Foucault, Structuralism, and the Ends of History



There is no doubt that our century has witnessed a widespread rebellion against historical consciousness, and that in consequence of this rebellion history can no longer lay claim to the central intellectual position to which it aspired in the nineteenth century, when "orthodox" historiography—by which I mean the tradition of professional academic historiography initiated by Ranke—came into being. If, as is suggested by the work of a multitude of poets and philosophers and by the reflections of some historians, we are currently undergoing a crisis of historical consciousness, it is clear that the crisis has been going on for some time. Already, in the works of various late-nineteenth and early-twentieth-century writers, the preoccupations and procedures of professional historians were being roundly condemned,1 while the First World War and the devastating events that followed, by destroying the intellectual respectability of the idea of progress, knocked the foundations out from under the historicist assumptions that had dominated nineteenth-century thought and thus turned the rebellion against historical consciousness into a general revolt.2 Under the aegis of the ideas of progress and of organic or dialectical development— under the aegis, that is, of ideas stressing the continuity between past and present—it was easy to believe that history was a vitally important discipline; under the reign of discontinuity the inclination is to turn toward other disciplines more relevant in their subject matter or more creative in their practice. The characteristic response of twentieth-century historians to the threat of history's potential irrelevance has been to attempt to preserve its vitality by extending


1 For examples, see Hayden White, The Burden of History," History and Theory 5 (1966): 111-34.

2 On nineteenth-century historicism and its decline, see Maurice Mandelbaum, History, Man, and Reason: A Study in Nineteenth-Century Thought (Baltimore, 1971), pp. 41-138 and 369-70. I use the word historicism,, in MandelbaunVs sense, to denote the widespread nineteenth-century belief that a thing can be properly understood only if one views it in terms of the place that it occupies within some larger process of development.



the range of its subject matter. Meinecke, for example, believed that historical scholarship could regain its former intellectual impact by reaching out toward the history of ideas, while the founders of the Annales school, Lucien Febvre and Marc Bloch, focused their hopes on the reconstruction of the broad socioeconomic processes of the past and on a methodological rapprochement with social scientists interested in the study of the same processes within contemporary society. But if these extensions in the subject matter of history meant the modification or abandonment of the mainly political focus of earlier ^orthodox" historians, they did not represent a fundamental departure from the general assumptions on which orthodox historiography had been founded.

The same cannot be said, however, of the brilliant, speculative, and in some ways deeply disturbing writings of the contemporary French historian Michel Foucault (b. 1926), whose historiographical aims are very different from those animating orthodox historians. Such historians as Meinecke, Febvre, and Bloch wrote with the intention of revitalizing a historiographical tradition that they saw as basically sound, even though, in their view, narrowness and lack of imagination had prevented historians from realizing the full potential of their craft. Foucault's aim—or at any rate one of his aims—is the demolition of that tradition. For Foucault, who since 1970 has been "'professor of history and systems of thought" at the College de France, is one representative of a radically antihistorical trend in recent thought—a trend that, under the inadequate labels of k'structuralism' ' and ^poststructuralism," has been a highly important part of the French intellectual scene over the past fifteen or twenty years. Among other things, many of the writers who are part of this trend (it is hardly coherent enough to be called a movement) have vehemently attacked historical modes of apprehension and understanding. Most English-speaking historians will have at least a passing acquaintance with structuralism; that is to say, they will at least know that something called structuralism exists. They may also know that the thinker who is in large measure its fountainhead, the linguist Ferdinand de Saussure, advocated, in opposition to the largely diachronic, historical linguistics that had hitherto predominated, a synchronic linguistics that would concern itself not with the evolution of language over time but rather with the structure of language at a given point in time.3 They may likewise know that an


3 Ferdinand de Saussure, Cours de linguistique generate (first published, posthumously, 1916), translated as Course in General Linguistics, ed. Charles Bally and Albert Sechehaye, with the collaboration of Albert Riedlinger, trans. Wade Baskin (London, 1974).



important figure in the structuralism of the 1960s, the ethnologist Claude Levi-Strauss, contrasted the magical and totemic thought of savages, which in its refusal to believe that anything really changes is radically antihistorical, with the thought of modern man, and that he stressed the richness and power of the former while suggesting that the benefits of modern or "hot" societies—societies that are historical, that are always on the move—are hardly worth the price.4 And finally, if they are truly up-to-date they may be aware of the poststructuralist literary critic and philosopher Jacques Derrida, whose best-known work, De la Grammatologie, reads like an absurdist parody of everything that has ever gone under the name of intellectual history.5 It is via Foucault, however, that orthodox historians can best come to grips with the antihistorical trend in recent thought—with those thinkers who, far from regretting a crisis of historical consciousness, welcome and promote it. For whereas Saussure is a linguist, Levi-Strauss an ethnologist, and Derrida a critic and philosopher, Foucault claims emphatically to be a historian,6 and his enterprise closely approximates, at least in its outward form, the enterprise of orthodox historians.

Foucault's first historical work, published in 1961, was his His-toire de la folie, which dealt with the history of madness from the late Middle Ages to the present day. This was followed by Nais-sance de la clinique, which discussed the development of medicine in the late eighteenth and early nineteenth centuries, and by Les Most et les с hoses, by far Foucault's most substantial and important work of the 1960s, which presented an account of the history of Western thought from the Renaissance onward. These three works were summed up, defended, and modified in a methodological treatise, L 'Archeologie du savoir, a work that was little read but that has nevertheless gained something of a cult following. Then came a six-year period during which Foucault published no full-length books at all, a period of relative silence broken in 1975 with the appearance of Surveiller et punir,7 a study of modes of punishment and disci-


4 See esp. Claude Levi-Strauss, Tristes tropiques, translated under the same title by John and Doreen Weightman (London, 1973), and La Pensee sauvage (1962), translated anonymously as The Savage Mind (London, 1966). In the final chapter of The Savage Mind, "History and Dialectic;' Levi-Strauss attacks the idea of historical process and argues that history is necessarily discontinuous.

5 Jacques Derrida, Of Grammatology, trans. Gayatri Spivak (Baltimore, 1976), first published in 1967 as De la Grammatologie.

6 See, most recently, "Foucault: Non au sexe гоГ' (Foucault interviewed by Bernard-Henri Levy), Nouvel Observateur (mars 12-21, 1977), pp. 92 + , translated as "Power and Sex: An Interview with Michel Foucault," by David J. Parent Telos no. 32 (Summer 1977), pp. 152-61.

7 Michel Foucault, Folie et deraison: Histoire de la folie a Tdge classique (Paris,



pline from the eighteenth century to the present. This was quickly followed by the publication in 1976 of the first volume of an ambitious Histoire de la sexualite which, when completed, is to include six studies: La Volonte de savoir, which serves as an introduction to the whole work; La Chair et le corps', La Croisade des enfants\ La Femme, la mere et Vhysterique', Les Pervers; and Populations et races. And beyond this we can expect to see yet another work, entitled Pouvoir de la verite,8 which will deal, presumably, with the theme of the relationship between knowledge and power that has become pervasive in Foucault's more recent work. In short, Foucault is a highly productive writer, whose already substantial corpus promises to be supplemented by even more works in the future. Furthermore, the writings that Foucault has produced so far have been, almost without exception, both original and compelling in nature. It is reasonable to expect that Foucault's prolific output, combined with the brilliance and topicality of what he produces, will make him a figure of considerable influence; indeed, some journalists are calling him "the new Заг^е.''9 Certainly, if he can bring his planned works to fruition he will have produced an imposing corpus.

Yet orthodox historians—at least in the English-speaking world— have either ignored Foucault, regarding his work as totally irrelevant to their own, or have summarily dismissed him. For when the orthodox historian attempts to read Foucault he finds himself confronted by serious difficulties—difficulties of which historically illiterate readers will be entirely oblivious. Quite naturally, the historian turns to Foucault's works, as he turns to more orthodox works of history, in the hope that they will contribute to this own understanding of the historical past. In practice, this means that he expects them to add themselves, without creating excessive difficulties, without behaving in a noisy or unruly fashion, to the mass of historical knowledge that he already possesses. But typically it will


1961), translated—in an abridged version—as Madness and Civilization: A History of Insanity in the Age of Reason by Richard Howard (New York, 1967); Naissance de la clinique: Une archeologie du regard medical (Paris, 1963), translated as The Birth of the Clinic: An Archaeology of Medical Perception by A. M. Sheridan Smith (London, 1973); Les Mots et les choses: Une archeologie des sciences humaines (Paris, 1966), translated anonymously as The Order of Things: An Archaeology of the Human Sciences (New York, 1970); VArcheologie du savoir (Paris, 1969), translated as The Archaeology of Knowledge by A. M. Sheridan Smith (New York, 1976); Surveiller et punir: Naissance de la prison (Paris, 1975) translated as Discipline and Punish: The Birth of the Prison, by Alan Sheridan (New York, 1977).

8 Promised in Michel Foucault, La Volonte de savoir (Paris, 1976), p. 79n.

9 Ferdinando Scianna, "La nuova sessualita: Rivoluzionaria analisi del nuovo Sartre," VEuropeo (febbraio 18, 1977), pp. 49-53.



take only a few pages of reading to convince him that something is amiss and to generate in him a feeling of genuine puzzlement. For he will encounter what from his point of view appear to be three distinct sorts of statements. He will find, in the first place, statements that he simply cannot understand. Second, he will find statements that strike him as (at worst) plausible descriptions of, or (at best) brilliant insights into, the historical field in question. And finally, he will find statements that he is convinced can only be mistaken. At this point, puzzlement turns to irritation, with the historian-reader most likely concluding that to read Foucault is a waste of time. Should the historian-reader persist—and he almost certainly will not—he may come out of his reading with mixed views about Foucault, as Keith Michael Baker does in asserting, of Les Mots et les choses, that "Foucault's analysis of the underlying epistemological procedures of Enlightenment thought is as brilliantly suggestive as his characterization of the nature of this episteme is confusing,'40 or as Roger Hahn does, in pronouncing Foucault's Naissance de la clinique to be a "terrible book": 'Terribly annoying because of the impressionistic style, the faulty construction, the willful effort to create new concepts by manipulating traditional language, and the forced desire continually to transcend the banal"; yet at the same time "terribly perceptive and suggestive, in ways that are hard to express.'41 But it is perhaps more likely that he will come out of his reading uniformly hostile and ready to second the entirely negative judgments of Foucault that one finds expressed in, for example, George Huppert's attack on Foucault's reading of the Renaissance,12 or G. S. Rousseau's attack oh Foucault's reading of the Enlightenment.13 Indeed, only one English-speaking historian, Hayden White, has been clearly sympathetic toward Foucault's work. In a long article published in History and Theory in 1973, White attempted a wide-ranging interpretation of the rhetorical or "tropological" substratum of Les Mots et les choses,14 while his brief review of Surveiller et punir, published in the American Histor-


10 Keith Michael Baker, Condorcet: From Natural Philosophy to Social Mathematics (Chicago, 1975), pp. vii-viii.

11 Roger Hahn, Review of The Birth of the Clinic, American Journal of Sociology 80 (May 1975): 1503-4.

12 George Huppert, "Divinatio et Eruditio: Thoughts on Foucault,'' History and Theory 13 (1974): 191-207.

13 G. S. Rousseau, "Whose Enlightenment? Not Man's: The Case of Michel Foucault," Eighteenth-Century Studies 6, no. 2 (Fall 1972): 238-55.

14 Hayden White, "Foucault Decoded: Notes from Underground," History and Theory 12 (1975): 23-54; also available in White, Tropics of Discourse: Essays in Cultural Criticism (Baltimore, 1978), pp. 230-60.



ical Review in 1977, marked the first appearance of any of Foucault's books in the review columns of an English-language historical journal.15

The difficulties that orthodox historians have had in coming to terms with Foucault suggest that something more far reaching is involved here than a disagreement, between Foucault and his historian critics, over individual points of historical interpretation. Rather, these difficulties suggest that Foucault's enterprise is fundamentally different from the enterprise of orthodox historians, and that simply to condemn Foucault's portrayal of the past as mistaken or simply to praise that portrayal as insightful represents a mistaken attempt to assimilate Foucault to the structure of orthodox historiography. They suggest, in short, that the historiographical criticism of Foucault must concern itself with Foucault's general historiographical presuppositions before it turns to deal with his historical analyses. As White points out, to judge Foucault according to conventional historiographical standards is to commit a "category mistake,,,J6 for Foucault is not engaged in conventional historiography; on the contrary, he "writes 'history' in order to destroy it, as a discipline, as a mode of consciousness, and as a mode of (social) existence."17 Hearing this, the orthodox historian might be tempted to reject any meeting with Foucault as pointless, given the manifest lack of common ground between Foucault and his orthodox counterparts. But such a judgment would be just as misguided as the attempt to interpret Foucault as if he were simply another orthodox historian. For besides informing the orthodox historian about a writer of great potential influence representing an important strand in contemporary thought, a reflection on Foucault's enterprise will also serve to highlight the character of orthodox historiography. It is with these thoughts in mind that I propose in this paper to investigate and criticize the foundations of Foucault's historiography.

While it would be a mistake to attribute to orthodox historiography a paradigmatic unity of the sort that T. S. Kuhn has pointed to in the natural sciences, it is nevertheless true that orthodox historians do adhere to what J. H. Hexter has called the "reality


15 Hayden White, Review of Surveiller et punir, American Historical Review 82 (1977): 605-6.

16 Ibid.

17 White, "Foucault Decoded," p. 26. Edith Kurzweil, "Michel Foucault: Ending the Era of Man," Theory and Society: Renewal and Critique in Social Theory 4, no. 3 (Fall 1977); 395-420: "In America the historian is a relic, but France's Michel Foucault, a historian of scientific thought, has become a prophet. Of course he is not a conventional historian (p. 395).



-rule," the rule that historians cannot tell just any story about the past but must rather tell "the best and most likely story that can be sustained by the relevant extrinsic evidence." The historian seeks, in short, to render "the best account he can of the past as it really was."18 To be sure, the historian's belief in the objective and realistic nature of his enterprise is tempered by his recognition that all accounts of the past contain an irreducible and inexpungeable element of interpolation—that is, an irreducible and inexpungeable element of subjectivity; but he holds that it is the explanatory element in the historical account, not the interpretative element, that is basic, admitting a legitimate role for interpretation only when it becomes necessary to fill in the gaps in a fragmentary historical record or (conversely) to exclude certain facts, or categories of facts, where the historical record is too copious.19 In short, orthodox historians adhere to a "discovery" view of the past, holding that the past is there, a field of real entities and forces waiting for the historian to find; and they reject the opposing "construction" view of the past, which holds that, far from discovering and reporting the past, historians must be regarded as constructing or creating it.20 In accepting an irreducible interpretative element in history, orthodox historians recognize that the historical account is in part an invention of the historian, but they see it as an invention that, solidly grounded in the facts of history, rightly aspires to portray the past "as it actually was." The present concerns and commitments of the historian will enter into the historical account as part of its necessarily interpretative element. Such concerns may, for example, be especially important in suggesting problems for historical investigation, as one can see time and time again in the history of modern historiography. But the origin of a particular historical investigation is separable, in the orthodox view, from its scholarly validity, the historian having both the capacity and the duty to distance himself, in his work, from present concerns. Perhaps the best expression of this view is to be found in Hexters essay, "The Historian and His Day," in which he argues that the historian's commitment to the accepted procedures of historical study, combined with an immersion in the documents, enables him to attain a contact with the past


18 J. H. Hexter, "The Rhetoric of History," History and Theory 6 (1967): 3-13; quotes from pp. 5 and 11.

19 As Hay den White points out, in "Interpretation in History," New Literary History 4, no. 2 (Winter 1973): 281-314; also available in Tropics of Discourse, pp. 51-80.

20 On the distinction between these two views, see Jack W. Meiland, Scepticism and Historical Knowledge (New York, 1965), pp. 3-4.



that, in its immediacy, particularity, and vividness, rivals his contact with the present.21

The orthodox historian is strongly committed, furthermore, to the view that there is a clear distinction between getting things right and getting things wrong. In his elementary concern with getting things right, the orthodox historian signals his adherence to a view that has dominated the historical profession since its birth in the nineteenth century—namely, the view that history is at bottom a science, capable of realistically apprehending the world and of discovering a truth that is more than relative. Admittedly, historians are today less confident about the scientific status of history than they were at the end of the nineteenth century, but there still remains a basic commitment to the ideal of scientific history. Indeed, one of the most striking features of recent historiography has been its increasing scientization as historians have come more and more to draw on the concepts and methods of the social sciences.22 It is true that some historians, most notably Hayden White, have argued that history is founded on a poetic apprehension of the world that is entirely prescientific in nature.23 But this is very much a minority position which, in its assertion that the historical fact is really a poetic factum, is in contradiction to the ingrained realism of the vast majority of historians. For the orthodox historian, the evidence that he has so laboriously discovered and assessed has a reality of its own reflecting the reality of the past itself, and he sees his task as the construction of a historical account that will explain and interpret this actual past.

Foucault does not conform to the rough consensus on the nature of historical investigation that I have just sketched out. On the contrary, he stands in radical opposition to it. But the nature and bearing of this opposition only become clear when one perceives the connections between Foucault's view of history and his reading of Nietzsche, who is undoubtedly the most severe critic that the enterprise of orthodox historiography has ever encountered. I do not mean to suggest that Foucault is nothing more than a follower of Nietzsche; I do not wish to reduce the Foucaultian enterprise to the earlier Nietzschean enterprise. Nevertheless, it is clear that


21 J. H. Hexter, "The Historian and His Day," in Reappraisals in History (London, 1961), pp. 1-13.

22 For insight into recent developments in historiography, see esp. Felix Gilbert and Stephen R. Graubard (eds.), Historical Studies Today (New York, 1972), and Georg G. Iggers, New Directions in European Historiography (Middletown, Conn., 1975).

23 Hayden White, Metahistory: The Historical Imagination in Nineteenth-Century Europe (Baltimore, 1973).



Nietzsche has been the single most important influence on Foucault's work, and Foucault himself makes no secret of the debt, telling us in Les Mots et les choses, for example, that Nietzsche "marks the threshold beyond which contemporary philosophy can begin thinking again; and he will no doubt continue for a long while to dominate its advance.,,24 Yet Foucault's discovery of Nietzsche was slow and halting. Though the mark of his reading of Nietzsche is already present in his first book, Maladie mentale et psychologies published in 1954, it was to take almost twenty years more for him to arrive at a conception of historical investigation that was genuinely and thoroughly Nietzschean. Foucault began his historical work with the intention of carrying out what he referred to as an archaeology' using this word in the allegedly Kantian sense of k4he history of that which renders necessary a certain form of thought."25 Foucault's archaeology was, in essence, a hybrid and unstable combination of conventionally historiographical concerns with certain structuralist themes and preoccupations. The history of Foucault's career as a historian has been the history of his movement from a k'Kantian'' archaeology to a Nietzschean genealogy. But the passage from archaeology to genealogy was delayed, I shall argue, by his traversal of structuralism—a traversal that obscured, both for him and for us, the true nature of his historical vocation. The conflict, in the interior of the Foucaultian text, between structuralism and antistructuralism, between Apollo and Dionysos, even—if one will—between Plato and Nietzsche, has from all points of view been the most interesting and most revealing theme in Foucault's work to date. It is, furthermore, a theme that we must grasp if we are to understand the changing presuppositions that have underlain Foucault's various historical writings.

Foucault himself has vehemently denied ever having been a structuralist. For example, in the foreword to the English edition of Les Mots et les choses Foucault tells us that kkin France, certain halfwitted 'commentators' persist in labelling me a 'structuralist/ I have been unable to get it into their tiny minds that I have used none of the methods, concepts, or key terms that characterize structural analysis."26 It is entirely true that Foucault was never a structuralist


24 Foucault, The Order of Things, p. 342.

25 Michel Foucault, "Monstrosities in Criticism," Diacritics: A Review of Contemporary Criticism 1 (Fall 1971): 57-60; quote from p. 60.

26 The Order of Things, p. xiv; see also his "Monstrosities in Criticism/' p. 58, and the dialogue in The Archaeology of Knowledge, pp. 199-203.



in any narrow construction of the term, but in a broader sense he certainly was a structuralist—even though, as we shall see, there was always a fundamentally antistructuralist element in his thought as well. Foucault's vehement denial of structuralism comes down ultimately to questions of terminology. For "structuralism" is a word with so many meanings that it can hardly be said to have any meaning at all,27 and its broad connotative penumbra has often been permitted to substitute for the rigors of definition. Certainly, the vaguely perceived contents of this shadowy world of meaning might tend to suggest that the word ^structuralism" fits, without further qualification, the Foucaultian enterprise. Thus, it is well known that structuralism is intimately tied up with language, and when one looks at the Foucaultian text one will see that Foucault, too, is deeply concerned with language. Indeed, Foucault's reflections on language form the underlying theme of Les Mots et les choses. It is well known that structuralism, in its search for a stable object of investigation, concentrates on language, or langue, rather than on the human speaker—a concentration that seems to be paralleled by Foucault's attack on subjectivism and anthropologism. It is well known that structuralism is synchronic rather than diachronic in orientation, an orientation that is apparently paralleled by Foucault's preference for discontinuity in history and by his refusal to explain the transitions or "mutations" leading from one episteme to the next. It is well known that structuralist analyses are articulated in terms of "binary opposition"; and when one looks at Foucault one finds—or seems to find—a massive and all-embracing opposition between "the Same," dealt with in Les Mots et les choses, and "the Other," dealt with in Histoire de lafolie, Naissance de la clinique, and Surveiller et punir. And finally, it is well known that structuralism focuses on the concept of the sign, and when one looks at the Foucaultian text one finds a pervasive interest in signs and their permutations. Witness, for example, the chapter on "Signs and Cases" in Naissance de la clinique;28 witness also the close relationship between significatory change and epistemic change in Les Mots et les choses.29

But these parallels, which in the wake of the publication of Les Mots et les choses became journalistic commonplaces, betray a


27 As Francois Wahl puts it: "Let's say it frankly: when people ask us about structuralism, most of the time we don't know what they want to talk with us about'' (Oswald Ducrot et al., Qu'est-ce que le structuralisme? [Paris, 1968], p. 9).

28The Birth of the Clinic, pp. 88-106.

29 See, most important, The Order of Things, pp. 42-43. Note that, in a rare blunder, the translator has rendered signifiant as "significant" rather than as "signifier."



gross failure to attend to the subtleties of the Foucaultian text, and it is little wonder that Foucault, confronted in the late 1960s by a concerted attempt to confine him in a box marked "structuralism," should have reacted with angry repudiations of a term that he himself had used to characterize his work. In arguing that there is a structuralist element in the Foucaultian text I have, however, something more specific in mind than the almost meaningless parallels mentioned above; when I say that Foucault was a structuralist in a broad construction of the term, I do not mean to say that he was a structuralist in a vague and intellectually sloppy construction of the term. But what I do mean will become clear only through a careful examination of the meanings of structuralism. There are, of course, a variety of ways of "slicing" almost any synthetic concept, for articulate general concepts tend to be articulated at more than one point. But for our present purposes—and without denying the possibility of other analyses—I wish to distinguish between a narrower "structuralism of the sign" and a broader "structuralism of structure," each of which may in turn be construed in both a strict and a loose sense. The structuralism of the sign has its conceptual origins in Saussure's Cours de linguistique generate and more specifically in the Saussurean definition of the sign as the union of signifiant and signifie. But the import of Saussurean structuralism can be variously interpreted. Some analysts of the concept of structuralism adhere to a relatively strict, "linguistic" definition of the term, restricting it to intellectual enterprises conforming rather closely to the outlines of Saussurean and post-Saussurean linguistics. Other analysts adhere to a looser, "semiological" definition of the term, linking it not to linguistics but to Saussure's proposal for a science of semiology that would concern itself with the study of "the life of signs within social life."30

Perhaps the most rigorous attempt to see structuralism in a strict, linguistic sense is to be found in Philip Pettit's book, The Concept of Structuralism: A Critical Analysis.31 Structuralism, Pettit asserts, and here all analysts of the structuralism of the sign would agree, involves an attempt to extend certain Saussurean and post-Saussurean analytical procedures beyond linguistics, applying them to such areas as literary criticism, art criticism, social psychology, social anthropology, and the analysis of "cutomary arts" like fashion and cuisine. But Pettit interprets the structuralist model of language very narrowly, arguing that anyone who is extending this


30 de Saussure (n. 3 above), Introduction, chap. 3, sec. 3, p. 33.

31 Philip Pettit, The Concept of Structuralism: A Critical Analysis (Berkeley, 1975).



model beyond linguistics may think of doing so in terms of three, and only three, analogies: structural phonology (as in Jakobson), generative syntax (as in Chomsky), and differential semantics (as proposed by Pettit himself).32 Each of these analogies, Pettit argues, requires that the nonlinguistic object being analyzed contain some element that corresponds to the sentence in language. But since none of the nonlinguistic objects upon which structural analysis has been attempted in fact contains such an element, Pettit concludes that the structuralist model, though it may have some heuristic value in fields outside linguistics, does not in any proper sense "nY' any of those fields.

Though Pettit mentions Foucault only in passing,33 preferring to concentrate his attentions on Levi-Strauss, there is never any doubt that Foucault does not conform to the kind of strict Saussurean model that Pettit articulates. And indeed, the vaguely Saussurean parallels, mentioned above, between the structuralism of Saussure and the Foucaultian text turn out on further examination to be almost entirely specious. True, Foucault's reflection on language is an extremely important part of his work, but this reflection owes far more to Mallarme (mediated through Blanchot and other French literary critics) than it does to Saussure.34 It is true that Foucault attacks subjectivism and anthropologism, but this is bound up with his strong reaction against the idealism of Husserlian phenomenology and of Sartrean existentialism and indicates a debt to Nietzsche, not to Saussure.35 It is true that Foucault has tended to emphasize the discontinuous in history, and that this has sometimes made it appear—particularly in Les Mots et les choses—that he is engaged in something that resembles, in its temporal orientation, Saussure's synchronic linguistics. But once again the appearance is entirely deceptive, for Foucault's emphasis on discontinuity is part of his attack on subjectivism, "continuous history1' being, in his view,


32 Ibid. p. 29.

33 Ibid., pp. 68-69.

34 For the Mallarmean theme in Foucault, see The Order of Things, esp. pp. 43-44, 81, 305-6, 382-84. On the importance of Blanchot for Foucault, see Raymond Bellour, "Deuxieme entretien avec Michel Foucault: Sur les facons d'ecrire This-toire," Les Lettres frangaises, no. 1187 (juin 15-21, 1967), pp. 6-9: "Cest Blanchot qui a rendu possible tout discours sur la Iittёrature,, (p. 8). Foucault also tells us here that "I differ from those who are called structuralists in that I am not greatly interested in the forma! possibilities presented by a system such as language. Personally, I am haunted rather by the existence of discourses, by the fact that utterances have taken place. . . ."

35 Thus, Foucault sees Nietzsche as having been the first to awaken us from "the anthropological sleep," the first to tear us free from "the anthropological fiekT {The Order of Things, pp. 340-43; see also pp. 306-7 and 322).



"the indispensable correlative of the founding function of the subject."36 As for Foucault's alleged interest in binary opposition, the fact is that though at a certain moment in his career Foucault was attracted by the idea of constructing "a whole series of binary divisions which in their own way would have re-minted the great division 'reason-unreason' that I had tried to reconstitute with regard to madness,"37 this idea was never really worked out, remaining an entirely subsidiary theme in his oeuvre as a whole.

We are left, then, with our final parallel, the fact that the concept of the sign, which is centrally important for Saussure and on whose basis he wanted to construct semiology, also functions as an important concept within the Foucaultian text. At this point we move from the strict, linguistic reading of the structuralism of the sign to the looser semiological reading. For, in fact, many of those who have called themselves structuralists are far more interested in the science whose outlines Saussure did not articulate than in the science whose outlines he actually did articulate. Indeed, in recognition of this, Pettit admits the rough interchangeability of the terms "semiology'' and "structuralism,"38 even though he goes on to discuss structuralism in terms of a strictly linguistic model. But other analysts, in their attempts to define the limits of structuralism, take Saussure's semiological intentions more seriously. One such commentator is Francois Wahl, whose essay, "La Philosophic entre l'avant et l'apres du structuralisme" (included in the volume Qu est-ce que le structuralisme?), is an important attempt to come to grips with the nature of the structuralist phenomenon. Like Pettit, Wahl identifies structuralism and semiology, asserting that "under the name of structuralism are grouped the sciences of the sign, of systems of signs/'39 But unlike Pettit, he does not go on to assert that the practice of a science of signs requires a strict conformity between the structure of the object being analyzed and the structure of language. On the contrary, Wahl is willing to allow the possibility of structural analysis wherever the object being analyzed passes through a structuring linguistic grid. For example, Wahl tells us that "the most diverse facts of anthropology" can be the object of structuralist analysis, "but only insofar as they pass through the facts of language—that they are caught within the institution of a


36 The Archaeology of Knowledge, p. 12.

37 Lucette Finas, "Entretien avec Michel Foucault: lLes rapports de pouvoir passent a Tinterieur des corps/ " Quinzaine litteraire 247 (Janvier 1-15, 1977): 4-5; quote from p. 5.

38 Pettit, p. 33.

39 Ducrot et al., p. 10.



system of the type Signifiantlsignifie and lend themselves to a communicative network—and that they receive from this their structure. "40 In short, structuralism, for Wahl, deals with structures; but it deals with structures only insofar as they have acquired their structure from their passage through a system of signs.

For Wahl, then, the sign is the absolutely critical defining element in structuralism: where the sign is, there also is structuralism, regardless of the absence of such linguistic elements as the sentence. Thus, in attempting to distinguish what is "not yet" structuralism from what is "no longer" structuralism, Wahl tells us that "wherever the sign is not yet conceived as being in an absolutely fundamental position, thought has not yet taken note of structuralism. Wherever the primacy of the sign is disputed, wherever the sign is destroyed or deconstructed, thought is no longer in the orbit of structuralism."41 On this reading of structuralism, where does Foucault stand? Since Wahl's account of Foucault's relationship to structuralism is based on a reading of Les Mots et les choses, a brief summary of the general thesis of that work is here in order. The book is set within the context, and between the limits, of an event that is stunning in its gratuitousness, namely, the presence, retreat, and return of language.42 The central protagonist of Les Mots et les choses is "language"—by which Foucault means, not language in the ordinary sense of the word, but rather language in a very Mallarmean sense: that is, language insofar as it has an autonomous and self-referring existence, freed from subjection to anything outside language. The mirror image of language, which appears when language disappears and disappears when language appears, is discourse. Again, Foucault employs the word "discourse" in a special sense, derived from the epistemological and linguistic writings of the Ideologues, Condillac, and ultimately Locke. Discourse, for Foucault, is language from which all self-reference, all inner play, all metaphorical distortion are eliminated. The sole function of discourse is to serve as a transparent representation of things and ideas standing outside it.43 Hence, language and discourse are totally


40 Ibid., pp. 10-11.

41 Ibid., p. 304.

42 For the outlines of this theme, see The Order of Things, pp. 42-44, 303-7, 382-87.

43 For references to discourse, see ibid., pp. 81, 236, 304, 311, 385-86. It should be noted that in L'Archeologie du savoir Foucault uses the word "discourse" much more broadly, to include—it would seem—virtually every systematic use of language. See esp. The Archaeology of Knowledge, p. 80: "... instead of gradually reducing the rather fluctuating meaning of the word 'discourse,' I believe that I have in fact added to its meanings: treating it sometimes as the general domain of all statements,



antithetical: in language, the "direction of meaning" is entirely inward; in discourse, it is entirely outward.44 Where "language" disappears, as Foucault argues it did at the beginning of the seventeenth century, all that remains of language is "its function as representation: its nature and its virtues as discourse."*5 Conversely, when language returns—and Foucault asserts that it returned at the end of the eighteenth century, though it has not yet regained its unity—then discourse disappears.46

Foucault's account of the disappearance and return of language is closely connected with an account of signs and signification. This is especially true of his account of the disappearance of language, which he relates to a fundamental change, at the beginning of the seventeenth century, in the structure of the sign. From the Stoics to the Renaissance, the system of signs in the Western world was, Foucault asserts, a "ternary" one, signifiant and signifie being linked together by a "conjuncture," that is, by a relationship of resemblance of one sort or another. But at the beginning of the seventeenth century the system of signs became "binary," with a purely arbitrary relationship between signifiant and signifie. It was this change, Foucault asserts, that signaled the disappearance of language from the world and its replacement by a supposedly transparent discourse.47 Unfortunately, quite apart from the question of the historical accuracy of what Foucault here argues—and as I said at the beginning of this paper I am not concerned here with whether Foucault is right or wrong in what he says about the past—his account of signs and signification remains unclear, even after one has gone to the considerable effort of learning his somewhat idiosyncratic terminology and of grasping the architectonics of his work. The locus of the problem is to be found in Foucault's failure to explain clearly his distinction between representation and significa-


sometimes as an individualizable group of statements, and sometimes as a regulated practice that accounts for a certain number of statements; and have I not allowed this same word 'discourse,' which should have served as a boundary around the term 'statement/ to vary as I shifted my analysis or its point of application, as the statement itself faded from view?" See also ibid., pp. 107, 117, 169.

44 I borrow the expression "direction of meaning" from Northrop Frye, Anatomy of Criticism: Four Essays (Princeton, N.J., 1957), pp. 73-74.

45 The Order of Things, p. 81.

46 Ibid., pp. 303-4, 385-86. For more on this theme, see Michel Foucault, Language, Counter-Memory, Practice: Selected Essays and Interviews, ed. Donald F. Bouchard, trans. Donald F. Bouchard and Sherry Simon (Ithaca, N.Y., 1977), esp. the four essays—"A Preface to Transgression," "Language to Infinity," "The Father's 'No,' " and "Fantasia of the Library"—that the editor has classified under the general rubric of "Language and the Birth of 'Literature.' "

47 The Order of Things, pp. 42-43, 27-30.



tion. Representation, he argues, is characteristic of the Classical episteme of the seventeenth and eighteenth centuries; signification is characteristic of the modern episteme that began in the late eighteenth century and that is now, he suggests, on the verge of its demise. Yet Foucault never makes it fully clear what the distinction between representation and signification is; nor does he make it fully clear what the implications of this distinction are for the concept of the sign, which remains binary in structure throughout both the classical and the modern epistemes. The drift of Foucault's account suggests that he sees the two concepts as variants of each other, since both exist under the aegis of the binary sign and in an economy in which language either does not exist (representation) or exists in a fragmentary state only (signification).48 They are, however, incompatible variants, for if Foucault does not tell us precisely what it is that distinguishes the two, he does tell us that the "universal extension of the sign within the field of representation precludes even the possibility of a theory of signification."49

In considering the question of whether Foucault is a structuralist, Wahl concentrates on what he sees as the inadequacies in Foucault's account of representation and signification. In the first place, Wahl distinguishes—and distinguishes clearly—between the two, holding that whereas representation involves a "doubling," within the order of language, of what is outside language, signification involves not doubling but difference, with the meaning of the sign being determined—in classic Saussurean terms—by the difference between it and all other signs.50 With signification, then, language constitutes a genuine and autonomous structure in which an alteration in one signifying element will necessarily alter, through the play of difference, every other signifying element, whereas with representation the "structure" of language is only a doubling of what is not language. Laying great stress on a passage in which Foucault suggests that "the binary theory of the sign" and "a general theory of representation" are linked together in an inextricable relation that "probably extends up to our own time,"51 Wahl condemns Foucault for failing to see that representation and signification are mutually exclusive, and especially for failing to grasp the fundamentally differential nature of the sign. Because Foucault had failed to grasp


48 On representation, signification, and their mutual relations, see ibid., esp. pp. 63-67, 208-11, 303-4.

49 Ibid., p. 65.

50 Saussure (n. 3 above), pt. 2, chap. 4, sec. 4: "Dans la langue il n'y a que des difference s."

51 The Order of Things, p. 67.



the fundamentally differential nature of the sign he had also failed, according to Wahl, to grasp the fundamentally systematic structure of language: "To persist in thinking of the sign within representation is not only to forbid oneself the means of reinstating the formal organization that constitutes the semiological edifice as such: ... it is in truth to resist this organization, in practice to contradict it and from that point on to deny the sign, at the very moment that one seems ready to accord to it its founding place. . . . The 'primacy of representation, the structure of language. Furthermore, elsewhere in entails the denunciation of representation."52 In consequence, Foucault remains, according to Wahl, "on this side of the sign, on this side of discourse, on this side of structure."53 Foucault is to be counted among those who have "not yet" arrived at structuralism. There is, I think, ample reason for agreeing with Wahl that Foucault is not a structuralist in WahFs definition of the term. But the problem is not that Foucault is not yet a structuralist in this sense; it is rather that Foucault is "no longer" a structuralist—that he lies beyond, and not short of, the structuralism of the sign. For WahPs treatment of Foucault fails to recognize that Foucault did indeed hold representation and signification to be incompatible; and while Foucault never raises the issue of difference, his assertions of the post-Classical fragmentation of language are a clear indication of his belief that the structure of things no longer establishes, as in representation, the structure of language. Furthermore, elsewhere in his discussion of Foucault Wahl gives a reading of Les Mots et les choses that, if it were correct, could certainly be taken as placing Foucault under the rubric of the structuralism of the sign; a reading that suggests not that Foucault was not yet a structuralist but that he was a structuralist without knowing it. According to Wahl, Foucault leaves the concept of the sign—which Wahl defines, following Barthes, who follows Saussure, as a relatio between two relata — "curiously in the shadows," even though this concept is, over the length of Les Mots et les choses, shown to be the element that governs the epistemic mutations.54 The configuration of knowledge that makes up any given episteme necessarily implies, Wahl argues, a whole series of interrelations. Each figure within the grid of a configuration, he asserts, functions as the representative of other elements and at the same time as the representative of the configuration in general. On account of these mutual relations, "the


52 Due rot et al., p. 339.

53 Ibid., p. 349.

54 Ibid., p. 306.



episteme, like every order, envelops a semiology."55 Within any given episteme the relations, and hence the signs, are of a given type. As long as the relationes between the relata retain the same nature, the episteme remains the same. Thus, Marx remains within the same episteme as Ricardo because, however much he attacks Ricardo's bourgeois presuppositions, he maintains the same relationship between "the surface circulation of values" represented by the movement of commodities and of their values and "the profound, un-representable fact of the activity that produces them: labor."56 But when the nature of the relationship between the relata changes, then the episteme changes: "The edifices of knowledge topple . . . and there is a change of episteme . . . when the assigned relation of the sign to what it signifies changes: when 4o signify' no longer signifies the same thing."57

It seems to me that if we were to accept this reading of Foucault we would have to acknowledge that he indeed conforms to the structuralism of the sign in its loose sense; for here the sign does appear to be in "an absolutely fundamental position," even though Wahl is right in pointing out that Foucault makes no use of the Saussurean conception of difference. But this reading is in my view an incorrect one, for it falls prey to a misleading metaphorics of depth, of which I shall have more to say below. It is only because Wahl sees the concept of the episteme in terms of depth, order, and firm foundations, and not in terms of dispersion and exteriority, that he is able to give a semiological reading of Foucault, that he is able to assert that the episteme "envelops" a semiology. I here touch on the Dionysian antistructuralist element in Foucault, and more specifically on the fact that for Foucault there are no firm foundations, no original, transcendental signifie to which all signifiants can ultimately refer. And given the absence of a signifie there can be no sign. The episteme stands, in short, beyond the firmly founded world presupposed by the Saussurean conception of the sign.

But structuralism need not be confined to a linguistic or to a semiological sense. For one can detach structuralism from any indenture to the sign, taking structure itself to be the defining feature of structuralism. Both Pettit and Wahl recognize the possibility of a "structuralism of structure," even though they reject such a definition for the purposes of their own analyses. Thus, Pettit tells


55 Ibid., p. 307.

56 Ibid., pp. 308-9.

57 Ibid., p. 309.



us that "I give quite a specific sense to 'structuralism1: unlike some commentators, I do not take it to embrace every science . . . which claims to investigate 'structures.1 ,,5S Wahl, for his part, evokes Levi-Strauss, who would take as the object of the structural sciences "whatever 'has the character of a system,1 that is, any ensemble in which one element cannot be modified without bringing about a modification of all the others. . . -11 As Wahl points out, such a definition would mean that "everything that touches on the idea of structure . . . would fall under the rubric of structuralism. . . Z'59 Where does Foucault stand in relation to structuralism when we broaden the idea of structuralism to take in "everything that touches on the idea of structure?11

The answer to this question depends, of course, on how this structuralism of structure is defined, for the structuralism of structure, like the structuralism of the sign, can be taken in both a strict and a loose sense. Perhaps the best known example of a strict construal of the structuralism of structure is provided by Jean Piaget in Le Structuralisme.60 Whereas Pettit and Wahl base their analyses of structuralism on linguistics and semiology, respectively, Piaget bases his analysis on a congeries of sciences, including not only linguistics but also mathematics, physics, biology, psychology, and anthropology. The effect of this broadening of the field is to rob the concept of the sign of any decisive role within the concept of structuralism. In contrast to Pettit and Wahl, Piaget makes no reference to the sign in his definition of structuralism. For Piaget, structuralism is concerned with structure, and a structure is a "system of transformations.11 Implicit within the Piagetian definition of structure are three ideas. In the first place, a structure, for Piaget, is not a mere aggregate; it is not an accidental collection of elements and their properties. On the contrary, it is a whole, whose elements are subordinate to laws, in terms of which the structure qua whole or system is defined. In the second place, a structure, for Piaget, is subject to transformations, brought about by the play of its governing laws. And finally, a structure, for Piaget, is self-regulating, that is, the transformational laws of the structure "never yield results external to the system nor employ elements that are external to it."61 In short, a structure necessarily entails self-maintenance and closure: it operates according to its own inner system of laws, a


58 Pettit (n. 31 above), p. 33.

59 Ducrot et al. (n. 27 above), p. 10.

60 Jean Piaget, Structuralism, trans. Chaninah Mischler (London, 1971).

61 Ibid., p. 5.



system of laws that never transforms the system into something other than what it is.

In the course of what is, after all, a quite brief book, Piaget pursues the theme of structuralism through a wide array of sciences, showing that his definition of structure is applicable to "groups" and "parent structures" in mathematics, to organisms in biology, to perceptual totalities in psychology, to kinship groups in anthropology, and so on. In each of the fields he examines Piaget is able to find, without much difficulty, investigators who have adhered to a basically "structuralist" methodology. But when—at the end of the book—he finally turns to Foucault, he is unable to find structuralism in the sense in which he defines it. Piaget tells us that Foucault's concept of the episteme seems at first glance to be a promisingly structuralist notion, for it suggests the discovery of "strictly epis-temological structures that would show how the fundamental principles of the science of a given period are connected with one another. . . ."62 But unfortunately Foucault is simply not scientific enough in his approach to carry out this program; instead of developing an appropriate methodology for his enterprise—instead of inquiring, for example, into the criteria for determining when a new episteme can be said to have come into existence or for judging the validity or invalidity of alternative interpretations in the history of science— Foucault relies on "intuition and . . . speculative improvisation."63 Foucault's epistemes, according to Piaget, turn out to be idiosyncratic inventions rather than a genuine attempt to discover the epistemological foundations of the history of science—for Foucault has "no canon for the selection of an episteme's characteristics; important ones are omitted and the choice between alternative ones is arbitrary."64 Foucault's epistemes, in consequence, are not systems of transformation at all, and his structuralism, which in Piaget's view retains all of the negative features of structuralism—such as the devaluation of history and genesis and contempt for functional considerations—without its positive features, can justly be called a "structuralism without structures."65

There is absolutely no doubt that if we take structuralism, as Piaget does, to be essentially a form of scientific methodology, then Foucault is not a structuralist. But the structuralism of structure can be defined in a much looser sense, a sense that is at bottom metaphysical rather than scientific. It can be defined, that is, in the


62 Ibid., p. 132.

63 Ibid.

64 Ibid., pp. 132-33.

65 Ibid., pp. 134-35.



sense proposed by Derrida, as the Apollonian element in the Nietzschean conflict between Apollo and Dionysos. As is well known, Nietzsche argued, in The Birth of Tragedy, that Greek culture at its height was the product of a peculiar and delicate union of the calm, clear, lucent spirit of Apollo and the frenzied, extravagant, esctatic spirit of Dionysos. The Apollonian spirit is the spirit of temperance, moderation, and justice, a spirit that demands the strict observance of the limits of the individual, of the principium individuationis', the Dionysian spirit is the spirit of hubris, of mystical jubilation, of the shattering of the principium individuationis in a savage and ritual unity. As might be supposed, the Apollonian and Dionysian spirits differ radically in their attitude toward forms: the Apollonian spirit teaches the acceptance and retention of forms, while the Dionysian spirit teaches their destruction and re-creation. Each spirit is equally necessary to the existence of a living culture. As Nietzsche puts it, "It is Apollo who tranquillizes the individual by drawing boundary lines, and who, by enjoining again and again the practice of self-knowledge, reminds him of the holy universal norms. But lest the Apollonian tendency freeze all form into Egyptian rigidity, . . . the Dionysian flood tide periodically destroys all the little circles in which the Apollonian would confine Hellenism."66

All primitive peoples, Nietzsche asserts, are amply endowed with Dionysian forces, but the Greeks had, in addition to Dionysos, "the proud, imposing image of Apollo, who in holding up the head of Gorgon to those brutal and grotesque Dionysian forces subdued them."67 The result was the sublimation of these forces into art and culture, the highest expression of which was the tragedy of Sophocles and Aeschylus. But with Socrates this vital collaboration between the Apollonian and the Dionysian, Nietzsche argues, was broken. For Socrates—the bearer, according to Nietzsche, of a degenerate Apollonianism, of an Apollonianism appearing in the guise of "logical schematism"68—was the great exemplar of what Nietzsche calls "theoretical man"—the man who believes in logic, in science, and in conscious knowledge, the man who believes that "thought, guided by the thread of causation, might plumb the farthest abysses of being. . . ,"69 Theoretical man is deeply suspici-


66 Friedrich Nietzsche, "The Birth of Tragedy" and "The Genealogy of Morals," trans. Francis Golffing (Garden City, N.Y., 1956), sec. 9, p. 65; see also sec. 21, p. 128, and sec. 25, p. 145.

67 Ibid., sec. 2, p. 26.

68 Ibid., sec. 14, p. 88.

69 Ibid., sec. 15, p. 93.



ous of the irrational sources of being, of knowledge, and of creativity, holding that culture must be based on conscious intelligence rather than on instinct. Hence, he has a great faith in science, in "the god of engines and crucibles," in "the forces of nature put in the service of a higher form of egotism."70 Hence, too, he opposes the irrational powers of Dionysian art, believing as he does that the beautiful and the reasonable should be made to coincide. The whole of Western culture is caught, Nietzsche argues, within the net of this theoreticism, this rationalism, this scientism; from the time of Socrates onward, the man of theory has been the ideal of Western thought.

But if Nietzsche holds that ever since the great age of Greek tragedy the logic of Socrates has dominated Western culture he also holds that this logic is always on the brink of its own collapse. For logic has its outer limits, its periphery beyond which it cannot move, and it also has an inner core that it cannot grasp. Logic does not extend itself indefinitely but rather "curls about itself and bites its own tail," and even within the circle we have "no way of knowing" how the area "is ever to be fully charted."71 But the man of theory, because he believes that "a culture built on scientific principles must perish once it admits illogic,"72 refuses to recognize the necessarily illogical accompaniment of logic. Nietzsche and, even more, Derrida see their task as that of alerting their fellows to what they allege to be the ultimate illogicality of Western culture. Indeed, Derrida's works, taken together, constitute a single, concerted attack on "logocentrism," on what he regards as the blindly logical orientation of Western thought.

I cannot here deal with Derrida4 s variations on this Nietzschean theme. Suffice it to say that for Derrida structuralism, in the sense of Apollonian formalism, is intimately tied up with the whole of logocentric culture. The most obvious indication of this relationship is to be found, in Derrida's view, in the metaphorical biases and determinations of structuralism. In the first place, structuralism in the Derridian sense is biased toward—or determined by—a metaphorics of light. Indeed, it is this metaphorics of light that links Apollo—the sun god; the god of light; "the 'lucent' one," as Nietzsche calls him;73 the god who stands over "the plastic, Apollonian arts," as opposed to "the non-visual art of music inspired by


70 Ibid., sec. 17, p. 108.

71 Ibid., sec. 15, p. 95.

72 Ibid., sec. 18, p. 112.

73 Ibid., sec. 1, p. 21.



Dionysos"74—with what Nietzsche refers to as "the great Cyclopean eye of Socrates."75 It is in Plato, Nietzsche tells us, that we see most clearly the "gigantic driving wheel of logical Socratism";76 and it is no accident, Derrida holds, that the whole of Platonic philosophy is based on the opposition of light and dark, of which the myth of the cave is only the most obvious indication. Nor, Derrida maintains, is it any accident that nearly all our expressions for thought are connected with visual metaphors: thus, "theory" comes from the word theoria, meaning a looking at, a comtemplation; while the word "idea" comes from eidein, meaning "to see." Indeed, Derrida goes so far as to maintain that "this metaphor of shadow and of light (of showing-oneself and of hiding-oneself)" is "the founding metaphor of western philosophy as metaphysics."77 From Plato onward, Derrida argues, Western philosophy has been indentured to a heliocentric metaphysics that has subjected Dionysian force to Apollonian form. Force, according to Derrida, cannot be thought in terms of eidos, that is, in terms of "form visible to the metaphorical eye," for to grasp "the structure of a becoming, the form of a force," is already to destroy their quality as becoming and as force.78 Indeed, for Derrida the whole project of understanding, of searching for meaning (sens), is thoroughly Apollonian in nature, for understanding requires "the repose of the beginning and of the end, the peace of a spectacle, a horizon, or a face."79 And the entity that Derrida calls "modern structuralism" is, he maintains, an integral part of this larger Apollonian project. Modern structuralism grew up in the shadow of phenomenology, which lacks, according to Derrida, any concept that would permit it to conceive of intensity or of force. This inability to conceive of force has been carried over into modern structuralism, which is biased toward—or determined by—a force-excluding metaphorics of space that in its form and in its implications is closely connected to the central philosophical metaphor of light. As Derrida points out, the notion of structure "refers only to space, morphological or geometrical space, an order of forms and of places."80 The very idea of a center or of an end, without which structure cannot be thought, is an exclusion of Dionysian revel; for, he argues, "the concept of centered structure


74 Ibid., sec. 1, p. 19.

75 Ibid., sec. 14, p. 86.

76 Ibid., sec. 13, p. 85.

77 Jacques Derrida, "Force et signification," in L'Ecriture et la difference (Paris, 1967), p. 45.

78 Ibid.

74 Ibid., pp. 44-45. 80 Ibid., p. 28.



is . . . the concept of a founded play, constituted on the basis of a founded immobility and of a reassuring certitude, itself out of the game."81 Modern structuralism, then, is only the most recent manifestation of the persistent Apollonianism of Western philosophy.

It is hardly necessary to point out that Derrida's thesis (if it is a thesis) or his position (if it is a position) deserves a considerable effort of exegesis and of criticism. Indeed, the distinction between the Apollonian and the Dionysian, and Derrida's elaborations upon that distinction, cry out for exegesis and criticism. Nevertheless, I do not wish at this point to investigate, or even to comment on, the validity of Derrida's conception of structuralism or the validity of the broader Nietzschean assertions that underpin it. I am quite aware of the ragged and illogical opening which this omission leaves in my argument. But my concern here is with the text of Foucault, not with the text of Derrida. I do not wish to ask the potentially destructive question, "Is there any logical basis for the distinction between the Apollonian and the Dionysian?" I wish rather to ask, in an entirely heuristic spirit, the potentially illuminating question, "Is Foucault a structuralist in the metaphysical, or antimetaphysical, sense proposed by Derrida?"

The answer to this question is to be found in an examination of the metaphorics of the Foucaultian text, for when we look at Foucault's works—and more specifically at the works that I would consider to be the most structuralist in nature, namely, Naissance de la clinique and Les Mots et les choses—we find precisely the sort of metaphorics that Derrida has identified as central to the "adventure of the look"82 that in his account constitutes structuralism. For both works are dominated by the theme of looking at space, with the inevitable admixture of a visual and a spatial metaphorics that such a theme implies. Thus, Naissance de la clinique bears the subtitle (iune archeologie du regard medical" and begins with the announcement that "this book is about space, about language, and about death; it is about the act of seeing, the gaze,"83 a statement that is amply confirmed in the rest of the book, in which vision, visibility, invisibility, and space are obsessively recurring motifs. I cite, for example, the following passage from the conclusion, in which Foucault looks back upon the book as a whole: "This book . . . concerns one of those periods that mark an ineradicable chronological threshold: the period in which illness, counter-nature, death, in


81 "La structure, le signe et le jeu dans le discours des sciences humaines," in L'Ecriture et la difference, p. 410.

82 "Force et signification," p. 3.

83 The Birth of the Clinic, p. 31.



short, the whole dark underside of disease came to light, at the same time illuminating and eliminating itself like night, in the deep, visible, solid, enclosed, but accessible space of the human body. What was fundamentally invisible is suddenly offered to the brightness of the gaze, . . . doctors . . . approach the subject of their experience with the purity of an unprejudiced gaze . . . the forms of visibility . . . have changed . . . the abyss beneath illness has . . . emerged into the light of language . . . the patient . . . enveloped in a collective homogeneous space."84

In Les Mots et les choses the metaphorics, though it tends, as in "Las Meninas," to shift from the gaze observing to the space observed, is just as obsessive as in Naissance de la clinique. To enter into the world of Les Mots et les choses is to enter into a world whose fundamental metaphor is the metaphor of arrangement in space; it is to enter into a world that is strangely silent and unmoving, into a frozen world of penetrating glances and arrested gestures. A cursory examination of the prefatory matter of Les Mots et les choses is enough to impress upon the reader the prominence of this metaphorics. Foucault tells us, for example, that in the "Classical age" the "space of knowledge" was "arranged in a totally different way from that systematized in the nineteenth century by Comte or Spencer" (p. xi). He tells us that he had taken a risk in "having wished to describe not so much the genesis of our sciences as an epistemological space specific to a particular period" (p. xi). He asks us where the strange typologies given in Borges's Chinese encyclopedia could be juxtaposed, except in the "non-space of language," in the "unthinkable space that language spreads before us" (p. xvii). He talks about "the table upon which . . . language has intersected space" (p. xvii); about the "space of order" within which knowledge was constituted (p. xxii); about "configurations within the space of knowledge" (p. xxii); not to mention the evocation of such spatial and visual figures as the "relation of contained to container" (p. xvii); and "common ground" (p. xvi); and "sites" (p. xvii) and "the already 'encoded' eye," that is forcibly confined by "linguistic perceptual, and practical grids" (pp. xx-xxi). This metaphorics, with its visual and spatial bias, dominates the whole of Les Mots et les choses, from the initial analysis of "Las Meninas" to the terminal evocation of the erasure of man.

Thus, in Naissance de la clinique, and above all in Les Mots et les choses, Foucault portrays for us—without, I would argue, being fully conscious of what he is doing—a lucent, Apollonian world. In


84 Ibid., pp. 195-96.



short, Foucault conforms in these works to structuralism in the Derridian sense; he is, in Derridian terms, in complicity—albeit a complicity that is entirely unintended—with the very "logocentric" culture whose claim to absolute validity he wishes to contest. Derrida himself has been less than explicit in applying his critique of logocentrism to the works of Foucault. True, in a critique of Histoire de la folie written in 1963 and entitled "Cogito et histoire de la folie," Derrida did hold that Foucault was in complicity with logocentrism, arguing that though Foucault claimed to have written a history of "madness itself . . . before any capture by knowledge" his claim was erroneous, for Foucault was no more able than anyone else to escape from the language of reason: "All our European language, the language of all that has participated ... in the adventure of western reason, is the immense delegation of the project that Foucault defines under the species of the capture or the objectification of madness. Nothing in this language and no one among those who speak it can escape from the historical culpability . . . that Foucault seems to want to bring to trial."85 In the same essay, Derrida claimed to detect in Foucault a "structuralist totalitarianism" that had carried out "an act of enclosure of the Cogito ... of the same type as that of the violences of the classical age."86 But Derrida does not seem to have engaged in any formal and explicit critique of Foucault's more obviously structuralist works, confining himself, in a 1966 essay on the structuralism of Levi-Strauss, to the observation that "the movement of all archaeologies" is in complicity with the attempt to "center" structure, with the attempt to place structure upon a foundation that is itself out of play.87 Nevertheless, despite the lack of explicit connection, the applicability of the Derridian critique of structuralism to the structuralist enterprise of Foucault is beyond question. Wahl was right in observing in Qu est-ce que le structuralisme? that "the schema of structuralism that Derrida attacks is more or less the same as the one to which Foucault adheres . . . ,"88 for Foucault's metaphorical bias—his privileging of sight over sound and of stasis over movement—clearly links him at the most basic level to Apollonian formalism and to all the logocentric themes, the themes of origin and end, of arche and telos, that Apollonian formalism implies.


85 "Cogito et histoire de la folie," in L'Ecriture et la difference, p. 58.

86 Ibid., p. 88.

87 "La structure, le signe et le jeu," in L'Ecriture et la difference, p. 410. See also Derrida's brief comments on the "general theory of epistemes" in "L'Archeologie du frivole," in Etienne Bonnot de Condillac, Essai sur Vorigine des connaissances humaines, precede de "L'Archeologie du frivole" (Paris, 1973), pp. 26-28.

88 Ducrot et al., p. 419.




But is Apollonian formalism the fundamental element in Foucault's work? Can it be said unequivocally that Foucault adheres to structuralism in the sense defined by Derrida? I think not. For Foucault has always been fascinated by Nietzsche, and he has been fascinated by precisely those elements in Nietzsche that tell against the apparent Apollonianism of his visual and spatial metaphorics. But, as I have already said, Foucault's discovery of Nietzsche was slow and halting, and it is only in his later work that his Nietzscheanism comes to the fore.

At least insofar as his approach to the historical world is concerned, Foucault's encounter with Nietzsche has been, I would assert, threefold. The early Foucault tended to see Nietzsche as an exemplar of what Foucault has called "the experience of madness."89 Foucault's early training was in philosophy and in psychology, and he quickly developed an interest in psychopathology. His first book, Maladie mentale et psychologie (1954), was an attempt to rescue insanity from the allegedly dismissive category of "mental illness." In Foucault's view, reason cannot fully know itself unless it engages in a "great tragic confrontation"90 with its opposite, unreason. Hence, unreason is both the mirror image and the furthest extent of reason. But modern culture has done its best to confine and exclude madness—to deny the reality of its existence—thus making the tragic confrontation impossible. A few great spirits, however, including Holderlin, Nerval, Roussel, Artaud, and Nietzsche, have had the true "experience of madness," and these spirits hold the promise that homo psychologicus will one day disappear and that the tragic confrontation with madness will once more take place.

Maladie mentale et psychologie already contains a historical thesis, namely, that madness was once free and that its confinement is only a recent development.91 The Histoire de la folie is an attempt to work out this thesis in detail and thereby to come to grips with the true reality of madness. In it Foucault proposes to return to "that zero point in the course of madness at which madness is an undifferentiated experience, a not yet divided experience of division


89 On "the experience of madness," see Histoire de la folie, pp. i, v, vi-vii, ix, 34-35, 44, 47, 51, 57, 135, 166, 411, 424-25, 459.

90 Michel Foucault, Mental Illness and Psychology [Maladie mentale et psychologie], trans. Alan Sheridan (New York, 1976), p. 75.

91 See esp. ibid., chap. 5, "The Historical Constitution of Mental Illness."



itself"; and starting from this zero point he proposes to write a history, "not of psychiatry, but of madness itself, in its vivacity, before any capture by knowledge."92 In short, Foucault's concern with the experience of madness leads him to see his historical task in much the same way as do orthodox historians, that is, as an attempt to come as close as possible to the reality of the past. To be sure, Foucault holds that orthodox historians have failed to come to grips with the stammering and inarticulate reality of madness, for they have written of madness in the language of that very psychiatry that has attempted, through capture and exclusion, to deny madness. But to hold that orthodox historians have in fact failed in their project is not to deny that the Foucaultian project and the orthodox historical project are here essentially the same. It might further be objected that Foucault's project is much more than the uncovering of the historical reality of madness, for his ultimate concern is with the revaluation of madness in the present. But the project of the orthodox historian also exceeds, through interpretation, the representative project. To take another tack, it might be held that it is in fact Foucault who has failed to come to grips with the historical reality of madness. Thus, we have already seen Derrida's objection to Histoire de la folie, while to the Anglo-American reader the work has the highly artificial flavor of a Hegelian Geistesgeschichte, with the pecularity that Foucault, influenced by Bachelard and Can-guilhem, is careful to reject the idea that the events of his history are arranged in any progressive order.93 But whether Foucault succeeds in coming to grips with the experience of madness is here irrelevant; what is important is that this is what Foucault claims to do, and in so claiming he aligns himself with the classic project of orthodox historiography, which has always asserted that its primary concern is with the provision of a record of objective events and structures. But by the early 1960s Foucault came to see something else in Nietzsche at least equal in significance to the experience of madness, and this new element led him to abandon the view that the historian's project is that of seeking out the solidity of a past reality. What Foucault now saw in Nietzsche is revealed most clearly in his essay, "Nietzsche, Marx, Freud," delivered as a lecture in 1964 but published only in 1967 after the publication of Les Mots et les choses. At least until recently it was customary to read Nietzsche, Marx, and Freud on the model of "depth" interpretation; that is, on


92 Histoire de la folie, pp. i, vii.

93 On this point, see J. J. Brochier, "Prison Talk: An Interview with Michel Foucault," trans. Colin Gordon, Radical Philosophy 16 (Spring 1977): 10-15.



the model of a search for "deep structures." It was customary, in short, to read these thinkers as being engaged in an attempt to find the will to power underlying the moral idea, the social force underlying the ideological fetish, the latent wish underlying the manifest dream. But this is not the way that Foucault comes to read these thinkers: he does not see them as having found a system of interpretation that would link a deceptive superstructure to the firm and comforting reality of a "base." True, he does assert that Nietzsche, Marx, and Freud added the dimension of depth to the field of interpretation. But this depth must be understood, Foucault maintains, not in the comforting terms of "interiority" but rather in the disturbing terms of "ех1епоп1у."94 For in pursuing their descending course, Nietzsche, Marx, and Freud had discovered, according to Foucault, that there is no solid and objective truth that can serve as a point of termination, no final signifie in which all signifiants find their culmination. On the contrary, they had discovered that every interpretandum is already an interpretans—that interpretation does not illuminate some "thing" which passively allows itself to be interpreted, but rather seizes upon an interpretation already in place, "which it must upset, overturn, shatter with hammer blows."95 Thus, Foucault asserts, Marx interpreted not relations of production but rather the interpretation of relations of production. Freud discovered, under the symptoms that his patients exhibited, not the concrete, historical reality of traumas but rather anxiety-charged phantasms, which were already interpretations of historical reality. And finally, above all, Nietzsche demonstrated, through his analysis of language, that there is no signifie originel; for words, which are always invented by the higher classes, do not indicate a signifie but rather impose an interpretation. In consequence, depth itself, now reconstituted as "an absolutely superficial secret,"96 is shown to be a deception, and the task of interpretation, which would otherwise have ended in the discovery of a foundation, becomes an infinite task of self-reflection.

One would expect this rejection of depth interpretation—a rejection which, despite Foucault's attempts to introduce Marx and Freud into the equation, owes much more to Nietzsche than to the other two thinkers97—to have an immediate and profound effect on


94 Foucault, "Nietzsche, Marx, Freud," in Nietzsche, Cahiers de Royaumont, Philosophic no. 6 (Paris, 1967), pp. 183-200. Foucault's use of the term "exteriority" has much to do with his reading of Blanchot (see Bellour [n. 34 above], pp. 7-8, and Foucault, "La pensee du dehors," Critique [Paris], no. 229 [Juin 1966], pp. 23-46).

95 Ibid., p. 189.

96 Ibid., p. 187.

97 And, though I cannot pursue the matter here, I feel compelled to point out that



Foucault's historiography. For the principle of exteriority, in its assimilation of interpretandum to interpretans, of signifie to sig-nifiant, is necessarily antithetical to any attempt to come to grips with the brute reality of the past, with the past "wie es eigentlich gewesen ist." And indeed, Foucault's adoption of the principle of exteriority separates his later works from Histoire de la folie. Thus, in L'Archeologie du savoir he singles out for criticism his use, in Histoire de la folie, of the concept of "experience," which, he asserts, had kept him "close to admitting an anonymous and general subject of history"98—which had kept him close, that is, to the orthodox conviction that the historian stands in some sense outside the movement and uncertainty of history and hence is able to view, with an objective eye, the actual reality of the past. In what is clearly a decisive modification of his earlier project, Foucault tells us, in L'Archeologie du savoir, that "in the descriptions for which I have attempted to provide a theory, there can be no question of interpreting discourse with a view to writing a history of the referent. . . . We are not trying to reconstitute what madness itself might be. . . ."" In short, the later Foucault repudiates the project of Histoire de la folie, arguing that "the stage of 'things themselves' " must be suppressed and that "for the enigmatic treasure of 'things' anterior to discourse" there must be substituted "the regular formation of objects that emerge only in discourse."100

Nietzsche's position on these fundamental issues of truth and interpretation is not as clear cut as Foucault suggests. Jean Granier, in his Probleme de la verite dans la philosophie de Nietzsche (Paris, 1966), and John T. Wilcox, in his Truth and Value in Nietzsche: A Study of His Metaethics and Epistemology (Ann Arbor, Mich., 1974), both investigate the complex problem of whether, and in what ways, Nietzsche believed in truth—which, to use our present terminology, is equivalent to the problem of whether Nietzsche believed in an interpretandum, in a signifie. For a convenient sampling of much of the recent "radical" Nietzsche literature, some of which inclines toward Foucault's view of Nietzsche, see David B. Allison, ed., The New Nietzsche: Contemporary Styles of Interpretation (New York, 1977). For what is perhaps Nietzsche's clearest expression of the theme of the absence of foundations, see Beyond Good and Evil, ed. and trans. Marianne Cowan (Chicago, 1955), paragraph 289, p. 230: "[The anchorite] will suspect behind each cave a deeper cave, a more extensive cave, a more extensive, more exotic, rich world beyond the surface, a bottomless abyss beyond every bottom, beneath every 'foundation.' Every philosophy is a foreground-philosophy: this is an anchorite's judgment. There is something arbitrary in the fact that the philosopher stopped here, that he looked back and looked around, that here he refrained from digging deeper, that he laid aside his spade. There is, in fact, something that arouses suspicion! Each philosophy also conceals a philosophy; each opinion is also a hiding place; each word is also a mask."


98 The Archaeology of Knowledge, p. 16. The translator renders the French experience as "experiment," which is not what Foucault means here.

99 Ibid., p. 47.

100 Ibid., pp. 47-48.



Foucault already acknowledges this modification in his project— this turning away from discourse as a system of signs pointing outward or downward to a signified to a discourse that would systematically form the objects of which it speaks—in the work that immediately followed Histoire de lafolie, namely, Naissance de la clinique. In the preface to the latter book he rejects the classic conception of depth interpretation, which he here refers to under the name of "commentary." As Foucault defines it, commentary "questions discourse as to what it says and intended to say; it tries to uncover the deeper meaning of speech that enables it to achieve an identity with itself, supposedly nearer to its essential truth. . . ."101 Foucault goes on to assert that this activity conceals a strange attitude toward language—an attitude that admits, by definition, an excess of the signifie over the signifiant, holding that it is possible, through a depth analysis, to read the signifie within the signifiant's gaps. To speak about the thought of others, he asserts, has traditionally been to analyze and bring to light the signifie. But, Foucault asks, "must the things said ... be treated exclusively in accordance with the play of significant and signifie, as a series of themes present more or less implicitly to one another?" And is it not possible, he asks, "to make a structural analysis of discourses that would evade the fate of commentary by supposing no remainder, nothing in excess of what has been said, but only the fact of its historical appearance?"102 As early as Naissance de la clinique, then, one finds evidence of the (Nietzschean) principle of exteriority—a principle whose tendency is to turn Foucault away from the reality of the past—within his work. There is thus a fair element of truth in G. S. Rousseau's observation, in "Whose Enlightenment? Not Man's: The Case of Michel Foucault," that Foucault "has evolved a long way from the Foucault of Histoire de la folie, in which he was tied to solid facts and still concerned with historical accuracy"103— though Rousseau does not perceive that this evolution means that Foucault ultimately requires a different type of criticism than the sort he undertakes, one that concerns itself with the theoretical foundations of Foucault's enterprise as a whole.

Nevertheless, the first work of Foucault's to be written under the systematic influence of the principle of exteriority was not Naissance de la clinique, which still remains largely within the historio-graphical orbit of Histoire de la folie, but rather Les Mots et les


101 The Birth of the Clinic, p. xvi.

102 Ibid., p. xvii.

103 Rousseau (n. 13 above), p. 239.



choses, published three years later. But if in Les Mots et les choses Foucault now came to do history under the aegis of the principle of exteriority, seeing his task as the analysis of discourse and not as an attempt to get down to the reality of the past, his employment of that principle remained inconsistent, largely because he continued to conceive his work in terms of the visual and spatial metaphorics that we evoked above. Time and time again Foucault's metaphorics of space involves images of depth and firm foundation that suggest, with great insistence, that despite his apparent adherence to the principle of exteriority he is still involved in depth interpretation in the classic sense, still involved in the attempt to move from what is visible and superficial to what is invisible, profound, and certain. Thus, Foucault speaks, in Les Mots et les choses, of "the fundamental codes of a culture" and of an "order that manifests itself in depth" (p. xx). He tells us that "it is on the basis of this order, taken as a firm foundation, that general theories as to the ordering of things ... are constructed" (p. xxi). He tells us that a culture "finds itself faced with the fact that there exists, below the level of its spontaneous orders, things that are in themselves capable of being ordered, that belong to a certain unspoken order" (p. xx). He tells us that "what I am attempting to bring to light is the epis-temological field, the episteme in which knowledge . . . grounds its positivity and thereby manifests a history which is not that of its growing perfection, but rather that of its conditions of possibility" (p. xxii; my italics in all quotations). And finally, Foucault's predilection for a metaphorics of depth is revealed by his use, throughout the work, of geological metaphors; for although Foucault is ostensibly engaged in an "archaeological" investigation, the archaeological metaphor, with its distant and ambiguous connotations of depth, tends to give way to geological metaphors, with their unequivocal connotations of depth; thus, we find Foucault speaking of erosion (p. 50), of shocks (p. 217), of strata (p. 221), and of "our silent and naively immobile ground . . . that is once more stirring under our feet" (p. xxiv).

If we are to read Foucaultian archaeology according to this metaphorics of depth, then the task of the historian, for Foucault, must be seen as an attempt to approach the past through the strategy of a "symptomatic" reading. The historian attempts, that is, to discover what the manifest discourse of men "really" means, a task that is accomplished by finding, in its gaps and silences, symptoms of the latent discourse underlying and determining it. Of course, one must be careful to note that since Foucault rejects, as subjectivist,



the unities of the book, the oeuvre, and the author,104 one is concerned here not with the discourse of individuals but with the discourse of entire periods—not with what Ricardo, or Lamarck, or Bopp really meant or intended but with the underlying meaning of the episteme itself. On this reading of Foucault, the task of the historian-archaeologist as the grounding of the signifiant in the signifie is reconstituted, for the historian-archaeologist is now seen as attempting to bring "a plethora of elements signifiants" into relation with a "single signified In this way, "one substitutes for the diversity of the thing said a sort of great uniform text, which has never before been articulated, and which reveals for the first time what men 'really meant.' ,405 This "uniform text," this "single signified this latent, underlying meaning to which all superficial discourse is linked, is nothing other than the episteme.

Yet such a reading of Les Mots et les choses, however convincing it might seem at first glance, clearly does not conform to Foucault's own reading of the work. For Foucault asserts in L'Archeologie du savoir, which he implies was written partly in order to repair "the absence of methodological sign-posting"106 in Les Mots et les choses, that it was not his intention that the episteme should be taken as a "basic" or "fundamental" category underlying the intellectual productions of a given historical period. He argues that his procedure in Les Mots et les choses was not "totalitarian"; he was not trying to show that "from a certain moment and for a certain time" everyone thought in the same way; he was not trying to show that beneath surface oppositions "everyone accepted a number of fundamental theses."107 Most of Foucault's readers had seen the classical episteme, for example, as an attempt to characterize the whole of seventeenth- and eighteenth-century thought, and there is a great deal in the text of Les Mots et les choses to support such an interpretation.108 But Foucault now asserts that the classical episteme of Les Mots et les choses was "closely confined to the triad being studied"—that is, to natural history, general grammar, and the analysis of wealth—and is valid "only in the domain specified."109


104 See "What Is an Author?" in Foucault, Language, Counter-Memory, Practice, pp. 113-38, and The Archaeology of Knowledge, pp. 21-27, 92-96, 122.

105 The Archaeology of Knowledge, p. 95.

106 Ibid., p. 16.

107 Ibid., pp. 148-51.

108 See, e.g., The Order of Things, p. 168: "In any given culture and at any given moment, there is always only one episteme that defines the conditions of possibility of all knowledge. ..."

109 Ibid., p. 158.



Other areas of analysis—and this triad, he states, is "only one of many describable groups'410—would yield other episfemes. In fact, Foucault no longer uses the term episteme at all, preferring such expressions as "discursive formation'' and "discursive regularity," expressions that give no suggestion of a distinction of depth. For Foucault's rejection of a metaphorics of depth is now, in LArcheologie du savoir, unequivocal. Thus, he tells us that "we do not seek below what is manifest the half silent murmur of another discourse," that it is not a question of finding "a secret discourse, animating the manifest discourse from within.'411 Indeed, Foucault now distinguishes between analysis and interpretation, telling us that the "analysis of statements avoids all interpretation";112 that is, it avoids all attempts to move from the exterior to the interior, from the manifest to the latent, from the statement to the intention. And in thus refusing to repeat in the opposite direction the work of expression, discursive analysis finally escapes, according to Foucault, from the domination of the subject, of the cogito.

We thus move from Les Mots et les choses to L Archeologie du savoir, the work in which Foucault sets out to examine "the problems of method raised by . . . 'archaeology.' "113 Foucault's essential concern in this extremely complex, difficult, and—I shall argue—self-contradictory work is with the problem of accommodating the project of an archaeology, which goes back to the beginning of his career as a historian, to the principle of exteriority, which entered into his work only after the publication of Histoire de la folie. There is a clear contradiction between the two: the concept of an archaeology, with its implication of a search for origins, is hardly consistent with a principle that denies the existence of origins. Foucault therefore attempts, in L Archeologie du savoir, a modification of the concept of archaeology. The essence of this modification is to be found in his altered view of the relationship between past and present: indeed, for a time the working title of UArcheologie du savoir was Le Passe et le present: Une autre archeologie des sciences humaines.114 The concept of the episteme, as presented in Les Mots et les choses, had suggested that Foucault was engaged in the project of constructing a "portrait" of the past. And since (leaving aside some of the inconsistencies in his account)


110 Ibid., p. 159.

111 Ibid., pp. 28, 29.

112 Ibid., p. 109.

113 The Order of Things, p. xxii n.

114 Foucault, "Reponse a une question," Esprit 36 (1968): 85-74.



he more or less completely denies that there is any relationship between one episteme and the next, he manages more or less completely to divide the past from the present. Thus, the Renaissance and classical epistemes are presumably entirely foreign to those of us who live under the aegis of the modern episteme. In UArcheologie du savoir, however, the governing concept is not the episteme but rather an entity that Foucault calls the "archive." The archive, for Foucault, is not, as one might immediately suppose, the totality of the texts that happen to have been preserved by a civilization, as a kind of accidental detritus lying passively in libraries and other repositories. It is rather "the first law of what can be said, the system that governs the appearance of statements as unique events"; it is "the general system of the formation and transformation of statements"; or, as Foucault puts it in an article published while he was working on UArcheologie du savoir, it is "the play of rules which determines in a culture the appearance and disappearance of statements (e nonces), their remaining (remanence) and their erasure, their paradoxical existence as event and as thing."115 There is clearly much that one could say about this concept and about the role that it plays within UArcheologie du savoir. But given that our concern is centered on the presuppositions of Foucault's historio-graphical enterprise, there is only one point that it is essential to make here, namely, that whereas the concept of the episteme, at least as Foucault presented it in Les Mots et les choses, seemed to be a concept that referred to specific historical periods, the archive is something that remains a more or less permanent determinant of any given culture. It is, in short, a kind of transtemporal constant—a constant whose "never completed, never wholly achieved uncovering . . . forms the general horizon to which the description of discursive formations, the analysis of positivities, the mapping of the enunciative field belong."116 And as a transtemporal constant it provides a linkage between the present and the past—a linkage that reveals Foucault's enterprise to be the portrayal not of the past but rather of the complicities between past and present created by a common discourse.117


115 The Archaeology of Knowledge, pp. 129, 130; "Reponse au cercle d'epistёmologie,,, Cahiers pour I 'analyse, no. 9 (ete 1968), pp. 9-40; quote from p. 19.

116 The Archaeology of Knowledge, p. 105.

117 I take Foucault's use of the term archive in UArcheologie du savoir to be fundamentally different from his use of the term in his 1966 interview with Raymond Bellour, where he speaks of "the general archive of an epoch at a given moment.'' In the 1967 interview with Bellour, the archive becomes "the accumulated existence of discourse" (see Raymond Bellour, "Entretien avec Michel Foucault," Les lettres francaises, no. 1125 [mars 31-avril 6, 1966], pp. 3-4; quote from p. 3).



It is interesting to note that a few commentators saw L'Archeologie du savoir as a work of the utmost importance, one of its American reviewers going so far as to call it k4he most noteworthy effort at a theory of history of the last 50 years . . . truly a work of great magnitude."118 But the more general reaction toward the work has been one of puzzlement rather than of enthusiasm. It is easy to see why this should be so, for it is an excruciatingly difficult book to make sense of. Admittedly, Les Mots et les choses is also a difficult book. But once one has grasped its remarkably simple architecture, and once one has taken account of the fact that Foucault uses some deceptively ordinary words (such as k language" and "discourse") in senses that are highly specialized, then— assuming that one has some background in the subjects of which Foucault speaks—things begin to fall into place. But with L'Archeologie du savoir this never really happens. To be sure, Foucault puts forward some interesting and provocative ideas, particularly when in part 4 (esp. pp. 135-77) he compares the "archaeology of knowledge" with the conventional history of ideas. Nonetheless, the book never seems to form, as Les Mots et les choses most assuredly does, a coherent whole. In consequence, the reader who manages to puzzle his way through it is apt to come out of his reading with a feeling of dissatisfaction or even of overt discontent; for having gone to the book because he believed that this work, at least, would let him know what Foucault is up to, he finds that he knows no more about the foundations and motivations of Foucault's enterprise than he did before.

The manifest failure of the work to form a coherent whole is in part the consequence of Foucault's own deliberate ironism—an ironism that is ironically compounded by his exclusion (pp. 109-10) of the polysemia that is irony's precondition. Notwithstanding this exclusion, Foucault makes it clear at various points in the text that he is writing in an ironic mode—saying one thing but meaning another, making apparently definitive statements that he knows he will contradict tomorrow. In a revealing passage at the end of the introduction to L Archeologie du savoir, an invented reader asks Foucault whether, after so many changes of position in the past, he is going to change his position yet again: "Are you already preparing the way out that will enable you in your next book to spring up somewhere else and declare as you are now doing: no, no, I am not where you are lying in wait for me, but over here, laughing at you?"


118 Mark Poster, Review of The Archaeology of Knowledge, Library Journal 97 (1972): 2736. On the French side, and at greater length, see Gilles Deleuze, "Un nouvel archiviste," Critique (Paris), no. 274 (mars 1970) pp. 195-209.



To this, Foucault replies that he would not take so much trouble and pleasure in writing if he were not preparing a labyrinth within which he might lose himself: "Do not ask me who I am and do not ask me to remain the same: leave it to our bureaucrats and our police to see that our papers are in order" (p. 17). Similarly, soon after informing us of a mode of analysis that will be concerned neither with signifiants nor with signifies, neither with words nor with things, he asserts that "words and things," besides being "the entirely serious title of a problem," is also the "ironic title of a labor that modifies its own form, displaces its own data, and reveals, at the end of the day, a quite different task" (p. 49).

The ironism of L'Archeologie du savoir resides most especially in the fact that whereas it appears to be a rigorously objective attempt to articulate a new scientific methodology it is actually an attempt to demolish everything that has hitherto gone under the name of science. On an overt level, the book has all the trappings of a discourse on method. In the first place, it explicitly and repeatedly advertises itself as a methodological treatise, as a work concerned not with mere "questions of procedure" (which are to be relegated to later empirical studies) but rather with "theoretical problems" (see, e.g., pp. 10-11, 13, 15, 16, 21, 38, 79). In the second place, it begins with a methodical doubt, with an apparent refusal to accept as true anything that is not known to be so; more specifically, it begins with a refusal to accept as valid the various sorts of unity and continuity to which historians usually accede unquestioningly (see esp. pp. 21, 31, 79). In the third place, it proceeds by the formulation of definitions, by the throwing up of hypotheses, by the suggestion of possible directions of research, by the pointing out of consequences, and by the discovery of rules (as can be seen by examining any page in parts 2 and 3 of the book). Fourth and last, it ends by turning to "possible domains of application," within which the "general theory" of archaeology can be put to use and against which the "descriptive efficacy" of "the notions that I have tried to define" can be measured (p. 135).

But when one looks more closely at the book, its supposed "method" and "theory" turn out to be disturbingly elusive. Most importantly, one finds that it is extremely difficult to give any real and determinate content to the major concepts of archaeology, whose apparently rigorous definitions turn out to be almost infinitely elastic. This applies above all to the concept of discourse (see n. 43 above), which defines the framework within which the "archaeology of knowledge" operates, but also to the various other concepts that litter its pages—such as the discursive formation, the rules of forma-



tion, the statement (ёпопсё), the historical a priori, and the archive. Closely connected with this difficulty is the astonishing frequency with which Foucault uses "neither/nor" constructions at crucial points in his argument (see, e.g., pp. 55, 63, 70, 75; but several hundred instances could be listed). Insofar as L'Archeologie du savoir can be said to have a general thesis, I take it to be that the uncovering of the archive can be carried out only by an analysis of "discursive practice'' that is concerned neither with the internal play of signifiants, as are the practitioners of (Mallarmean) literature, nor with the external reference of signifies, as are the practitioners of (orthodox) historiography (for relevant discussions, see pp. 47-49, 62-63, 99, 109, 111), but what the uncovering of the archive is concerned with—since it is concerned neither with words nor with things—is never made clear.

To be sure, some commentators have managed to ignore the disturbingly "unmethodical" aspects of L'Archeologie du savoir and have instead insisted on treating it as if it were the discourse on method that it appears to be. But those who take it at face value are usually forced to acknowledge that the Foucaultian method is strangely defective. This is the case, for example, with a French commentator, Francois Russo, who looks at the book from "a purely positive point of view," treating it as if it were an objective, technical attempt to contribute to the methodology of the history of science.119 Not surprisingly, Russo manages to find a multitude of contradictions in Foucault's proposed methodology, and he is forced to conclude that though the work "has furnished analyses and opened perspectives of the greatest interest," as a systematic methodology it is a failure.120 In a certain sense, however, by taking Foucault seriously Russo fails to take him seriously enough. Admittedly, Russo perceives that L Archeologie du savoir, far from being the neutral, objective work that it claims to be, proceeds from a clear "parti pris," in that it is intended to serve the Foucaultian thesis of the "death of man."121 But—like most commentators on the book—he fails to see that in the Foucaultian scheme of things the death of man also means the death of history, of science, of theory, and of method. For Foucault, as I have already asserted, is


119 Francois Russo, "LArcheologie du savoir de Michel Foucault," Archives de philosophic: Recherches et documentation 36 (1973): 69-105, esp. 105.

120 Ibid., p. 105. For Russo's detailed and systematic exposure of the contradictions and insufficiencies of Foucault's proposed methodology, see pp. 91-105. For an earlier attempt to view Foucault as a (failed) "positivist," see Sylvie Le Bon, "Un positiviste desespere: Michel Foucault," Temps modernes 22 (1967): 1299-1319.

121 Ibid., p. 105.



concerned with the promotion of cultural crisis; he is concerned with fostering the mutation whose intimations he claims, in Les Mots et les choses, already to perceive. He is concerned, in short, with an essentially Dionysian project—that is, with the breaking of the Apollonian forms of science, "the little circles in which the Apollonian would confine Hellenism." And U Archeologie du savoir, in its grotesque explication of the procedures of Apollonian science—in its "cautious" and "stumbling" affectation of scientific humility, in its articulation of principles "so obscure that it has taken hundreds of pages to elucidate them," in its creation of a "bizarre machinery" and its development of a "strange arsenal," in its determined pursuit of a thesis that is "difficult ... to sustain" (pp. 17, 135, 109)—is more than ironical; it is, in fact, a parodistic imitation of what it seeks to destroy, an attempt to out-methodologize Descartes himself. Most of the book's contradictions and obscurities can be linked to its ironical and parodistic intentions.

But a further contradiction, devolving not from these intentions but rather from the utter impossibility of the reconciliation that the book seeks to bring about, remains. For the opposition between archaeology and exteriority is not an opposition that can be overcome by the deft reworking of concepts; on the contrary, it is absolutely definitive in nature. However much Foucault might struggle to prove otherwise, archaeology is not the science of the archive but rather the science of the arche—that is, of the ancient, the primitive; and in its implications of a search for the firm reality of the past it sets out to find something whose existence the principle of exteriority denies. Once more, the substance of Foucault's work lags behind th cutting edge of its irrational intention. In a review written soon after its initial appearance, an Althus-serian commentator hailed U Archeologie du savoir as "a decisive turning-point in Foucault's work."122 But in its retention of archeology it was not so much a turning point as an impasse.

In fact, the "decisive turning point" in Foucault's work came after L Archeologie du savoir, and it involved the abandonment of the entire "bizarre machinery" of that work—an abandonment so complete that Foucault was able to say of his next full-length work, Surveiller et punir, "C'est mon premier livre."123 The transition in Foucault's work, which can be situated in the years from 1970 to


122 Dominique Lecourt, Marxism and Epistemology: Bachelard, Canguilhem and Foucault, trans. Ben Brewster (London, 1971), p. 189.

123 Francis Ewald, "Anatomie et corps politiques," Critique (Paris), no. 343 (decembre 1975), pp. 1228-65.



1972, has implications going far beyond the realm of historiography, and I cannot deal in detail here either with the transition itself or with the posttransition writings. Suffice it to say that according to Foucault's own account—an account confirmed by his most recent works—the transition involved an alteration in his conception of power. From the beginning of his intellectual career, Foucault had been concerned with the problem of social controls, as his writings on psychiatry, on madness, and on somatic medicine amply demonstrate. But the conception of power on which those writings were based was a purely negative conception: power, for the archaeological Foucault, was an entity whose importance was to be found in the fact that it "excludes," "represses," "censors," "abstracts," "masks," and "conceals."124 This conception of power had served Foucault—and had apparently served him well—in his investigations of the mental asylum and of the hospital. But after the publication of L'Archeologie du savoir Foucault turned to the study of the prison; and here he found—or claimed to find—phenomena that a purely negative conception of power could not accommodate. Most importantly, the institution of the prison had ostensibly been founded in order to repress delinquency; but almost from its very foundation there had been unceasing complaints that, far from repressing delinquency, it was only serving to encourage it. How, Foucault asks, can one account for the fact that for nearly 150 years criminologists have talked of the "failure" of the prison and yet the prison still exists? The answer to this question, he asserts, is that the prison has not failed. For the ostensible aim of the prison was not its real aim; the prison was in fact founded in order to encourage delinquency, and thus to provide a rationale for the construction of the vast apparatus of control and discipline without which the alleged freedoms of bourgeois society could not exist. In other words, power is here seen to be not a negative but a positive phenomenon: "Power produces; it produces reality; it produces domains of objects and rituals of truth."125 The new attitude toward power of which Surveil-ler et punir gives such eloquent testimony is even more firmly embodied in the more recent Volonte de savoir, which Foucault has lately designated as the first book in which he really liberates himself from the search for "things themselves in their primitive vivacity," the first book in which he fully frees himself from the idea that power is "bad, ugly, poor, sterile, monotonous, and dead."126


124 Discipline and Punish, p. 194.

125 Ibid.

126 "Foucault: Non au sexe roi," p. 113 (English trans., Telos, no 32, p. 158).



Foucault is saying nothing new here; for commitment to the productivity of power is the supremely Dionysian insight, well known to "that Dionysian monster, Zarathustra."127 In asserting power to be a creative force Foucault has now distanced himself from the Apollonian structuralism of his earlier work in which the excluded, the suppressed, the censored, the abstracted, the masked, the hidden, was alone en jeu, and power—that is, the center from which these operations were created—was hors jeu. That is to say, Foucault has now rejected the Apollonian conception of centered structure that dominated, willy-nilly, his earlier work.128 He has acceded, in essence, to the criticisms of Derrida—which is not to say that he was "influenced" by Derrida (though he may very well have been), but only that the element criticized by Derrida in the "early" Foucault is precisely the element against which the later Foucault rebels.129

Not surprisingly, this rebellion against structuralism brings with it an alteration in the metaphorics of the Foucaultian text. Foucault does not abandon a visual and spatial metaphorics, for at least in Surveiller et punir this metaphorics still plays a prominent role. But the movement beyond structuralism as Derrida defines it does not entail the dropping of visual and spatial metaphors; indeed, without these metaphors coherent discourse would be impossible. Rather, it requires a consciousness of both the existence and the implications of structuralist metaphors; it requires that in employing metaphors of space, of foundation, or of structure one recognize that these are indeed metaphors and nothing more. This recognition is a prominent feature of Surveiller et punir. To explain adequately how Foucault here goes beyond the structuralist metaphorics of his archaeological period to a consciously antistructuralist metaphorics would take us far beyond the limits of this paper. But some indication of the alteration can be gained through a brief comparison of this book with Naissance de la clinique, the archaeological work with which it is most closely linked. As we have already seen, the latter is replete


127 Nietzsche (n. 66 above), The Birth of Tragedy, preface to 1886 edition, p. 15.

128 On the theme of decentering in Foucault, see The Archaeology of Knowledge, pp. 12-13, and esp. "Theatrum Philosophicum," first published in 1970 and conveniently available in translation in Foucault, Language, Counter-Memory, Practice, pp. 165-96. This essay is an important early manifestation of the "later" Foucault.

129 I might further add that there are important differences between Foucault and Derrida. In the "nouvelle edition" of Histoire de la folie (Paris, 1972), p. 602, Foucault attacks Derrida for reducing "discursive practices" to "textual traces," and for teaching that there is "nothing outside the text"—which amounts to an attack on Derrida for his tendency to consign himself to an aesthetic realm that is "above," and in large measure indifferent to, history. But fundamentally Foucault inhabits the same aesthetically inventive and irrealistic territory as Derrida.



with visual and spatial metaphors; indeed, the work is constructed around the concept of the regard medical—the "'medical gaze." But Foucault here took the concept of "regard" in an entirely negative sense, as if the "regard" were a passive observer gazing from a fixed point of view upon an objective field of knowledge, and he saw his task as that of reconstituting the space that this gaze had surveyed. In Surveiller et punir, metaphors of vision and of space are employed in an entirely different way.130 Here these metaphors do not describe a rigid and unmoving field existing at some time in the past; they describe, rather, an active field of conflict in which, Foucault maintains, we are all engaged. Foucault's concern in Surveiller et punir is with the disciplinary systems—the systems of micro-pouvoirs—which, he asserts, exist beneath the surface of bourgeois society and control our behavior without our knowledge. These disciplinary systems, Foucault holds, depend upon a regime of observation, surveillance, and inspection whose model Foucault finds in the Panopticon of Bentham. The exercise of discipline, Foucault asserts, "presupposes a mechanism that coerces through the play of the glance {par le jeu du regard)."131 Furthermore, this disciplinary power "is exercised through its invisibility; at the same time it imposes on those whom it subjects a principle of compulsory visibility."132 In short, the gaze is not a passive entity but an active force engaged in its own strategy of domination, and Foucault's counterstrategy in Surveiller et punir is to reverse the game by depriving the gaze of the invisibility it has so long cultivated. Vision, which had provided the framework of Foucault's earlier work, is now to be exposed in all its operations—or so Foucault claims.133 At this point we arrive at the essential core of Foucault's historical project (insofar as it can be said to have an "essential core"); for, as the reorientation in his metaphorics suggests, he claims now to be concerned not with the Apollonian portrayal of dead past—a past that, as far as we are concerned, exists in a state of "Egyptian rigidity"—but rather with the active play of forces in the present.134


130 In L'Archeologie du savoir, Foucault had already expressed his dissatisfaction with the concept of the "regard medical" (see The Archaeology of Knowledge, p. 54n, and Deleuze, p. 202 [n. 118 above]).

131 Discipline and Punish, p. 170 (translation altered).

132 Ibid., p. 187.

133 The most recent ultra criticism of Foucault denies the radicality of Foucault's critique (see Jean Baudrillard's brief essay, Oublier Foucault [Paris, 1977], where Foucault's conception of power is held still to be a "structural notion" p. 53). I, too, deny the radicality of Foucault's critique—but not for the reasons that Baudrillard adduces.

134 A claim likewise denied by Baudrillard, who accuses Foucault of "nostalgia" (ibid., p. 87) and thus reveals the unbridgeable gulf that separates his critique of Foucault from the critique toward which I am here aiming.



Indeed, I would argue that despite the antiquarian suggestions of archaeology Foucault never at bottom had the orthodox historian's passion for the objective apprehension of the past, even in Histoire de la folie. But I would also argue that there was a sense in which he did not know that his concern was not ultimately with the past at all, and that it was only with the reorientation in his conception of power that he came to see this. Certainly, only in his more recent writings and utterances can one find unequivocal expressions of an (allegedly) presentist concern.135 Thus, in an interview given in 1971, Foucault informs us that "it is a question, basically, of presenting a critique of our own time, based upon retrospective analyses"; and he goes on to explain that "what I am trying to do is grasp the implicit systems which determine our own most familiar behavior without our knowing it. I am trying to find their origin, to show their formation, the constraint they impose upon us; I am therefore trying to place myself at a distance from them and to show how one could escape." For Foucault, it is now explicitly a question of shaking things up, of putting into play—en jeu—"the systems that quietly order us about."136 In a more recent interview, given in 1975, Foucault emphasizes even more strongly the total insertion of his works into the context of the present: "Writing interests me only insofar as it enlists itself into the reality of a contest, as an instrument of tactics, of illumination. I would like my books to be, as it were, lancets, or Molotov cocktails, or minefields; I would like them to self-destruct after use, like fireworks." It is necessary, Foucault asserts, for historical analysis to be a real part of "political struggle"—not that it attempts to give such struggles a "guiding thread" or a "theoretical apparatus," but rather that it "constitutes" their "possible strategies."137 It is in Surveiller et punir that this concern first comes fully into play. Foucault tells us in this work that "I have learnt not so much from history as from the present" that "punishment in general and the prison in particular belong to a political technology of the body"; and he goes on to say that it is of the prison in its actuality "that I would like to write the history," an enterprise which he characterizes, not as "writing a history of the


135 Admittedly, in the 1966 interview with Bellour he asserts that "it is not a fault when these retrospective disciplines find their point of departure in our present situation"; but this is hardly different from the orthodox historian's belief in the problem-generating capacity of the present (see Bellour, "Entretien avec Michel Foucault," p. 3).

136 John K. Simon, "A Conversation with Michel Foucault," Partisan Review 38 (1971): 192-201. t

137 Jean-Louis Ezine, "[Entretien avec] Michel Foucault,' Nouvelles Litteraires, no. 2477 (mars 17-23 1975), p. 3.



past in terms of the present," but rather as "writing the history of the present."138

Foucault's claim to be concerned with the present brings us finally back to Nietzsche, for Foucault identifies this concern with the Nietzschean conception of genealogy. As far as I know, his first reference to the affinity between his work and the historico-critical project of Nietzschean genealogy occurs in his 1967 interview with Raymond Bellour, where he asserts that archaeology "owes more to the Nietzschean genealogy than to structuralism properly so called."139 But it was only when he came to see himself as unequivocally a presentist that he stopped characterizing his work as archaeology and began to characterize it as genealogy instead. Thus, in Surveiller et punir he tells us that the book is intended as "a correlative history of the modern soul and of a new power to judge; a genealogy of the present scientifico-legal complex from which the power to punish derives its bases, justifications and rules, from which it extends its effects and by which it masks its exorbitant singularity."140 And in a 1975 interview in which he comments on Surveiller et punir he tells us that "if I wanted to be pretentious, I would give 4he genealogy of morals' as the general title of what I am doing."141 With this transition from archaeology to genealogy—a transition that some of Foucault's reviewers seem intent on obscuring142—Foucault has finally acknowledged his own lack of interest in the past.

But what does Foucault's genealogy entail besides a radical rejection of the past? Let us proceed by indirection; let us proceed, that is, historically. Nietzsche's presentist, genealogical view of history was articulated within the context of a culture whose dominant mode of intellectual apprehension was historical. In the form of two complementary but nevertheless distinct historicisms, historical modes of thought played a central role in nineteenth-century intellectual life. One of these historicisms, which found its archetypal manifestations in the work of Hegel and of Comte—so different and yet in their underlying approaches to history so similar—was cen-


138 Discipline and Punish, pp. 30-31. See, more recently, "Foucault: Non au sexe roi," p. 113: "C'est la que commence le vrai travail, celui de l'historien du present4' (English trans., Telos, no. 32, p. 159).

139 Bellour (n. 34 above), p. 9.

140 Discipline and Punish, p. 23.

141 "Prison Talk: An Interview with Michel Foucault," p. 15; see n. 93.

142 Thus, in Gilles Anquetil, "Le Nouveau Pacte de Faust" (review of La Volonte de savoir), Nouvelles Litteraires, no. 2564 (23-30 decembre 1976), p. 9, the reviewer refers to Foucault's studies of madness and of the clinic as "genealogies"; see also "Prison Talk," p. 10.



tered on the idea of development, on the idea of an ordered, lawful movement from stage to stage in the historical process. The fundamental assumption of this type of historicism, namely, that "an adequate understanding of the nature of any phenomenon and an adequate assessment of its value are to be gained through considering it in terms of the place which it occupied and the role which it played within a process of development,"143 was the most nearly universal element in nineteenth-century thought, coming closer to giving a unity to the intellectual history of that century than any other theme. The other form of historicism, which was associated with the emergence of the modern academic discipline of history, was much less pervasive, but its impact on the academic milieu within which Nietzsche worked was nevertheless immense. This second form of historicism was centered on the idea that every historical entity possesses its own unique and incomparable value, an idea that, divested of its idealist origins, came to underpin the view that history must be scientific and objective in nature. The historicism of development and the historicism of individuality worked together to raise the value of a specifically historical consciousness. Though the elements of historicism had certainly been present in Western thought before the nineteenth century, it was only in that century that historical modes of thought moved to the center of the intellectual stage—that history became, as it were, essential to knowledge, essential to intellectual life in general.

Nietzsche reacted strongly against this rise in the value of the historical, and in his Use and Disadvantage of History for Life, an essay written soon after The Birth of Tragedy and serving as a kind of coda to it, he sharply attacked what he conceived to be the hypertrophy of historical culture in his own time. Since we are concerned in this essay not with Nietzsche but with Foucault, I can deal with Nietzsche's views on history only very briefly. Suffice it to say that the essential theme of The Birth of Tragedy—namely, the theme of the incessant struggle between Apollo and Dionysos—was carried over into The Use and Disadvantage of History, for Nietzsche associates the historical culture of his own time with Socratic theoreticism which, with its bias toward science and logical understanding, had destroyed myth and displaced poetry from its native soil. This is not to say that Nietzsche opts for a historical barbarism that would reject all knowledge of the past. For he believed that culture in the higher sense cannot exist without memory; but if that culture is to be a living culture it must know when to


143 Mandelbaum (n. 2 above), p. 42.



forget the past, when to strike out on its own: "This is the point that the reader is asked to consider; that the unhistorical and the historical are equally necessary to the health of an individual, a community, and a system of culture."144 Nietzsche's complaint against his own time was thus not that it was historical, but rather that it was too historical; it suffered not from history but from an excess of history. Time and time again Nietzsche complained that the study of history had become an end in itself, detached from the real needs of men. Historical knowledge, he asserts, streams in upon us from inexhaustible sources, but we have failed to digest this knowledge, we have failed to impose upon it our own self-created, life-endowing form. Against the reigning historicisms, which seemed to preach, respectively, subordination to the general process of history and subordination to the objective reality of the past, Nietzsche articulated a new, relativistic historicism that claimed to subordinate the past to the needs of the present and the future. This new historicism would attempt to restore "the clarity, naturalness, and purity of the connection between life and history. . . ,"145 It would recognize that we need history "for the service of the future and the present," that we need it "for life and action, not as a convenient way to avoid life and action.'446 It would likewise recognize that the true understanding of history is vouchsafed, not to those who passively observe history, but rather to those who actively use it, linking it instinctively to their own needs and actions in the continuing present. For in Nietzsche's view, "You can only explain the past by what is highest in the present. . . . Only he who is engaged in building up the future has a right to judge the past."147

What are the consequences of seeing history in these radically presentistic terms? Most obviously, presentistic history must necessarily be perspectival in nature; it must give us, not the truth of the past, but a point of view on the past. And indeed, in The Genealogy of Morals Nietzsche rejects the claim of "our modern writers on history" to be "a mirror of events,"148 attacking their scientific pretentions as nothing more than a fearful attempt to deny, through the assertion of a single truth, the multiple truths of things. In the


144 Friedrich Nietzsche, The Use and Abuse of History, in Complete Works, ed. Oscar Levy (Edinburgh, 1909), 2: 10. It should be noted that the Levy edition does not accurately render the German title of this work, Vom Nutzen und Nachteil der Historie fur das Leben.

145 Ibid., p. 30.

146 Ibid., pp. 30, 3.

147 Ibid., pp. 55, 56.

148 Nietzsche (n. 66 above), The Genealogy of Morals, 3d essay, sec. 26, p. 293.



same work, Nietzsche vehemently condemns "the hallowed philosophers' myth of a 'pure, will-less, painless, timeless knower' "; for the concept of such a knower presupposes "an eye such as no living being can imagine, an eye required to have no direction, to abrogate its active and interpretative powers—precisely those powers that alone make of seeing, seeing something. All seeing is essentially perspective, and so is all knowing."149 Foucault, too, recognizes the perspectivism of a presentistic historiography, as he indicates in an essay that is extremely important for an understanding of his genealogical apprehension of history, "Nietzsche, la genealogie, l'histoire," first published in 1971. In this essay he contrasts orthodox historiography with Nietzschean genealogy. "The history of the historians," Foucault declares, "gives itself a point of support outside of time; it claims to judge everything according to an objectivity of the apocalypse; but it can do this only because it presupposes an eternal truth, a soul that does not die, a consciousness always identical with itself." In opposition to the "regard de fin du monde" cultivated by orthodox historiography, Nietzschean genealogy, according to Foucault, "does not fear to be a perspecti-val knowledge. . . . The historical sense, as Nietzsche understands it, knows itself to be perspective, and does not refuse the system of its own injustice. . . . Rather than feigning a discreet effacement in the face of what it is looking at, rather than seeking therein its law and subordinating each of its movements to it, it is a gaze that knows from where it looks as well as what it is looking at."150

But the nature of Nietzsche's perspectivism must be carefully attented to, for despite Foucault's account of "the history of the historians" few practising historians would deny the perspectival nature of their own work. On the contrary, orthodox historiography attributes to interpretation—that is, to the subjective viewpoint of the historian—a legitimate and indeed a necessary role in the historical account. It is not perspectivism as such, in which differing and apparently contradictory perspectives are taken to be simply the varying profiles of a single invariant reality, that distinguishes Foucault's version of Nietzschean genealogy from orthodox historiography but something much more radical, namely, a rejection of the conception of historical reality itself. For, at least in one of his modes,151 Nietzsche is doing far more than asserting the legitimacy


149 Ibid., sec. 12, p. 255.

150 "Nietzsche, Genealogy, History," in Foucault, Language, Counter-Memory, Practice, pp. 152, 157 (my translation, however).

151 Which is not, I would argue, his only mode of apprehending history. But a thorough exploration of the Nietzschean apprehension of history—an issue that is



of looking upon the hard reality of the past from a variety of angles. He is asserting that the past is not a hard reality—he is asserting that every supposed historical reality is merely a foregound, a mask, an arbitrary stopping point, covering up an infinitude of other "realities." In short, genealogy denies the existence of a res gestae that would be the object of the historical account, holding rather that each historical reality is only an excuse for our stopping at one point and not at some other point in the vast and unending play of interpretation. Thus, the genealogical answer to the burden of history is to be found not in a perspectival reinterpretation of historical reality in the hope of accommodating that reality to the needs and interests of the continuing present; it is to be found in the denial of historical reality, in the assertion that "historical reality" is a mere projection of present needs and interests.

History has always been taken to be a "representative" or "descriptive" verbal activity, an activity whose "final direction of meaning" is necessarily outward. In this sense it has been contrasted with myth, poetry, and literature in general, in which—at least if we accept the aesthetic views of Mallarme, Northrop Frye, and Foucault himself—the final direction of meaning is inward: for while history is normally intended to represent things external to it and, as history, has been valued in terms of the accuracy with which it does represent those things, in literature "questions of fact or truth are subordinated to the primary literary aim of producing a structure of words for its own sake. . . ,"152 In rejecting the concept of a historical reality separable from the needs and interests of the historian himself—the concept of a res gestae that the historia rerum gestarum seeks to double—one necessarily rejects the view that history is a representative activity. I cannot deal here with the full implications of such a reordering. Suffice it to say that Nietzsche's rejection of representation in history, which is of a piece with his rejection of the stilo rappresentativo in music,153 his rejection of naturalism in the drama,154 and his rejection of the truth-conveying function of language,155 means the rejection of history as history and its recreation as literature, as poetry, as myth. Hayden White


quite as complicated as, and closely related to, the issue of the Nietzschean apprehension of truth (see n. 97 above)—is beyond the resources of the present essay, where I am giving an incomplete and in some respects one-sided account of his attitude toward history.

152 p^ (n 44 above), p. 74.

153 The Birth of Tragedy, sec. 19, pp. 113-21.

154 Ibid., sec. 11, pp. 69-75.

155 See "On Truth and Lie in an Extra-Moral Sense," excerpted in The Portable Nietzsche, ed. and trans. Walter Kaufmann (New York, 1968), pp. 42-47.



conveys the nature of this alteration well when he asserts that Nietzsche's "metaphorical" historiography is "the means by which the conventional rules of historical explanation and emplotment are abolished. Only the lexical elements of the field remain, to be done with as the historian, now governed by 4he spirit of music,' desires. . . . The historian is liberated from having to say anything about the past; the past is only an occasion for his invention of ingenious 'melodies.' Historical representation becomes once more all story, no plot, no explanation, no ideological implication . . . that is to say, 4myth' in its original meaning. . . ."156

Foucault is, I believe, fully aware of these wider implications of the genealogical history that he has now adopted as his own; he is fully aware of the fact that his history is essentially fabulation and myth. Even in 1967, when he was still claiming to be an archaeologist, he was able to tell Raymond Bellour that Les Mots et les choses "is purely and simply a 'fiction.'" But this insight, which conflicted so radically with Foucault's scientific pretentions of that period—pretentions that attained, in L'Archeologie du savoir, an almost baroque intensity—was not followed up; indeed, the insight was immediately vitiated by Foucault's assertion that the fiction had not been invented by Foucault, but was an expression of the relationship between the epistemological configuration of our own epoch and the "whole mass of statements" emanating from the past.157 With Foucault's transition from archaeology to genealogy, however, which finally liberated him—if that is the word—from the "structuralism" of his earlier work, he has been able to achieve a more consistent conception of what his enterprise involves. Thus, a recent interviewer, asking him about the "fictional character" of La Volonte de savoir, evoked the following response: "As for the problem of fiction, it is, for me, a very important problem: I am fully aware of the fact that I have never written anything but fictions. I do not mean to go so far as to say that fictions are beyond truth Qiors verite). It seems to me that it is possible to make fiction work inside of truth, to induce truthful effects with a fictional discourse, and to operate in such a manner that the discourse of truth gives rise to, 'constructs,' something that does not yet exist, and thus 'fictionizes.' One 'fictionizes' a history from the basis of a political reality that makes it true, one 'fictionizes' a not yet existing politics from the basis of historical truth."158 Foucault's history, then, is a fiction.


156 White, Metahistory, p. 372.

157 Bellour (n. 34 above), p. 7.

158 Finas (n. 37 above), p. 6.



But it is not intended to be a frivolous fiction. Rather, it is intended—in an almost Sorelian sense—as a weapon in contemporary social and political struggles. For as Foucault stated in another interview, dating from 1974, "Memory is actually a very important factor in struggle ... if one controls the memory of the people, one controls their dynamism. ... It is vital to have possession of this memory, to control it, administer it, tell it what it must contain."159 It should be noted that Foucault is here speaking against what he sees as the surreptitiously conservative propagandizing of the recent mode retro in the French cinema. Nevertheless, these observations well convey, if in a negative fashion, what Foucault takes to be the central justification for his historical enterprise.

It is not given to us to view Foucault definitively—to view him with a regard de fin du monde', we must rather view him from the elusive and shifting standpoint of our own historical situation, which happens also to be the historical situation within which Foucault, our contemporary, lives and works. Nietzsche's genealogy was articulated within the context of, and as a reaction against, the nineteenth-century passion for the historical. More specifically, it was a protest against 44he modern historical education" that in his view accounted for "the premature grayness of our present youth" and for the impairment of the plastic, creative power of life.160 It was an attempt to counteract the simultaneous rigidity and confusion that nineteenth-century historical consciousness had allegedly induced and thus to free once more the springs of creativity. It was an attempt to use the "unhistorical" and the "suprahistorical" as antidotes to the "historical": the unhistorical being the power "of forgetting, and of drawing a limited horizon round one's self"; the suprahistorical being the power "that turns the eyes ... to that which gives existence an eternal and stable character, to art and religion."161 If we are to determine the value of Foucault's elaboration of Nietzschean genealogy for our own time and place we must do so at least in part in terms of the balance, within our culture, of the historical, the unhistorical, and the suprahistorical. Nietzsche was able, in The Use and Abuse of History, to refer to history as "a Western prejudice"162 and to complain of the burdens of a historical education. Does history still constitute a burden? I think not; for one


159 "Film and Popular Memory: An Interview with Michel Ро^аии," trans. Martin Jordin, Radical Philosophy, no. 11 (Summer 1975), pp. 24-29; quotes from pp. 25 and 26. The original appeared in Cahiers du cinema.

160 Nietzsche, The Use and Abuse of History, pp. 89-90.

161 Ibid., p. 95.

162 Ibid., p. 15.



of the most striking features of the intellectual history of the West in our century has been the turning aside from history and from historical modes of thought. The historical experience of the West has worked to make our culture less ''historical," perhaps, than at any other point since the beginning of the Renaissance. The major intellectual movements and fashions of the twentieth century have all been nonhistorical in their orientation while there has been a dramatic shrinkage in the historical branches of disciplines, such as philosophy and literary criticism, that have traditionally had an important historical component. Even among persons of great intellectual attainment history has tended to become an irrelevancy; it has tended to become a storehouse from which to draw examples at will, abstracted from the actual contexts of their creation.

It is easy to suggest reasons for this devaluation of the historical, though more difficult to assess their relative weights and to determine the complicities between them. First of all, our experience of glut, our experience of the sheer weight of the historically given, has tended to turn us against history. In the late eighteenth century, within the framework of an ideology that emphasized the universality of reason, Individualitat seemed marvelously liberating; by the late nineteenth century it was already beginning to seem oppressive. In the second place, our experience of cultural multiplicity, by which I mean not only the widely publicized work of ethnologists but also the infamous "knowledge explosion" with its tendency to infinite scholarly fragmentation, has destroyed the conception of a common humanistic culture, which was often adduced by traditionalists as the primary reason for the study of history.163 In the third place, our experience of sheer destructiveness, on a more massive scale than has ever been seen before—the decimation of entire generations and of entire races, for example—has destroyed the conception of historical progress that underpinned so much of nineteenth-century his-toricism. And finally, there is our experience of the cumulative technological revolution of the last seventy years or so—a revolution that has altered our environment and our conditions of life in a radical and historically discontinuous way.

But whatever the reasons, historical culture no longer occupies an important place within the literate culture of our time. There is, to be sure, some evidence of a popular hunger for history and for the sense of reality that history can bring.164 This hunger proceeds,


163 See, now, the traditionalist at bay: "Troy will always be, in the foreseeable future, an integral part of the Western cultural heritage'4 (Frye, p. 102; my italics).

164 On this point, see John Lukacs, "The Future of Historical Thinking," Salmagundi 30 (Summer 1975): 93-106.



however, not from the addictive craving diagnosed by Nietzsche, but from a more elemental sense of lack. Foucault is wrong—or, perhaps better, no longer right—when he tells us that "in our culture, at least for the last few centuries, discourses hang together {syenchainent) on the mode of history" and that "in a culture such as ours, all discourse appears against the background of history (apparait sur un fond d'histoire)."165 In arguing for a mythical, presentistic, genealogical view of history, Nietzsche was taking upon himself the task of thinking "thoughts out of season"; he was following in the footsetps of "the great 'fighters against history/ "166 In arguing for a mythical, presentistic, genealogical view of history, Foucault is thinking seasonable thoughts, not unseasonable ones, and it is the orthodox historian who, paradoxically enough, is the fighter against history. I do not mean here to condemn Foucault, for he is a man of much brilliance, who frequently illuminates the landscape in unexpected ways. Nevertheless, in opposing Apollonian culture he is behind the times. He is engaging in an immense con game. He is trying to set fire to the ashes of the library at Alexandria. Let us get what entertainment we can from the spectacle; but let us remember that that is precisely what it is—a spectacle, a play, a performance.

Do I mean, then, that we should not take Foucault seriously? You misread me. To be sure, he should not be taken seriously as a historian. That is to say, we should recognize, and we should inform others, that Foucault is not interested in the interpretation of the past. To read Foucault's myths as if they were a portrayal of the past itself—to read them as if they bear some necessary and comprehensible relationship to anything that actually happened in the past—is to confirm and strengthen the widespread historical illiteracy of the present day. And yet if Foucault should not be taken seriously as a historian, he most emphatically should be taken seriously as an indication of where history now stands. The popular hunger for history—one might almost say the human hunger for history—is something to which orthodox academic historiography finds it almost impossible to respond. For we are faced by a paradox. Even as orthodox historiography has been expanding the range of its subject matter and rendering its methodology more and more technical and sophisticated, two countermovements have been occurring: the higher intellectual foundations of history have been crumbling, and its accessibility and immediacy have been declining. It is ominously significant that many historians have trouble justifying their vocation


165 Bellour (n. 34 above), p. 9.

166 Nietzsche, The Use and Abuse of History, p. 74.



in terms comprehensible to those who are not themselves historians—indeed, that such justifications are frequently couched, in terms that are entirely intra muros and negative, as the defense of the reality of the past against the misinterpretations of other disciplines. Foucault's work is symptomatic of a "higher" culture in which history as a science can no longer justify itself because the knowledge of the past as such appears to have no independent cultural value or purpose, and of a "lower" culture which history as a science does not reach. Detached from both, the orthodox historian finds himself unable to justify his analytical vocation—unable to justify his penchant for subjecting myth to the rule of reality, to the Apollonian rule of science. The solidity of the past gives way—in Foucault and in his followers—to the ersatz reality of myth.

And this takes us to the term of our criticism, which lies beyond history. Admitting for the moment that there is a genuine element of liberation in Foucault's opting for the free play of the interpretation of interpretation rather than for the circumscribed work of the interpretation of things, and admitting that in a culture bent down under the weight of a historical factuality—if ours were such a culture—the Foucaultian option might well perform a valuable contrapuntal function, is it not true that this option entirely lacks the radicality it claims? For is it not true that it fails to touch the roots—that precisely because it is myth it renounces the attempt to plumb the reality of human social life, which is the realm within which all change must ultimately be effected? I do not deny that myth may be an instrumentally useful stimulus to social action; I merely deny that it is a substantively rational guide for social action. Those who reject the distinction between myth and science do so at their peril—and at ours. At best, they confine themselves within a rhetoric that has no issue upon the real world of social action: at worst, they reap the whirlwind. Caveat emptor. Foucault's mythification of the past is also at the same time a mythification of the present. I hold, with Foucault and with Nietzsche, that the historian's concern with the reality of the past—if it is nothing more than that—is trivial. This is why I hold, for reasons now transcending the historical, that Foucault must not be taken seriously—and at the same time must be taken very seriously indeed.


The Journal of Modern History, Vol. 51, No. 3 (Sep., 1979), pp. 451-503

Hosted by uCoz