(This is a developed version of the paper delivered at Cheiron '94. It has not been presented or published in this extended form. Minor editorial changes were made to this article in October, 1998.)
Although he has been a notable figure in history and philosophy since he burst on to the intellectual scene in the 1960s with the publication of Folie et déraison (translated in part under the title Madness and civilization), Michel Foucault's importance has steadily grown since his premature death a decade ago. Foucault has long been touted in France as the apparent intellectual successor to Sartre in terms of scholarly influence and popular recognition, but his influence in the English-speaking historical, philosophical, and social-scientific world has now grown to substantial proportions as well. This claim is evinced by the large and ever-growing selection of English books now available that aim to explicate, revise, and criticize his work, and is capped by the recent publication of two major English-language biographies (Macey, 1993; Miller, 1993). His importance to historians of social and behavioral science is twofold: (1) he is the most discussed European historian of psychiatry, human science, criminology, and sexuality this century, and (2) his controversial historiography holds the promise of bearing much fruit yet. This paper is divided into two main parts. In the first I show that Foucault's approach was not, contrary to widespread opinion, particularly Marxist, structuralist, or irrationalist, at least in any straightforward way. In the second, I examine the origins of his early historiography -- the archaeology -- in mainstream French history and philosophy of science. The focus is on the influential works of Gaston Bachelard and Georges Canguilhem. At the end of the paper I briefly comment on the relation between the archaeology and Foucault's later historiographical approach -- the genealogy.
What Foucault Is Not
There is a strong tendency among traditional Anglo-Americans scholars to regard Foucault as a Marxist, a structuralist, or a full-blown Nietzschean irrationalist. Because of the widespread disdain with which each of these positions is held in the English-speaking world, such characterizations of Foucault have contributed to a summary dismissal of his work by a large sector of the English-language intellectual community. Unfortunately, they are misguided views, serving more to clarify the failure of Anglo-Americans to distinguish between various strands of the Continental intellectual tradition, than they do any important aspect of Foucault's work. Moreover, they are views that subordinate Foucault's historical contributions to his philosophical orientations.
Concerning his having been a Marxist, Foucault dropped his membership in the Communist Party (which he joined under the direct influence of leading Marxist historian Louis Althusser) in the 1950s, within two years of joining. Of course, not being a practicing political Marxist does not mean that one's scholarly historical analyses are not Marxist; Marxism has many manifestations.
Still, many commentators -- Marxist and otherwise -- have highlighted a series differences between Foucault's positions and those of mainstream Marxist historians and social commentators. Whatever Marxist tendencies Foucault may have had as a graduate student in the 1950s, a careful comparison of the original 1954 edition of Maladie mentale et personalité and the 1962 edition (translated under the title of Mental illness and psychology) reveals a sharp reduction in Marxist historical analysis, as Gutting (1989) has shown. Moreover, as Foucault distanced himself from Marxism, the Marxists rejected Foucault and his work. As Macey (1993, pp. 175-179) has documented in detail, it was precisely the Marxists who were most savagely critical of Les mots et les choses when it swept to the best-sellers list in France during the summer of 1966. Marxists of almost almost all stripes -- from the hard-liners of the PCF to the Sartrean existentialists -- acused him of being an apologist for the bourgeoisie and neo-capitalism. There can be little doubt as to why Marxists took this view. Foucault own opinion of Marxism was put succinctly in the book: "Marxism exists in nineteenth-century thought as a fish exists in water; that is, it ceases to breathe anywhere else" (cited in Macey, 1993, p. 177).
This provocative comment notwithstanding, however, an examination of Foucault's major works reveals relatively few references to either Marx or to his followers. In an extended comparison of Foucault's work with that of noted Marxist theoretician Antonio Gramsci, Barry Smart (1986) has pointed out that "no sustained discussion of Marxist theory of politics is to be found in Foucault's corpus" (p. 157). Along similar lines, Mark Poster (1989, p. 104) has written that Foucault's is to be regarded as "a response to the crisis of Marixism" (italics added). He continues that "the positions of postmodernism" -- of which he takes Foucault's work to be "the main reference point" -- "are significant to the extent that they present a critique of Marxism and therefore indicate paths that may be taken to get beyond its current theoretical impasses." In particular, Poster believes that Foucault's work evades the "totalizing" and "reductionist" tendencies inherent in Marxist analysis. In less jargonistic terms, whereas Marx and his followers have tried to give an account of everything solely in terms of labor and production, Foucault's critiques are both more regional in nature, and allow a wider range of conceptual bases from which to work; e.g., the analysis of the rise mental hospital need not be grounded in the same beliefs and practices as those on which the analysis of the rise of the human sciences, or of the rise of the prison, are based.
To take an example directly from Foucault's work, one of the most important lessons of the book Discipline and punish (1975/1977), his history of the penal system, is that power is not something that one class simply wields en bloc over another, but that it forms a network of relations that travel, as he says, "downwards," upwards," and "sideways" throughout society. As Arnold Davidson (1986) notes, Foucault's analysis of power, "will lead one to view power not as the homogeneous domination of one group or class over another, but as a net-like, circulating organization" that involves everyone in society (p. 226). Contrary to conventional Marxist theory, he saw power not only in negative terms -- used to repress and oppress -- but for its positive possibilities as well; viz., its role in constituting new objects of knowledge. Though critical of contemporary society, Foucault, unlike the Marxists, prescribed no specific program of action, promised no utopia, and was sometimes as critical of "the proletariat" as he was of "the ruling class" and "the bourgeoisie." Mark Philp (1985) has argued that Foucault "distrusts Marxism" (p. 76) precisely because of its reliance on a system -- a technology -- for the improvement of society. Such systems, according to Foucault, always contain implicit within them the seeds for doing more harm than good, particularly if they become orthodoxies. The first volume of Foucault's History of sexuality (1976/1978), for instance, can be seen as nothing less than a 160-page critique of the standard Marxist analysis of sexual repression, that is widely, but Foucault believed wrongly, thought to have stretched from the Victorian era to the present day.
Claims such as these have led some critics, like J. G. Merquior (1985), to declare him to have been an anarchist, rather than a Marxist. This, too, however, is a crude overstatement of a subtle and complicated situation. Defending Foucault against Merquior's attack, Colin Gordon (1986) has advanced the view that Foucault's work constitutes, in fact, a "Nietzschean challenge to Marxism" (p. 735). This may be the most accurate description of Foucault's relation to Marx, but implicit within it is the beginning of the all-too-common tendency to dismiss Foucault as an irrationalist, which I discuss below. On the other side of the Marxist coin, Frankfurt Critical Theorist Jürgen Habermas (1981) has gone so far as to call Foucault a "new conservative." Little has changed, it seems, since the Marxist criticisms of 1966. Though this is clearly a characterization born of Habermas' tendentious pro-Enlightenment views and Marxist origins in the Frankfurt School, Foucault's libertarian tendencies seem to outweigh in the eyes of many, any Marxist tendencies he may have once been thought to harbor.
It is instructive, if somewhat unconventional, to compare Foucault's position, relative to European Marxists, to Karl Popper's, relative to Anglo-American positivists. Europeans, being disinclined to make fine distinctions among "scientific" philosophers, tended to regard Popper as a positivist, though he was strongly critical of the logical positivist program, offering "falsificationism," or -- as it was sometimes called, tongue-in-cheek, "logical negativism" -- in its stead. Still, Popper's argument with European "sociological" philosophers was widely, if inappropriately, called the "positivist dispute." Similarly, Anglo-American philosophers are disinclined to make fine distinctions among European "sociological" philosophers and, thus, tend to regard Foucault as a Marxist. He was, however, quite cautious of traditional Marxists approaches. As Barry Smart (1986) has put it, "his work may...be seen as a response to, or in effect as a critique of, fundamental elements of both Marxist analysis and socialist political strategy" (p. 157). In any case, it is clear that Foucault was not a Marxist in any direct or straightforward way.
Second, regarding structuralism, there are elements of Foucault's work of the 1960s that are reminiscent of structuralist analysis. Like Marxists, structuralists are an ilk of whom Anglo-American scholars are widely disdainful, and among whom they often do not care to make distinctions. Among psychologists, the best-known example of this attitude is probably to be found in Maher and Maher's (1982) American Psychologist article on the allegation that Foucault asserted the existence of real "ships of fools" -- boats packed full of exiled madmen roving the rivers of 14th century Europe. About their failure to find independent confirmation of Foucault's apparent claims for the existence of ships of fools, they wrote:
Foucault and his followers have an ideological axe to grind: It is important to the structuralist view that social behavior accord with theoretical patterns of symbolism, and it is important to the scapegoat theory of mental illness that it be demonstrated dramatically. The image of the storm-tossed soul cast adrift from rational society fills the bill perfectly. Since real ships of fools did not exist, it was necessary to invent them. (p. 760)
This sort of name-calling is hardly justified, but is common enough. "Structuralism" does not, for Maher and Maher, seem to denote a specific, much less a possibly legitimate, form of inquiry. It simply serves as a term of derision, invoked to denote those whose fantasies have run away with them.
Unfair as this might be to structuralists, it is entirely beside the point, as well. Foucault certainly cannot be plausibly considered to have been a structuralist as early as the 1950s, when Madness and civilization was being written in Sweeden, Poland and Germany. What is more, even his later work of the mid-1960s can only be regarded as having been structuralist in the broadest possible sense -- a sense so broad that it would probably include figures as strikingly different in their approaches as Sigmund Freud and Noam Chomsky. Maurice Henry's famous cartoon, depicting him lecturing Barthes, Lacan, and Lévi-Strauss, all clad in grass skirts having a picnic on the grass, had more to do with the four being in common revolt against the Sartrean humanist orthodoxy of the day, as with them really subscribing to any common methodological doctrine.
Alan Megill (1979) has distinguished four, hierarchically arranged, types of structuralism, and puts Foucault only in the most general category. Moreover, although it might come as a surprise to most Anglo-Americans, it is common among contemporary European scholars to regard structuralism as being akin to positivism, and a failure for the same reasons positivism is. Given Foucault's well-known misgivings about traditional positivist epistemologies, and his explicit criticism of structuralism (along with most other major European approaches of the 20th century) in the closing pages of his book, The order of things, it is surprising that anyone would think him a structuralist in any strong sense.
Whereas the structuralists sought to establish ahistorical abstract structures that apply to a wide variety of situations, Foucault attempted only to describe the relationships within the human-scientific discourse of given historical periods. There was no attempt to "lift" such structures from where they were found and apply them to the discourses of other periods, as Lévi-Strauss attempted to do with cultural forms. Indeed, as with the Marxists, many leading structuralists rejected Foucault's approach. Jean Piaget (1968/1970), for instance, upbraided Foucault for indulging in an incoherent "structuralism without structures." Mistaking Foucault's intent, he criticized the historical knowledge structures called epistemes, described in The order of things, for being "neither necessary nor permanent; they simply follow one another in the course of history" (p. 132). Foucault responded to what he called such "half-witted commentators" in the preface of the English edition of The order of things (1966/1970), stating explicitly that he had used "none of the methods, concepts, or key terms that characterize structural analysis" (p. xiv). Clearly, structuralism was not Foucault's expressed aspiration.
In a 1967 interview, Foucault said, "my archaeology owes more to Nietzschean genealogy that to so-called structuralism" (cited in Mahon, 1992, p. 106). This has lead many to carelessly assume, however, that Foucault adopted all aspects of Nietzsche's philosophy, and bring us to the matter of Foucault's alleged "irrationalism." Irrationalism is a position that is at least as difficult to define as structuralism (and Marxism, for that matter) but, to a first approximation, it entails that the conventional standards of reason, widely thought to apply to all forms of intellectual discourse, are not just to be called into question, but explicitly rejected; to be replaced, it is important to note, not by an improved set of guidelines, but by no definitive standards at all. It is an explicit move to Sophism.
A position very much like this is often attributed to Nietzsche who, after all, saw the Socratic revolution as one of the darkest moments in human intellectual history. It is important to note, however, that many commentators do not see Nietzsche's central position as all this simple. Debate over what Nietzsche "really" said or "really" meant continues to the present day. In fact, it seems to be currently accelerating, if anything. It is interesting that among those who would criticize Foucault for being Nietzschean, almost all adopt a very radical reading of the works of both men. Among those who praise Foucault for being Nietzschean, however, there is a broader latitude of readings.
It is true that Foucault paid unabashed tribute to Nietzsche from the beginning of his career, invoking his writings in Madness and civilization (1961/1965) as an example of the conversation between madness and reason that he believed has been largely suppressed from the time of the Enlightenment to the present day. (Interestingly the very foundation of this position -- viz., that madness is reason's a priori "other," waiting in the wings to be rediscovered -- was soon implicitly rejected by Foucault in favor of a view more akin to social constructivism.) Foucault later explicitly revived Nietzsche's historiographical approach -- the genealogy -- and raised it to new heights of influence and popularity. Genealogy rejects the traditional hermeneutic view that, through historical investigation, one can return to and discover the primordial essence of things. Instead, as Arnold Davidson (1986) has put it, "when genealogy looks to beginnings, it looks to accidents, chance, passion, petty malice, surprises, feverish agitation, unsteady victories, and power" (p. 224). Historical beginnings, thus, are not great moments of revelation that we later must struggle to recapture in their full glory, à la the early Heidegger; they are, on the contrary, messy, corrupt, often ignoble squabbles that must be negotiated, refined, and "smoothed over" for a region of consensus to appear -- for the possibility of a region of knowledge to arise.
In any case, whatever Foucault's allegiance to the genealogical approach might have been, calling upon important aspects of Nietzsche's historiography, on the one hand, and calling for the wholesale overthrow of reason, on the other, are two very different things. Foucault, contrary to widespread belief, never explicitly did the latter. His most radical period was during the early 1970s, when he was first formulating his theory of "power/knowledge." Contrary to the mistaken beliefs of many critics and would-be supporters alike, however, this theory was never -- even in its most radical phase -- the simplistic claim that people in power unilaterally declare what is to be considered "true." Nor even that power and knowledge are the same thing. Power and knowledge were thought to be in a symbiotic relation to each other: "...there is no power relation without the correlative constitution of a field of knowledge, nor any knowledge that does not presuppose and constitute at the same time power relations" (1975/1977, p. 27). Put, perhaps, more plainly, the exercise of power depends on a domain of knowledge for it to operate on (e.g., one cannot set out to exercise power over criminality per se unless and until one is in a position to conceive of such diverse activities as murder, theft, prositution, and military desertion as constituting a unified domain of knowledge). Conversely, one cannot conceive of such activities as constituting a single domain unless one can "discover" or otherwise establish lines of relation drawing them together. But, of course, once this has been accomplished, they are by that very act left open to being dealt with as superficially different manifestations of a single entity.
Foucault is routinely regarded as being closely akin to such radical "textualists" as Baudrillard, Lyotard, and Rorty, all of whom are deeply committed to Nietzsche's rejection of the traditional distinction between philosophy and poetry. This is intellectual company with which Foucault was not always comfortable to be closely associated. In turn, these postmodern thinkers have been critical of Foucault for not being extreme enough -- at least not explicit enough about his alleged extremity (e.g., Baudrillard, 1977/1987; Rorty, 1991; see also Norris, 1994, for an account of some of the main differences). Even the simple label "postmodern," however, was "a term which [Foucault] viewed with some suspicion" (Macey, 1993, p. 462). Near the end of his career -- perhaps frustrated with widespread and facile attempts to assimilate his work to much more radical positions with which he was not in full agreement -- Foucault even allied himself somewhat with the Enlightenment tradition itself, describing his work as a critical extension of Kant's famous call for the use of individual reason in order to further one's freedom from the oppression of the church, state, and other powerful social institutions. In Foucault's own words, "I have been seeking to stress that the thread that may connect us with the Enlightenment is not faithfulness to doctrinal elements, but rather the permanent reactivation of an attitude -- that is, of a philosophical ethos that could be described as a permanent critique of our historical era" (1984, p. 32, italics added). That is, "enlightenment" is not, nor ever was, a matter of dogmatic adherence to one or another specific logical system or scientific method but, rather, a particular attitude or ethos; one that values inquiry and encourages criticism of the status quo. One is tempted to say that, by the end of his career, Foucault realized that he was not so much opposed to the Enlightenment, per se, as to what it had hardened into during the intervening centuries in the hands of people with minds not so broad as, say, Kant's. This is quite a distance from Nietzsche's out-and-out rejection of the entire Western philosophical tradition from Socrates to his own time.
Jürgen Habermas, perhaps the most prominent and explicit defender of Enlightenment thinking on the European philosophical scene, described this apparent shift in Foucault's position as "surprising" (1986, p. 104). Foucault, however, recognizing that Habermas (see, e.g., 1981, 1985/1987a, 1985/1987b) had long taken him to be far more radical than he really was, once joked that he was a little more in agreement with Habermas than Habermas was with him (cited in Ingram, 1994, p. 237). Even Habermas' most important English-language expositor (McCarthy, 1991) has argued that Foucault was closer to Habermas on a host of fundamental issues than he was to any other significant intellectual grouping.
I will not argue any further against claims that Foucault is any strong or direct sense a Marxist, a structuralist, or an irrationalist here -- this has been done often enough by others elsewhere. Given that he is not, however, Foucault seems to be left without an intellectual home; appearing to us as an inexplicable and singular event for which we have no points of reference; a case of intellectual "spontaneous combustion." Without the epithets of Marxism, structuralism, and irrationalism with which he is so often easily dismissed, he seems to appear before us out of thin air, making dark and Delphic pronouncements about madness, medicine, and the human sciences.
What Foucault Is
This is, of course, not true. Foucault was a representative of the tradition in which he was raised as much as any other scholar. The problem is that the tradition out of which Foucault emerged, though widely respected on the Continent, is a virtual unknown to Anglo-American scholars. Interestingly, it is neither purely philosophical nor purely historical, but a tradition of scholarship that combines elements of what we in North America know as the history of science with elements of what we call the philosophy of science. In the Anglo-American tradition, the history of science and the philosophy of science have typically been considered to be two relatively independent projects. The underlying justification for this separation has been that history is descriptive -- it tells us how things were -- and philosophy is normative -- it tells us how things ought to be. Since things are rarely what they ought to be, history was thought to have little of interest to say to philosophy, except perhaps to supply it with past examples of "bad" scientific practice to be held up to critical scrutiny.
It was really only with the publication of Kuhn's Structure of scientific revolutions (1962) that the English-speaking world came to see the tremendous value that history can have for philosophy, and vice versa. In France, however, the mutual connection of history and philosophy has long been recognized, and it is most evident in the work of the French historian-philosophers of science such as Gaston Bachelard and Georges Canguilhem. It was primarily of this tradition that Foucault was a product. The remainder of this paper explores some of the significant parallels between Foucault's work, on the one hand, and that of Bachelard and Canguilhem, on the other.
As early as the 1920s Bachelard was arguing for the importance of the history of science to a proper understanding of its philosophy. Among his most famous claims was that the history of science is littered with epistemological ruptures -- points at which a great deal of what scientists think to be true is quickly abandoned in favor of a new structuring of knowledge within a scientific domain, or even a redistribution of territory among the domains themselves. Between these ruptures were said to be "regions of rationality" (1949) -- periods of relative epistemological coherence and quiescence. Different regions of rationality -- even when concerned with the same objects -- were said to operate, as it were, by different rules; they privilege different scientific approaches and sanction different scientific procedures. If this is so, however, what is to keep science, ultimately, on the right track (viz., toward truth)? Although the idiosyncratic natures of different regions of rationality make them difficult to compare directly, Bachelard remained a strong advocate of the progressiveness of science. Therefore, he said, we are entitled to say that a given region of rationality constitutes an "improvement" on a previous one not by comparing their statements of discovery one-for-one, but rather by looking at the overall territory that is explained by each.
Truth and error, Bachelard argued, are not symmetrical with each other. Truths tend to organize themselves into epistemic structures whereas erroneous beliefs tend only to accumulate without strong connections to the epistemic structure being built. The organization of the set of "fundamental" truths in a given region of rationality strongly influences the way the science of a given time is organized; what methods are sanctioned, what outcomes are most easily accepted into the canon, etc. When the basic truths -- those most strongly entrenched -- change, so does the accepted structure of the science in question. Because of this, Bachelard claimed that "science in effect creates philosophy" (1934/1984, p. 3); that is, epistemology, and even ontology, generally follow scientific practice rather than direct it.
Foucault's famed notion of the episteme -- the "rules" of coherent scientific discourse in a given science at a given time, described in his book The order of things -- can now be seen as a fairly straightforward extension of Bachelard's "region of rationality." What Foucault was able to see clearly of Bachelard could only detect a glimmer -- perhaps because Foucault was studying the human sciences in which it is more apparent -- is that the very objects under study are themselves, to a greater or lesser degree, functions of the discourse of the region, not simply physical "chunks" of the world awaiting our attention. Without objects that naturally fall into scientific "kinds," Bachelard's defense of the progressive nature of science is called into question. Without the same kinds to be studied, we may no longer have a common basis for comparison. Contrary to widespread belief, however, this need not entail a radical sort of relativism. Foucault never categorically denied either that sciences can progress or that objective knowledge is possible. On the contrary, he argued that the natural sciences are much as Bachelard and Canguilhem (discussed below) characterized them.
"We should distinguish carefully," Foucault (1969/1972, p. 183) warned, "between scientific domains and archaeological territories: their articulation and principles of organization are quite different." Chief among the principles that he thought distinguish scientific from archaeological domains is the presence of "certain laws for the construction of propositions" (1969/1972, p. 187). The natural sciences were thought to be among those disciplines that abide by such laws, and have thereby passed the "threshold of scientificity," as he called it. Mathematics was thought by Foucault to have surpassed scientificity, having passed on to the next level above the natural sciences, over what he called the "threshold of formalization." Psychiatry, medicine, and the human sciences, by contrast, though "positive domains of knowledge" (1966/1970, p. 365), were thought by Foucault to have reached only the "threshold of epistemologization." Disciplines at this relatively low level of development still make "claims to validate (even [if] unsuccessfully) norms of verification and coherence" (1969/1972, p. 186), but have not yet become full-blown sciences because the construction of their propositions is not law-governed. It is important to note, however, that this failure does not mean the human sciences are to be rejected out of hand. Foucault specifically stated, "they are not, therefore, merely illusions, pseudo-scientific fantasies motivated at the level of opinions, interests, or beliefs..." (1966/1970, p. 365). The main question for Foucault was not whether the human sciences have epistemic "value" as such -- of course they do -- but, rather, whether they have the capacity, in principle, ultimately to cross over the threshold of scientificity, as the natural sciences did before them.
Foucault believed that disciplines make the transition from the level of epistemologization to the level scientificity by "freeing themselves from the background of social practices that [initially] makes them possible," (Dreyfus & Rabinow, 1983, p. 161). Mathematics, for instance, was said to have transcended its beginnings as a set of measurement techniques used by Ancient Greeks, and the natural sciences transcended their Medieval beginnings in techniques of judicial investigation (Foucault, 1975/1977, p. 225). Presumably other disciplines that currently reside only at the level of epistemologization, can also transcend their beginnings in mundane practices, becoming full-fledged sciences as well. Meteorology, for instance, seems a likely candidate (concerns about weather being a "chaotic" system notwithstanding). The point was, perhaps put most neatly when Foucault (1969/1972, p. 5) approvingly cited Althusser's (1965/1969, p. 168) comment that the birth of a "hard" science comes with a theoretical transformation "which establishes a science by detaching it from the ideology of its past and by revealing this past as ideological." The human sciences, however, seem to entail an additional difficulty that may block such a transition. As Dreyfus and Rabinow (1983, p. 163) put it, "if the human sciences claim to study human activities, then the human sciences, unlike the natural sciences, must take account of those human activities which make possible their own disciplines." That is, transcending their own origins in human practices may be impossible, in principle, for the human sciences.
Thus, the objectivity of science, per se, does not seem to have been a particularly important issue for Foucault. His critique was "regional" in nature, confined primarily to the human sciences, and their applications. To quote Dreyfus and Rabinow (1983) once again,
Foucault at times sounds -- and his critics frequently misread him here -- as if his intention was to situate all science as a mere product of power. This is false. Instead his goal has consistently been to isolate the interconnections of knowledge and power. Throughout his intellectual itinerary, it has been exactly those "pseudosciences" or "near sciences" -- fundamentally the humans sciences -- which he has chosen as his object of study. (p. 177)
Moreover, his motivation was not to repudiate these disciplines but, rather, to investigate their position relative to the modern episteme. Put another way, he called upon us to examine more closely those instances where we often casually assume that progress toward objective knowledge is being made without closely examining the historical content of the concepts underlying our scientific claims.
To return to the main argument, another aspect of Foucault's epistemes that might be counted as an advance over Bachelard's "regions of rationality" is that, whereas Bachelard saw the process as operating over the conscious beliefs of the scientists of a given time, Foucault extricated himself from speculation on the psychology of scientists by having epistemes reside in the discourse of the science itself, rather than in the scientists' heads. The episteme functions more as an external characterization of what is thinkable and comprehensible at a given time, rather than as a specific set of positive beliefs in itself.
Although the impact of Bachelard's work on Foucault was great, the influence of the work of his doctoral supervisor, Georges Canguilhem, was greater still. Whereas Bachelard's concern was almost exclusively with the physical sciences, Canguilhem dealt primarily with biology and medicine. Up to that time -- the mid-1940s -- the history of science tended to focus either on the personal intellectual development of "great" scientists, or on the exploration of the presumed "prehistory" of modern scientific discoveries and disciplines. The former suffers, of course, from the criticisms conventionally directed at Carlylian history: Where did the "great man" -- and it was men that were usually studied -- come from? What forces conspired to produced him? Would he have been as great in another place and time? The latter -- the attempt to trace the "prehistory" of modern science -- on the other hand, suffers from the Whiggish proclivity to read the events of the past through the theoretical and conceptual glasses of the present. It was this latter problem that Canguilhem sought primarily to address and rectify.
Thus, Canguilhem's project was the historical analysis of scientific concepts. Among his best known studies was that of the concept of the reflex arc (1977/1994). Though its discovery is commonly attributed to Descartes, this, in Canguilhem's view, is a mistake. Although Descartes did study the withdrawal of the limb upon noxious stimulation, his explanation of the phenomenon was not one that included the modern concept of the reflex; viz., neural energy being diverted from the sensory pathway to the motor pathway in the spinal cord. On the contrary, Descartes specifically argued that, "the excitation of the senses and the contraction of the muscle are two movements with no analogical relation" (cited in Gutting, 1989, pp. 35-36, italics added).
Common confusions in the history of science, such as that of believing that Descartes developed the modern concept of the reflex, derive from two main sources, according to Canguilhem. The first is, as he said, the "confusion of the identification or description of a phenomenon [which Descartes did] with its proper conceptual interpretation [which Descartes didn't]" (Gutting, 1989, p. 36). The second is that of equating similarity in the actual terms used with that of the underlying concepts they denote.
Foucault's book, Madness and civilization, originally written as a dissertation under Canguilhem's supervision, is, similarly, a history of a concept: "madness." What Foucault aimed to show was that the efforts of traditional histories of psychiatry to link the modern concept of "mental illness" with the various concepts of "madness" as they were understood during the Renaissance, the Enlightenment, and the 19th century, are misguided in exactly the same way as is the effort to locate the origin of the modern concept of "reflex" in the work of Descartes. To do so confuses the description of a phenomenon with its explanation and location in the conceptual network of a period, and is, in part, the result of similar terms being used to express different understandings in different times.
Similarly, in his later book The order of things, Foucault argued that, despite superficial similarities, a term such as "evolution" does not really mean the same thing in Enlightenment natural history as it does in 19th century biology. Thus, conventional efforts to trace "precursors" to Darwin by looking to those natural historians who used the term "evolution" are misguided, according to Foucault, in that they uncritically take terminological similarities, spanning relatively long periods of time, to constitute conceptual similarities. Perhaps ironically, it was the fixist opponent of Lamark's evolutionism, Georges Cuvier, who Foucault believed to be the first person to express an appropriately modern understanding of the concept of "life" to allow for the development of the modern understanding of "evolution." In this connection, it is important to note that when Foucault makes the famous and controversial remark that there was no such thing as "life" until the 19th century, he does not mean, of course, that there were no living things, but rather that "life," as a unified object of knowledge, was not constituted until then -- earlier studies of living things did not constitute recognizable studies of life itself.
To return to Canguilhem's work, his most famous book undertakes an analysis of the relation between the "normal" and the "pathological" in medicine. There Canguilhem surveys the strenuous attempts of 19th century biologists to assimilate the normal and the pathological; to put them on a single continuum. Such attempts usually consisted of the declaration (often unsupported) that the physiological changes underlying the occasioning of illness are merely quantitative changes, not the introduction of qualitatively new phenomena. Diabetes, for instance, is the result of dramatic increases in blood sugar. Canguilhem points out that such claims, whatever other problems they might have, are predicated on a conflation of different levels of medical analysis. Although diabetes may be the result of quantitative changes in the level of sugar in the blood, its experienced symptoms are qualitatively different from what the organism normally experiences.
This insight -- viz., that quantitative changes at one level of analysis can have qualitative effects at another -- is, I believe, crucial to understanding Foucault's belief in the apparently sudden shifts between epistemes in the history of science. Critics often talk of epistemes as though they appear on the scene sui generis, without any cause or adumbration. No reasonably close reading of Foucault's work, however, can support such an interpretation. In The order of things, for instance, he takes great pains to discuss the brief but important transitional periods, dwelling at length on transitional characters such as Lamark in biology, Adam Smith in economics, and William Jones in philology. Shifts in epistemes are perhaps best seen as being analogous to Gestalt shifts in perception: the changes in the underlying material may be relatively smooth and subtle, but at some point the old perceptual organization becomes untenable (the perception "explodes," as Husserl put it), and a new organization imposes itself. Thus, for instance, although the term "labor" had long been a featured part of the analysis of wealth, according to Foucault it was considered but one element among many to have value. When, at the end of the 18th century, the peculiar problems it posed became the focus of so much intense investigation, its position in the "disciplinary matrix" (to borrow a phrase from Kuhn) was changed, and it came to be recognized as the very basis of value itself. In short, a new concept came to be denoted by the term.
To return again to Canguilhem's analysis of the pathological in medicine, it is tempting to regard medical pathology as an objective biological fact -- viz., as the presence of "diseased" tissue. Canguilhem argued, however, that the matter is, in fact, much more complicated. Medical "normalcy" is not a purely objective biological matter. The decision concerning what is pathological and what is merely unusual is ultimately grounded in the organism's subjective experience of its state, rather than in just an objective description of it. This means that merely noting some aspect of the organism's physiology that is statistically unusual does not constitute the identification of a pathology unless it is associated, or ultimately will become associated, with a negative subjective experience. As Canguilhem put it, "pathology implies pathos, the direct and concrete feeling of suffering and impotence, the feeling of life gone wrong" (1943/1991, p. 137) Thus, Canguilhem was led to argue that, biological norms, rather than being a matter strictly for scientists, are a matter to be determined by the organism itself. As he points out, we call the doctor when we determine that we need medical help (1943/1991, p. 226). Of course it is true that the doctor can sometimes identify problems before we feel them, but this is only because past cases have shown that such physical anomalies as the doctor identifies will ultimately come to cause suffering.
It is an obvious, but by no means trivial, step from here -- Canguilhem's analysis of pathology -- to Foucault's study of madness. The difficulty is summed up in a limitation that Canguilhem himself notes: "It is understood that we are not dealing here with mental illness where the patients' ignorance of their state often constitutes an essential aspect of the disease" (p. 295, n. 37, italics added). So if the subjective state is not to be the criterion of mental illness, then what is? The crux of the matter lies, of course, in uncritically regarding madness as, literally, a sort of illness, and it is the development of precisely this contemporary assumption -- as well as those of earlier eras -- that Foucault attempts to discover in Madness and civilization.
Despite this crucial difference, the questions Foucault poses with respect to madness closely parallel Canguilhem's for illness. What have the relative roles of the mind, the body, the society, and the environment been thought to be, historically, in causing madness? Has it been thought to be primarily a matter of defective or perverted reason or will? Foreign elements, or the wrong balance of elements, in the body? Inadequate or harmful physical surroundings? Simple bad company? Put perhaps more directly, the question for Foucault was: what are the relative roles of mind, body, society, and environment not just in causing, but in actually constituting madness (i.e., in making the very concept possible in each age)? This may seem a tendentious question -- I seem to have assumed that it is a social construction a priori. In considering it, however, it is worth noting, for instance, that today, someone unable to "hold down a job" may well find him- or herself on the way to the clinic for a so-called "assessment," whereas the question of "holding down a job" had virtually no meaning in pre-capitalist economies where productivity was not the primary measure of a person's societal worth. Thus, even if one could find a "biological basis" for what might be termed "occupational incapacity," such a condition would only find itself an object of medical knowledge in a society where "occupational capacity" is not just usual, but the positive norm -- in a quasi-medical sense -- for all members.
So the question of what, exactly, the criteria for madness were in earlier times becomes somewhat more urgent. In what relation to the rest of society did the mad stand prior to industrialization? To use Continental terminology, what was madness's "other" -- what was it opposed to -- in other historical periods? Many of the most controversial aspects of Madness and civilization are precisely those points at which Foucault attempts to address this question.
I have said little about Foucault's later works: the history of the prison found in Discipline and punish, and the 3-volume History of sexuality. This is not because I believe these works to be unimportant. On the contrary, they are the crowning achievements of his relatively brief career. There are those who will argue, however, that the moderate characterization of Foucault I have put forward is true only of his early "archaeological" works; that it was only in the "genealogical" approach of his later works that the radical relativism commonly attributed to him comes to the fore. I disagree with this claim on a number of counts. First of all, it seems clear that Foucault was, if anything, attacked by critics even more savagely for his early works than for his later ones with respect to the alleged extremity of his position. For instance, one of Foucault's most vociferous critics, J. G. Merquior (1991), has little but contempt to pour on Foucault's early works, but he praises the second and third volumes of the History of sexuality for their moderation and lucidity. Merquior might have argued, like some others, that these last two works constitute yet another shift in Foucault's approach. Poster (1989), for instance, has said that these works constitute "Position 3" of Foucault's relationship to his work; the most "mature" phase (pp. 58-60). Davidson (1986), has said they signal a shift in emphasis from away from the "genealogy" toward what he calls the "ethics." In any case, the simple division of Foucault's works into "early" and "late" does little justice to their deep interdependencies. What is more, the implication that he simply repudiated his early work when he moved to the genealogy just doesn't stand up to scrutiny, as I argue below. If his expressed position was somewhat more radical when writing Discipline and punish and first volume of the History of sexuality, than in the periods before and after these, this hardly suffices to undermine the rest of his work. As I showed above, when discussing Kant, Foucault's final move seems to have been to achieve something of a rapprochement with the Enlightenment tradition. The search for Foucault's One True Position is, of course, fruitless (not to mention ironic), but the simplistic absolute perspectivalism often attributed to him (by both many critics and some would-be allies alike) fails to adequately characterize his work.
Second, and more importantly, the shift from the archaeology to the genealogy did not, contrary to the beliefs of many, really constitute a reversal or "flip-flop" in Foucault's basic stance. The supposed nature of the alleged turnabout is hard to pin down precisely, but it is sometimes argued that Foucault discovered the errors of his early "totalizing" structuralist ways around 1970, shifting only then to a Nietzschean-style perspectivism. From the first section of this paper, it should be fairly obvious what my response to this characterization will be. On the one hand, the influence of Nietzsche's insights on Foucault's extends back to Madness and civilization (1961/1965) and The order of things (1966/1970). It was hardly new with the works of the 1970s. One the other hand, Foucault was never really a structuralist to begin with, so he cannot have abandoned it for perspectivalism. Davidson (1986) describes the shift not as a replacement of the archaeology by the genealogy but, rather, as the integration of a "second axis" of analysis with the previous one, putting the archaeology "in a wider framework" (p. 227). More recently, Mahon (1992) has argued that Foucault saw the relationship between archaeology and genealogy as that between a method and its goal. As Foucault himself put it, "what I mean by archaeology is a methodological framework for my analysis. What I mean by genealogy is both the reason and the target of analyzing those discourses as events" (cited in Mahon, 1992, p. 105). In any case, it seems clear that the two approaches are not to be simply opposed to each other.
Having now presented and discussed the Foucauldian vocabulary in some detail, I can bring forward a passage by Foucault in support of the claim that the archaeology was integrated with, rather than replaced by, the genealogy, that would have been fairly obscure at the beginning of the paper. In The archaeology of knowledge, Foucault says, contra the frozen abstract edifices of structuralism, that "the episteme is not a motionless figure that appeared one day with the mission of effacing all that preceded it: it is a constantly moving set of articulations, shifts, and coincidences that are established, only to give rise to others" (1969/1972, p. 192). The problem for Foucault, then, was not that of finding something to replace the structuralist faith he is thought to have lost (a view considered and rejected above) but, rather, that of investing the results of archaeological study with the dynamic quality that, from early in his career, Foucault claimed distinguish his epistemes from structuralism's arid and timeless skeletons. The embrace of the notion of "power," central to genealogical analysis, was aimed precisely at solving this dilemma. Power was thought by Foucault to animate the "discursive formations" that were the products of his archaeological investigations, as well as allowing his purview to extend beyond -- or, more precisely, to return from -- the increasingly textual realms of his work of the mid-1960s, to realms of action and social systems of behavior (e.g., prison systems, sexuality). Summing the matter up, the genealogy did not replace the archaeology but, rather, was integrated with it to fill a lacuna in its descriptive horizon. This may not be a majority opinion but, as noted above, neither is it an interpretation idiosyncratic to myself alone. Similar views are held by several other moderate expositors of Foucault's work, such as Dreyfus and Rabinow (1983), Davidson (1986), Gutting (1989), and Mahon (1992). Even Foucault (1983) himself sought to disabuse those who thought him a radical relativist of their enthusiastic error. Discussing his own analyses of historical struggles against institutionalized power, he wrote:
There is nothing "scientistic" in this (that is, a dogmatic belief in the value of scientific knowledge), but neither is it a skeptical or relativistic refusal of all verified truth [italics added]. What is questioned is the way in which knowledge circulates and functions, its relations to power. In short, the régime du savoir. (p. 212)
Before closing, I think it is important to highlight the fact that I have not offered any evaluation of Foucault's works. There is much that is controversial in his analyses of madness, medicine, human science, the penal system, and sexuality. Before one can engage in fruitful evaluation of those works, however, it is important to have a clear understanding of what they actually claim. The point of this paper was to reveal the intellectual ground in which Foucault's ideas germinated and took root, not to provide a summary and review of his work. That ground is not, as many are inclined to believe, particularly Marxist, structuralist, or irrationalist. Foucault was, rather, an historian-philosopher of a sort that is central to the French intellectual tradition.
Althusser, L. (1969). For Marx (???, Trans.). London: Pantheon. (Original work published 1965)
Bachelard, G. (1949). Le rationalisme appliqué. Paris: Presses Universitaires de France.
Bachelard, G. (1984). The new scientific spirit (A. Goldhammer, Trans.). Boston: Beacon. (Original work published 1934).
Baudrillard, J. (1987). Forget Foucault (N. Dufresne, Trans.). New York: Semiotext(e). (Original work published 1977)
Canguilhem, G. (1994). The concept of the reflex (A. Goldhammer, Trans.). In (F. Delaporte (Ed.), A vital rationalist: Selected writings from Georges Canguilhem (pp. 179-202). New York: Zone. (Original work published 1977)
Canguilhem, G. (1991). The normal and the pathological (C. R. Fawcett & R. S. Cohen, Trans.). New York: Zone. (Original work published 1943, 1966)
Davidson, A. I. (1986). Archaeology, genealogy, ethics. In D. C. Hoy (Ed.), Foucault: A critical reader (pp. 221-233). London: Basil Blackwell.
Dreyfus, H. L. & Rabinow, P. (1983). Michel Foucault: Beyond structuralism and hermeneutics (2nd ed.). Chicago: University of Chicago Press.
Foucault, M. (1965). Madness and civilization: A history of insanity in the age of reason (R. Howard, Trans.). New York: Random House. (Original work published 1961)
Foucault, M. (1970). The order of things: An archaeology of the human sciences (A. Sheridan, Trans.) New York: Random House. (Original work published 1966)
Foucault, M. (1972). The archaeology of knowledge & The discourse on language (A. M. Sheridan-Smith, Trans.). New York: Pantheon. (Original works published 1969, 1971).
Foucault, M. (1973). The birth of the clinic: An archaeology of medical perception (A. M. Sheridan Smith, Trans.). New York: Random House. (Original work published 1963).
Foucault, M. (1976). Mental illness and psychology (A. Sheridan, Trans.). New York: Harper & Row. (Original work published 1954, 1962)
Foucault, M. (1977). Discipline and punish: The birth of the prison (A. Sheridan, Trans.). New York: Vintage. (Original work published 1975)
Foucault, M. (1978). The history of sexuality: An introduction (R. Hurley, Trans.). New York: Random House. (Original work published 1976)
Foucault, M. (1983). The subject and power. In H. L. Dreyfus & P. Rabinow, Michel Foucault: Beyond structuralism and hermeneutics (2nd ed., pp. 208-226). Chicago: University of Chicago Press.
Foucault, M. (1984). What is enlightenment? (C. Porter, Trans.). In P. Rabinow (Ed.), The Foucault reader (pp. 32-50). New York: Pantheon.
Gordon, C. (July 4,1986). Reassessing Foucault. Times Literary Supplement. p. 735.
Gutting, G. (1989). Michel Foucault's archaeology of scientific reason. Cambridge: Cambridge University Press.
Habermas, J. (1981). Modernity versus postmodernity. New German Critique (no. 22).
Habermas, J. (1986). Taking aim at the heart of the present. In D. C. Hoy (Ed.), Foucualt: A critical reader (pp. 103-108).
Habermas, J. (1987a). The critique of reason as an unmasking of the human sciences: Michel Foucault (F. Lawrence, Trans.). In The philosophical discouse of modernity: Twelve lectures (pp. 238-265). Cambridge, MA: MIT Press. (Original work published 1985)
Habermas, J. (1987b). Some questions concerning the theory of power: Foucault again (F. Lawrence, Trans.). In The philosophical discouse of modernity: Twelve lectures (pp. 266-293). Cambridge, MA: MIT Press. (Original work published 1985)
Ingram, D. (1994). Foucault and Habermas on the subject of reason. In G. Gutting (Ed.), The Cambridge companion to Foucault (pp. 215-261). Cambridge, Cambridge University Press.
Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press.
Maher, W. B. & Maher, B. (1982). The ship of fools: Stultifera navis or ignis fatuus? American Psychologist, 37, 756-761.
Mahon, M. (1992). Foucault's Nietzschean genealogy: Truth, power, and the subject. Albany, NY: State University of New York Press.
Macey, D. (1993). The lives of Michel Foucault. London: Vintage.
McCarthy, T. (1991). The critique of impure reason: Foucault and the Frankfurt School. In Ideals and illusions: On reconstruction and deconstruction in contemporary critical theory (pp. 43-75). Cambridge, MA: MIT Press.
Megill, A. (1979). Foucault, structuralism, and the ends of history. Journal of Modern History, 51, 451-503.
Merquior, J. G. (1985). Foucault. London: Fontana.
Miller, J. (1993). The passion of Michel Foucault. New York: Doubleday.
Norris, C. (1994). What is enlightenment? In G. Gutting (Ed.), The Cambridge companion to Foucault (pp. 159-196). Cambridge: Cambridge University Press.
Philp, M. (1985). Michel Foucault. In Q. Skinner (Ed.), The return of grand theory in the human sciences (pp. 65-81). Cambridge: Cambridge University Press.
Piaget, J. (1970). Structuralism (C. Maschler, Trans.). New York: Basic. (Original work published 1968).
Poster, M. (1989). Critical theory and poststructuralism: In search of a context. Ithaca, NY: Cornell University Press.
Rorty, R. (1991). Moral identity and private autonomy: The case of Foucault. In Essays on Heidegger and others (pp. ???-???). Cambridge: Cambridge University Press.
Smart, B. (1986). The politics of truth and the problem hegemony. In D. C. Hoy (Ed.), Foucault: A critical reader (pp. 157-173). London: Basil Blackwell.
This research was conducted with the assistance of a York University Faculty of Arts Research Grant, a York University President's Natural Sciences and Engineering Research Council Research Grant, and a federal Natural Sciences and Engineering Research Council Individual Research Grant. Reprints of this article can be obtained from the author at the Department of Psychology, York University, North York, Ontario, M3J 1P3, CANADA.