durkheim 1915, p. 239n.). He expressed his hope that the analysis of The Elementary Forms would put an end to the problem, which he stated as follows: "it has been assumed that according to our opinion, physical constraint is the essential thing for social life. As a matter of fact, we have never considered it more than the material and apparent expression of an interior and profound fact which is wholly ideal: this is moral authority" (1915, p. 239n). Durkheim goes on to argue that the sociological problem consists of seeking the forms of external constraint which correspond to the different forms of moral authority, which is precisely what he has done in The Elemen- tary Forms.

States on the functional Durkheim with a corresponding neglect of both

his sociology of knowledge and his epistemology.

This is markedly different from what happened in Europe. There, Durkheim's influence was primarily in terms of what Parsons had called the later work-the sociology of knowledge, that is, the logic of concepts as laid out in 1912 in the last chapter of The Elementary Forms, in 1913- 14 in the nineteenth chapter of the lectures on pragmatism, and in the long discussions of classification and totemism that have been misunder- stood as "systems" of ideas. LCvy-Bruhl's initial interpretation of Durk- heim's epistemology as a philosophy of the logic of abstract concepts in- fluenced LCvi-Strauss, who in turn incorporated this view into his structuralism, which then became an important influence on the develop- ment of French structuralism. Poststructuralism also inherits the interpre- tation of epistemology as a system of beliefs.

Durkheim's sociology of knowledge also influenced in some measure the semiotic arguments of Saussure (Jameson 1972; Aarsleff 1982; Alexander 1988). Jameson (1972, p. 27) argues that Saussure's distinction between langue and parole is drawn from Durkheim: "the theoretical advantages of this new model can be measured if we compare it to what seems to have been its source in the sociology of Durkheim." Jameson also cites

W. Doroszewski (1933, pp. 82-91) and Robert Godel (1957, p. 282) on this point. He writes: "not only does the latter's [Durkheim's] insistence on the representational nature of social facts strongly resemble Saussure's notion of signs . . . but the very thrust of Durkheim's thought, in its at- tempt to separate out the personal and individual from the objective and social, is quite consistent with the Saussurean distinction between langue and parole" (Jameson 1972, p. 29).

It is this so-called later period in Durkheim's work, which always had a large impact in Europe, that has become the focus of a contemporary revival of interest in the Unitedstates. Functionalism now being somewhat in disre- pute, the current trend in the United States is toward treating the apparent idealism in Durkheim's later work as the "real" Durkheim, and the later work is now seen as more relevant to contemporary sociology than his early work. Scholars prominent in the social constructivist approach, particularly in studies of science and culture, have over the last several decades begun to recognize Durkheim's sociology of knowledge, as articulated in The Ele- mentary Forms and Primitive Classification, as an important forerunner of their own position. As many sociologists in the postpositivist period have turned to cultural studies and embraced the dilemmas posed by postmod- ernism, poststructuralism, and semiotics, Durkheim's arguments regarding the importance of collective representations have seemed to make the later Durkheim relevant in a very modern sense.

Alexander and Collins (e.g., 1988, 1990) both argue that it is from the

later Durkheim that these schools of thought have developed. Alexander

(1988, pp. 10-11) writes: "It is, of course, the recognition of the crucial

distinctiveness of this later work which has allowed the Durkheimian

roots of contemporary cultural studies to be traced." Alexander (1988, p.

6) continues: "Both as theory and empirical investigation, poststructural-

ism and semiotic investigations more generally can be seen as elaborating

one of the pathways that Durkheim's later sociology opens up. Indeed,

they have demonstrated the importance of his later theory more forcefully

than any discipline in the social sciences more narrowly conceived."

Similarly, Steve Fenton (1984, p. 1) attributes the revival of interest in Durkheim in the 1970s and 1980s to a replacement of the earlier Durkheim by the later in American thinking: "In early American sociology his writ- ings were seen as granting an undue realism to social phenomena, and were thus believed to be antagonistic to American individualism and the voluntaristic tradition." Fenton cites Parsons (1968) and Hinkle and Hin- kle (1954). He continues, "The revival has been marked by a distinct move away from past interpretations of Durkheim as the cornerstone of social conservatism, and, whilst the new work has not shed all social conceptions about Durkheim's weaknesses as a sociologist, it does see these weak- nesses in a new light" (Fenton 1984, p. 1).

While this new appreciation of the allegedly idealist side of Durkheim's argument is in many ways an improvement over the many years of inter- preting Durkheim as a positivist, functionalist, idealist, dualist, rationalist, and realist, it is nevertheless based on the same misunderstanding of Durkheim's work as composed of later idealist and earlier positivist phases. When contemporary theorists argue that Durkheim was a precur- sor of poststructuralism, postmodernism, or pragmatism, they are merely following to its logical conclusion the line of reasoning begun with the idealist interpretation of Durkheim popularized by L6vy-Bruhl, Gehlke, Dennes, L6vi-Strauss, and Parsons that has continued in a relatively un- broken chain to the present. They continue to confuse the sociology of knowledge with the epistemology.

The current fascination with Durkheim's "later" work is centered on his sociology of knowledge and does not recognize his epistemological ar- gument. Consequently, current debates and dilemmas over Durkheim re- volve around weaknesses in treating Durkheim's sociology of knowledge as an adequate epistemological foundation for sociology (Bloor 1982; Alex- ander 1988; Collins 1988; Mestrovic 1993), not his functionalism, empiri- cal studies, or epistemology proper. This view of Durkheim's work "redis- covers" the side of Durkheim that has always been emphasized in Europe. It does not challenge the idealist interpretation of The Elementary Forms. It merely argues that the later idealism of his sociology of knowledge is better than the earlier functionalism. Because of this, the new interest in

Durkheim's sociology of knowledge has ironically reinforced the neglect

of his epistemological argument.

The problem with the whole debate is that there were never two Durk-

heims in the first place.24 Both the European and American interpretations

of Durkheim are merely two sides of the same argument resulting from

the idealist misunderstanding of The Elementary Forms. Both misinter-

pretations result from not seeing the distinction between Durkheim's soci-

oempirical epistemology and his sociology of knowledge. Durkheim's soci-

ology of knowledge without a foundation in his epistemology does

constitute an idealist theory of fictive reality. However, Durkheim clearly

intended for the sociology of knowledge to rest on his own unique episte-

mological basis.

Alexander (1988, p. 6) argues that while the later Durkheim is relevant to cultural studies, the later Durkheim also has to accept the same limita- tions on empirical validity as cultural studies. This is an argument that clearly contradicts Durkheim's own claims concerning the empirical va- lidity of his work. While it is certainly true that Durkheim has had a great and unacknowledged influence on the development of the sociology of knowledge and social constructivist thinking (Bloor 1982; Alexander 1988; Fenton 1984), these arguments are quite separate from his epistemology, which does not have to accept the same limitations and indeterminacies as cultural studies.

For social constructivists, a sociology of knowledge that posits an inde- terminate relationship between thought and reality and that is not capable of empirical validity is the best sort of epistemology that can be hoped for. Social consensus, structure, or shared practices, they argue, lead persons to believe certain things or think in certain ways. Because persons share the same beliefs, they act in ways that reinforce those beliefs. The resulting consensus creates the appearance of a valid relation between thought and reality where there can in fact be none. This makes scientific practice and everyday understanding possible within certain limits but rules out the possibility of empirical validity.

For Durkheim, however, the genesis of the categories of the understand- ing in enacted practice solved the problem of indeterminacy. Although Durkheim applauded the pragmatist attempt to overcome the dualism of thought and reality through a dynamic relation of action, he did not agree with the pragmatist theory of fictive reality. He argued that James's at- tempt was not successful because the argument remained individualistic.

2Wurkheim himself (1982) suggests what he calls a "watershed" in his thinking. He gives the year 1895 for this. But he says that what changed was his ability to deal sociologically with the question of religion, not a change in epistemology, method, or general theoretical direction as has been asserted (Lukes 1982, p. 259).

Into the flux of experience comes a problematic moment, a moment of tension. James argues that at these moments persons are brought to con- sciousness by the need to act. The action, or the need to act, converts the flux of empirical reality into conceptual reality. There is no longer a separation between thought and reality. They have merged in action. However, this method of overcoming dualism results in the argument that whatever action is taken is the truth, as long as it works, because there is no longer a relation to an underlying reality. Truth, logic, and conceptual coherence all give way in pragmatism, according to Durkheim, to the action-defined truth of each particular moment. Durkheim argued that valid knowledge could be explained and reason saved but that this could only be achieved if the focus of epistemological argument were removed from the individual actor and placed instead on shared enacted practices (not shared belief). He pointed out that this was as much a dynamic focus on action as James. The difference is that, for Durkheim, the actions are practices that are inherently collective (motions enacted by an assembled group), whereas for James they were individual.

While Durkheim did articulate a sociology of knowledge along social constructivist lines in the last section of The Elementary Forms, he did not intend this argument to stand as his epistemology. Rather, in the prior 400 plus pages of that work, he carefully outlined an epistemology that demonstrated the empirical validity of six key categories of the under- standing. His sociology of knowledge was only intended to explain the development of ideas beyond these six empirically valid categories and the direct knowledge of social forces that they constitute. He did not intend a social constructivist basis of validity for the categories themselves, which is nevertheless how they have been understood.

Durkheim's epistemology argues for a direct relation between the cate- gories and socioempirical reality. Knowledge and truth are not fictive. Cosmologies are fictive, but the social patterns underlying them are ulti- mately discoverable because the possibility of empirically valid knowl- edge underlies the fictive reality. Thus, Durkheim's sociology of knowl- edge does not rest on an idealist epistemology. The "objectivity" of social facts is just as important an argument for the Durkheim of The Elemen- tary Forms as it was for the Durkheim of The Rules of the Sociological Method. On the other hand, the argument was never positivist, even in the early work, as the two-Durkheim argument assumes, because the "givens" were always socially constructed phenomena resting on a founda- tion of socioempirically valid categories of thought (Durkheim 1915, p. 465). The argument gains in sophistication with the theory of enacted practice articulated in The Elementary Forms but does not change in es- sence, certainly not from a positivist to an idealist position. There only appear to be two Durkheims when the empirical arguments and the sociol- ogy of knowledge are not both seen in the context of his socioempirical epistemology.


A failure to appreciate Durkheim's epistemological argument has impor- tant implications for both sociology and philosophy. For sociology the im- plications are most obvious. Misunderstanding the epistemology leads to serious misunderstandings of Durkheim's entire corpus of work. The thinker whom most sociologists credit with being in some important sense a founder of the discipline has been fundamentally misunderstood. In building on Durkheim's work while neglecting his epistemology, the disci- pline has generally placed itself in an untenable epistemological position, as philosophers have repeatedly pointed out (Winch 1958; Rorty 1979; Turner 1994). What is more important, work that Durkheim would have seen as central to the discipline, for instance, studies of shared enacted practice (interaction) at various levels, has been relegated to the sidelines in the quest to uncover alleged Durkheimian "structures" or "social facts" or, the logic of conceptual or narrative systems.

The two halves of Durkheim's argument have been practiced sepa- rately in different disciplines and subdisciplines and on two different con- tinents, leaving both without epistemological coherence. It matters a great deal to the discipline as a whole whether Durkheim saw social facts as external constraining entities in their own right, as his position has gener- ally been interpreted, or rather as enacted practices that impose moral constraints via "feelings" resulting from the shared enactment of practice. In the second case, the details of the enactment of shared practice are their structure, not the invisible norms and rules posited by the traditional interpretation. Invisible norms and rules can only be revealed via abstract conceptualization, quantitative measurement, and modeling. The details of enacted practice, on the other hand, are open only to a detailed qualita- tive approach. The implications for both the theoretical understanding of "structure" and for preferred methods of research are vast. In order to be consistent with Durkheim's argument, sociology would have to assume an actual order in the enactment of each individual case rather than trying to establish tendencies that must be modeled.25

As important as these implications are, the implications for philosophy are potentially greater but less clear. If it is possible to establish a valid

Z5 See Garfinkel's (1988)discussion of Parsons's plenum for an extended consideration of the practice of modeling a hypothetical order in traditional sociology vs. the assump- tion made by Garfinkel that each individual case displays a witnessable order that is available for research.

epistemological argument on the basis of studies of enacted social practice,

current arguments in epistemology, particularly those of social construc-

tivists, will need to be revised, as Durkheim recognized in 1912. Sociology

and a sociological theory of shared enacted practice would replace the

current vogue for theories of practice based on individual action, systems

of belief, or collective paradigm, and sociology would find itself at the

heart of the epistemological debate, determining the criteria of validity,

instead of at its periphery.

Objections to Durkheim's epistemology will, and should, remain. While a good case for his position can be made, it is ultimately more important to see his work as initiating an important line of argument that, while not explicitly recognized, has always been critical to the sociological enter- prise. It would be a mistake to evaluate the argument on the basis of a completeness that it was incapable of achieving given the limitations of research methods at the time. Because Durkheim's epistemology depends to a large degree on the empirical details of actual shared enacted prac- tices-details that constitute the witnessable enactment of social facts but also details that the research practices of the time shed relatively little light on-it should not be surprising if there is a certain incompleteness in the argument and a degree of inaccuracy with regard to details. Modern technological aids to data collection and the sophistication of contempo- rary field work when contrasted with the lack of attention to such details at the turn of the century guarantee that this will be the case. It is to modern studies of shared enacted practice that we should look for an eval- uation of the potential of Durkheim's view, studies that are generally seen as having no epistemological or general theoretical implications whatso- ever, which, however, turn out to be at the center of the Durkheimian project properly understood.

One problem any argument on a Durkheimian model will have to face is the fact that the analysis depends heavily on the enactment of what Durkheim calls ritual interactions, which (at least in formal institutional terms) play a less prominent role in modern industrial society. Contempo- rary theorists like Garfinkel (1988), Goffman (1959), and Collins (1988, 1990) have argued that informal local orders have replaced formal rituals as the source of order and meaning in modern society. Goffman can be interpreted as arguing for an "interaction order" quite separate from insti- tutional social practices (Rawls 1987). Garfinkel has focused on the achievement of intelligibility in and through local orders that are identical with the practices that enact them. Collins argues for the importance of what he calls "ritual interaction chains" in sustaining social solidarity and selfhood through time. Durkheim's notion of enacted practice may need to be interpreted along similar lines as including mundane enactments of "interaction order" in every day life.26 It will certainly need to be clearly distinguished from institutionalized belief systems.

For Durkheim, the experience of moral force was the most important feature of enacted practice for providing direct experience of the six cate- gories. Mundane enactments of local interaction orders would seem to fulfill this requirement. There are moral obligations at the level of "inter- action order" (as distinguished from institutional levels of orders) wherein everything depends on the mutual commitment to enacted practice (Rawls 1987). This gives interaction a moral dimension and implicates shame, blame, and trustworthiness (Rawls 1990). The original sacred character of formal ritual practices may have played a role in the development of the initial category of classification, dividing the world in two morally in order that the original division would have moral force and thereby socioempirical validity. However, the essential ingredient of enacted prac- tice was always the perception of moral force by participants, which is in principle separable from religious or institutional constraints. The experi- ence of the mutual creation of moral force in and through the enactment of shared practice and the experience of mutual obligation to the enact- ment are both characteristics of "interaction order" practices (Rawls 1987, 1990). Durkheim's own suggestion that professional associations might play a moral role in modern society corresponding to totemic ritual in traditional society runs along similar lines.


Aarsleff, Hans. 1982. From Locke to Saussure. Minneapolis: University of Minnesota Press. Alexander, Jeffrey C. 1982. The Antinomies of Classical Thought: Marx and Durk- heim. Berkeley and Los Angeles: University of California Press. ,ed. 1988. Durkheimian Sociology: Cultural Studies. Cambridge: Cambridge University Press. Alexander, Jeffrey C., and Steven Seidman, eds. 1990. Culture and Society. Cambridge: Cambridge University Press. , Allcock, John, ed. 1983. Introduction to Emile Durkheim's Pragmatism and Sociology. Cambridge: Cambridge University Press. Bloor, David. 1976. Knowledge and Social Imagery. London: Routledge & Kegan Paul. . 1982. "Durkheim and Mauss Revisited: Classification and the Sociology of Knowledge." Journal of the History and Philosophy of Science 13 (4):267-97. . 1983. Wittgenstein: A Social Theory of Knowledge. New York: Columbia University Press.

26 Interaction orders are not random, individually initiated, or "negotiated" orders. They are patterned, recurring, and mutually expected orders, the regularity of which is depended on by participants for achieving intelligibility (Rawls 1987).

Collins, Randall. 1988. "The Durkheimian Tradition in Conflict Sociology." Pp. 107- 28 in Durkheimian Sociology: Cultural Studies, edited by Jeffrey C. Alexander. Cambridge: Cambridge University Press.

. 1990. "Ritual Interaction Chains." Symbolic Interaction 12 (1): 1-32. Coser, Lewis A. 1971. Masters of Sociological Thought: Ideas in Historical and Social Context. New York: Harcourt. Cuvillier, Armand, ed. 1955. Introduction to mile Durkheim's Pragmatism and Soci- ology. Cambridge: Cambridge University Press. Dennes, W. R. 1924. "The Methods and Presuppositions of Group Psychology." University of California Publications in Philosophy 6 (1): 1-182. Doroszewski, W. 1933. "Quelques remarques sur les rapports de la sociologie et de le linguistique: Durkheim et F. de Saussure." Journal de Psychologie 30:83-91. Durkheim, Emile. (1892) 1960. Montesquieu and Rousseau. Ann Arbor: University of Michigan Press. . (1893) 1933. The Division of Labor in Society, translated by George Simpson. New York: Free Press. . (1895) 1982. The Rules of the Sociological Method, translated by W. D. Halls, edited by Steven Lukes. New York: Free Press. . (1897) 1951. Suicide: A Study in Sociology, translated by John A. Spaulding and George Simpson. New York: Free Press. . (1898) 1953. "Individual and Collective Representations," translated by

D. F. Pocock. In Sociology and Philosophy. Glencoe, Ill.: Free Press. . 1912.Les Formes e'le'mentaires de la vie religieuse: Le Systime tote'mique en Australie. Paris: Librairie Felix Alcan. . (1912) 1915. The Elementary Foms of the Religious L$e. New York: Free Press. . (1912) 1995. The Elementary Forms of Religious Life, translated by Karen Fields. New York: Free Press.

. (J913) 1960. "Dualism of Human Nature and Its Social Conditions." Pp. 325- 40 in Emile Durkheim, edited by Kurt H. Wolff. Columbus: Ohio State University Press.

. (1913-14) 1955. Pragmatism and Sociology, translated by J. C. Whitehouse, edited by,John B. Allcock. Cambridge: Cambridge University Press. Durkheim, Emile, and Marcel Mauss. (1901) 1963. Primitive Classification, translated and edited by Rodney Needham. Chicago: University of Chicago Press. Fenton, Steve, with Robert Reiner and Ian Hamnett. 1984. Durkheim and Modern Sociology. Cambridge: Cambridge University Press.

Feyerabend, Paul. 1975. Against Method. London: NLB.

Frazer, John George. 1887. Totemism. Edinburgh. . 1889. "The Origin of Totemism." Fortnightly Review 65:647-65. , 1910. Totemism and Exogamy. London: Macmillan. Fuller, Steve. 1989. Philosophy of Science and Its Discontents. San Francisco: West- view.

Garfinkel, Harold. 1988. "Evidence for Locally Produced, Naturally Accountable Phe- nomena of Order, Logic, Reason, Meaning, Method, etc., in and as of the Essential Haecceity of Immortal, Ordinary Society." Sociological Theory 6 (1): 103-9.

Gehlke, Charles Elmer. 1915. "Emile Durkheim's Contribution to Social Theory." Columbia University Studies in Economics, History, and Public Law 63 (1): 7-187. Giddens, Anthony. 1971. Capitalism and Modern Social Theory. Cambridge: Cam-

bridge University Press. . 1977. "The Individual in the Writing of c mile Durkheim." Pp. 273-91 in Studies in Social and Political Theory. New York: Basic Books. Godel, Robert. 1957. Addendum to Les Sources manuscrites du cours de linguistique ge'ne'rales de F. de Saussure. Geneva: Droz.

Godlove, Terry. 1986. "Epistemology in Durkheim's Elementary Forms of Religious Life." Journal of the History of Philosophy 24 (3): 385-401. Goffman, Erving. 1959. Presentation of Self in Everyday Life. New York: Doubleday Anchor. Goldenweiser, A. A. 1915. "Review of The Elementary Forms." American Anthropolo- gist 17:719-35. Hilbert, Richard A. 1992. The Classical Roots of Ethnomethodology: Durkheim, We- ber, and Garjinkel. Chapel Hill: University of North Carolina Press. Hinkle, R. C., and Hinkle, G. J. 1954. The Development of Modern Sociology. New

York: Random House. Hirst, Paul Q. 1975. Durkheim, Bernard and Epistemology. London: Routledge. Hughes, H. Stuart. (1958) 1977. Consciousness and Society. New York: Random

House. Hume, David. (1739) 1978. A Treatise of Human Nature. Oxford: Oxford University Press. . (1 777) 1975. Enquiries Concerning Human Understanding and Concerning

the Principles of Morals. Oxford: Clarendon Press.

James, William. (1907) 1955. Pragmatism. New York: Meridian Books.

Jameson, Frederic. 1972. The Prison-House of Language. Princeton, N.J.: Princeton University Press. Kant, Immanuel. (1781) 1965. The Critique of Pure Reason. New York: St. Martin's Press.

Kohnke, Klaus Christian. 1991. The Rise of Neo-Kantianism: German Academic Phi- losophy between Idealism and Positivism. Cambridge: Cambridge University Press. Kuhn, Thomas. 1962. The Structure of Scientific Revolutions. Chicago: University of

Chicago Press. LaCapra, Dominick. 1972. d mile Durkheim: Sociologist and Philosopher. Ithaca, N.Y.: Cornell University Press. Latour, Bruno, and Steve Woolgar. 1979. Laboratory L$e: The Social Construction of Scientific Facts. London: Sage. Lehmann, Jennifer. 1990. Deconstructing Durkheim. London: Routledge & Kegan Paul. LCvy-Bruhl, Lucien. (1910) 1966. How Natives Think. New York: Washington Square Press.

. (1922) 1966. The "Soul" of the Primitive. Chicago: Henry Regnery.
LCvi-Strauss, Claude. (1958) 1963. Structural Anthropology. New York: Basic Books.
Locke, John. (1690) 1959. An Essay Concerning Human Understanding. New York:

Dover. Lukes, Steven. 1973. mile Durkheim: His Life and Work: A Historical and Critical Study. Stanford, Calif.: Stanford,University Press. , ed. 1982. Introduction to Emile Durkheim's The Rules of the Sociological Method. New York: Free Press. Mestrovic, Stjepan. 1988.Emile Durkheim and the Reformation of Sociology. Totowa, N.J.: Rowman & Littlefield. . 1993. Durkheim and Poststructuralism. New York: Aldine DeGruyter. Nisbet, Robert. (1966) 1993. The Sociological Tradition. New Brunswick, N.J.: Trans- action. . 1974. The Sociology of mile Durkheim. New York. Oxford University Press. Parsons, Talcott. (1937) 1968. The Structure of Social Action. New York: Free Press.

. 1973. "Durkheim on Religion Revisited: Another Look at The Elementary Forms of the Religious Life." Pp. 156-87 in Beyond the Classics? Essays in the Scientific Study of Religion, edited by Charles Y. Glock and Phillip E. Hammond. New York: Harper & Row.

Quine, Williard Van Orman. (1966) 1977. The Ways of Paradox and Other Essays. Cambridge, Mass.: Harvard University Press. Rawls, Anne Warfield. 1977. "The Ordinary Language Philosophy of John Locke." Manuscript. Wayne State University, Department of Sociology. . 1987. "The Interaction Order Sui Generis: Goffman's Contribution to Social Theory." Sociological Theory 5 (2): 136-49. . 1989. "Language, Self, and Social Order: A Re-Evaluation of Goffman and Sacks." Human Studies 12 (1): 147-72. . 1990. "Emergent Sociability: A Dialectic of Commitment and Order." Symbolic Interaction 13 (1): 63-82. . 1996a. "Durkheim's Epistemology: The Initial Critique, 1915-1924." Sociological Quarterly 38 (I), in press.

. 1996b. "Durkheim's Sentiments and Hume's Passions: Emotion and the Va- lidity of Knowledge." Manuscript. Wayne State University, Department of Soci- ology.

Rorty, Richard. 1979. Philosophy and the Mirvor of Nature. Princeton, N.J.: Princeton University Press. Schaub, Charles. 1920. "A Sociological Theory of Knowledge." Philosophical Review 29 (4): 319-39. Sirianni, Carmen. 1981. "Justice and the Division of Labor: A Reconsideration of Durkheim's Division of Labor in Society." Sociological Review 32:449-70.

Stone, Gregory, and Harvey Farberman. 1967. "On the Edge of Rapprochement: Was Durkheim Moving toward the Perspective of Symbolic Interaction?" Sociological Quarterly, pp. 147-64.

Traugott, Mark. 1978. Introduction to mile Durkheim on Institutional Analysis, edited by Mark Traugott. Chicago: University of Chicago Press. Turner, Stephen. 1994. The Social Theory of Practices: Tradition, Tacit Knowledge, and Presuppositions. Chicago: University of Chicago Press. Tylor, Sir Edward Burnett. (1874) 1973. Primitive Culture: Researches into the Devel- opment of Mythology, Philosophy, Religion, Language, Art, and Custom. London:

J. Murray. . 1899. "Remarks on Totemism with Especial Reference to Some Modern Theo- ries Concerning it." Journal of the Royal Anthropological Institute 28:138-48. Vico, Giambattista. (1744) 1948. The New Science of Giambattista Vico. Ithaca, N.Y.:

Cornell University Press.

Winch, Peter. 1956. The Idea of a Social Science. London: Routledge & Kegan Paul.

Wittgenstein, Ludwig Josef Johan. (1953) 1958. Philosophical Investigations. Oxford:

Basil Blackwell.

  • Recommend Us