RESPONSIBLE BELIEVING

S. Joel Garver

CHAPTER FIVE

The Power of "Truth":
Foucault and Doxastic Practices

Section II: Part B

Foucault's Descriptive Account

Let's begin by taking a look at Foucault's notion of a "discursive formation." A discursive formation for Foucault is not a simple parallel of what a language game is for Wittgenstein or a paradigm for Kuhn and, though similar concerns are present in those writers, we will do well to set them aside for the time being. The closest parallel to Foucault's notion and analysis of discursive formations probably stems from the structuralist linguistics of Saussure and, perhaps, those who have various ly developed "structuralist" thinking (Hjelmslev, Levi-Strauss, Barthes, Canguilhem, etc.). Nevertheless, Foucault's thought breaks beyond even the old wineskins of structuralism (cf. Dreyfus and Rabinow 1983). We do well, then, to make a fresh start in turning to Foucault.

The idea of a discursive formation "in general denotes the conditions of possibility for webs of belief" (Alcoff 1993:99). We are not speaking here of broadly logical possibility, but rather of the actual possibility at a given time and within a particular social context of certain beliefs and "webs" of belief being generated, having semantic meaning, being epistemically understandable, being capable of undergoing challenge and defense, and counting as true, as justified, or as knowledge. This possibility, then, is better thought of as a possibility conditioned by various social, epistemic, cultural, political, and linguistic facts and practices. Foucault's analysis of these facts--what they are and how they function--is his theory of the discursive formation.

The basic point of Foucault's analysis is that these conditions of possibility are to be accounted for (wholly or for the most part) by "the relations between statements" that are immanent to actual discourse at a given time (Foucault 1972:31). The conditions of possibility, then, are not made determinate by the intrinsic meanings of statements, by authorial intentions, or by reference to the non-discursive world, but by "the interrelationships which exist between discursive entities" (Alcoff 1993:99). As Foucault writes, "the regularities [i.e., conditions of possibility] of statements is defined by the discursive formation itself. The fact of [a statement] belonging to a discursive formation and the laws that govern it are one and the same thing" (1972:116). In order to better understand this approach it will be helpful to examine several aspects of discursive formations.

Let us begin with the role of discursive formations in producing meaning. According to Foucault the meaning of a discursive entity is produced by its position within the particular discourse in which it functions and what it does within that discourse (since discursive entities involve "events and functional segments" in a system; 1975:xvii). He writes, "The meaning of a statement would be defined not by the treasure of intentions it might contain...but by the difference that articulates it upon the other real or possible statements, which are contemporary to it or to which it is opposed in the linear series of time" (1975:xvii).

We might think of this by analogy with a chess game. The "meaning" of a chess game is not simply a matter of where each piece is or what the players intend to do with them. Rather the "meaning" of the chess game consists in a dynamic relationship among the pieces, defined by what sorts of moves are permitted, how the possible moves of each piece differ from those of another, possible responses that are shaped by previous moves, and possible consequences of these factors.

In terms of a language, the meanings of words, and the meanings of larger units, we can most easily see how such an approach might function when considering words like "poor," "good," and "average." If we place these terms within a certain discursive context--for example, a teacher grading students--we can see what the words mean by how each is used differently within the context. We can also see how the meaning of "good" would change if we were to add "excellent" and "outstanding" to the grading system. It is the different functions of the words in this example, then, that determines their meanings (the example is drawn from Silva 1990:46). Foucault wants to extend this sort of analysis to cover not only all linguistic meaning, but further to analyze the possibility for truth, exclusion, and understanding.

In order for a statement to have a truth-value within a certain discursive formation, the statement must be somehow connected to other elements within that formation. If there is no connection, the statement can have no meaning. Thus a discursive formation will draw boundaries around what may be meaningfully expressed, asked, defended, challenged, disproved, and understood within its purview. Likewise, what sorts of beliefs will be formed by subjects within the discursive formation is delimited. Of course, there are different sorts of meaningfulness.

By "meaning" Foucault includes a continuum between what we might think of as semantic meaning on one end and epistemic meaning on the other (or what we might, more subjectively, think of as "significance"; cf. Ross 1980:115-118). As Alcoff writes "For Foucault, statements cannot be adequately analyzed as bearers of propositional content alone" (1993:102). Semantic meaning, as I am thinking of it, has to do with meanings of words as such and basic syntactical relations and transformation s of those words; both are defined, says Foucault, by the rules of the discursive formation in which they are set. Nevertheless, there is far more to Foucault's idea of meaningfulness than that.

Consider Alcoff's example, the statement "Mars is angry" (1993:101). While this statement may have some semantic meaning for us, its meaning is limited to that. We lack epistemic understanding of it: our present discursive formation makes it difficult for us to paraphrase it, to recognize its entailments, to produce evidence for its truth or falsity, or to list its truth conditions. The statement, therefore, lacks the right kinds of connections to the present discursive formation in order for it to be regulated in the ways necessary for it to have epistemic meaning or significance. Foucault's account of meaning, therefore, extends to these considerations as well.

But what sorts of relations do give rise to the kinds of meaningfulness Foucault envisions? Foucault considers these issues in terms of what he calls an "enunciative function" which

Thus robust meaningfulness is not a matter of mere grammar, logical consistency, or psychology, but rather is the product of a wide field of regulative functions. These four areas of the enunciative function (a referential, a subject, an associated field, a materiality) parallel the "four directions in which [the discursive formation] is analyzed (formation of objects, formations of the subjective positions, formations of concepts, formation of strategic choices)" (1972:116).

The differences between these two four-fold analyses are not important for our purposes. It will suffice to think of these two four-fold analyses as covering the same ground from different perspectives (the analysis of discursive formations focuses on the conditions for the existence of a discourse; the analysis of the enunciative function focuses on the conditions for the existence of regulated groups of signs or statements; 1973:38-9, 86- 88). Grouping the two sets together, then, let us examine two of these four areas in some detail: the objects/referentials and the subjects/subjective positions. These two areas have played central roles in western philosophical thought and so Foucault's conceptions may prove of the most interest in respect to them.

The first area concerns the existence of objects which are "referred" to by discourse, the "referentials." For Foucault, the objects of discourse are not things-in-themselves, utterly mind independent, existing wholly beyond or apart from discourse to which discourse refers. Whether this is true for Foucault in respect to all objects of discourse, or whether it is true in the same ways or to the same degree, is an open question. It is Foucault's view concerning at least and especially those objects in which he shows the most interest: social objects such as madness, medical diagnoses, rationality, the self as a social and ethical agent, sexuality and gender roles, and so on. Objects of discourse, Foucault says, are formed and regulated by several factors.

There are "surfaces of emergence" which "show where these individual differences [the individuation of objects of discourse]...may emerge, and then be designated and analyzed" (1973:41). Thus, something like mental illness does not exist as something simply referred to by discourse, but is "constituted by all that was said in all the statements that named it, divided it up, described it, explained it...and possibly gave it speech by articulating, in its name, discourses that were to be taken as its own" (1973:32).

There are also "authorities of delimitation" which are those to whom society has given authority to allow or disallow certain objects as belonging to a discursive formation. For instance, "medicine...became the major authority in society that delimited, designated, named, and established madness as an object" (1973:42). Such authority will not often exist in isolation, but will be joined with that of legality, religion, and criticism.

There are "grids of specification" which "are systems according to which the different [for example] 'kinds of madness' are divided, contrasted, related, regrouped, classified, derived from one another as objects of psychiatric discourse" (1973: 42). Thus grids provide systems of classification and relation.

Furthermore, these areas are not exhaustive: there is a place for nondiscursive factors--primarily practices--and their relations in the formation of objects and the discursive elements. What Foucault finally concludes is that we should "no longer...[treat] discourses as groups of signs...but as practices that systematically form the objects of which they speak" (1973:49). Statements achieve "reference" to these discursively constituted objects by means of "laws of possibility, rules of existence for the objects named, designated, or described within it, and for the relations that are affirmed or denied in it" (1973:91).

How are we to understand Foucault regarding the objects of discourse? Let's begin by clearing away what might be a certain misunderstanding. Foucault is not saying, as far as I can see, that reality is wholly "socially constructed" or a product of discourse. This is not idealism. Rather, Foucault is saying that we lack extra-discursive access to the world, but not in the sense that we necessarily lack real access to the world. Such access is mediated through human practices and emerges as a world of objects within discourse.

I think this is clear from Foucault's discussion of "the body" within discourse and, thus, within power relations. Foucault's writings concerning the human body are one the few places where we encounter him theorizing concerning an object in a way that portrays the object as more metaphysically basic or real (as we would traditionally think of that). Nevertheless, even there he describes the body as "the inscribed surface of events" (1977:148) and takes it to be his task to expose a body totally imprinted by history and the processes of history's destruction of the body. The body exists within discourse and within practices and there power relations "invest it, mark it, train it, torture it, force it to carry out tasks, to perform ceremonies, to emit signs" (1979:25). Such statements certainly take objects such as bodies as real and substantial, but they do not countenance these objects as ever outside of discourse or practices (or power, but that topic will come later). Rather, bodies are always "under a description" and not "the way they are in themselves."

Discourse, then, along with the practices that are tied to it, is the house of being. How we understand the details of Foucault's view is of less importance than to grasp the emphasis on discourse. After all, Foucault's views may be compatible with and shed light on any number of other views (e.g., they may correspond to Putnam's internal realism; they may provide a context in which to understand and relativize Kripke's notion of a rigid designator and how "initial baptisms" work; they may, despite all their "social constructionism," be compatible in certain limited areas and by some revision with the Thomistic and Scotistic notion that we can abstract real natures from experiences of particular natural objects).

Furthermore, most of what Foucault writes focuses on socially embedded objects of knowledge such as gender roles, madness, political concepts, criminality, and so on. Such objects are reasonably taken to supervene upon or emerge from not only the world as it is independent of the human intellect, but also from the sum of human discourse, practices, institutions, relations, beliefs, and so on, at a given time within a given context. We also take all sorts of statements about such objects to be true and false. Thus, it seems to me, that any philosopher is foolish to make "conformity between statements (beliefs, propositions, or what have you) and mind independent objects or reality" the measure of all truth.

Statements concerning intentional objects or within intentional contexts (e.g., "John believes the fish has gone bad," "The present construction of femininity involves disciplinary technologies that restrict women's freedom") are true or false, Foucault would say, in virtue of the shape of the present discursive formation. Whether or not it is true that John believes the fish to have gone bad depends upon whether or not the present discursive formation is such that what the statement says can be truly said to be the case (e.g., whether the present discursive formation authorizes us to take an intentional stance that attributes such a belief to John). Similarly, whether or not what was stated regarding the present construction of femininity is true also depends upon the present discursive formation (e.g., whether the beliefs, practices, subjective positions, power relations, and other aspects of the present discursive formation really do construct a form a femininity such that the properties in question are authorized to be attributed to it within the discursive formation).

These sorts of considerations, while in need of clarification and refinement, do not pose, I think, any threat to our common sense notion of what it is for a statement to be true (whether that's as simple as "what the statement purports to be, actually being the case" or whether we have a richer notion of correspondence). What is the case (or the one member of the correspondence relation) is, at least sometimes, dependent upon intentional features of the world. Perhaps, due to discourse's role as the "house of being," all statements essentially involve intentionality. In any case, ordinary notions of truth are reasonably taken to be safe.

Moreover, even if the objects of discourse are constituted within discourse, Foucault's suggestions are modest. He notes that he is not saying that every change in a discursive formation entails an entirely new domain of objects. Foucault writes:

And, as Alcoff adds, that "leaves open the possibility that most common sense beliefs and simple perceptual beliefs may remain constant through discursive changes" (1993:110). That constancy, moreover, might be best accounted for by causal connections to the world itself. For our purposes, however, more important than the metaphysics is the epistemology that Foucault suggests. In this respect, Foucault's views supplement Millar (how this is so I shall outline later).

Let us now move to Foucault's discussion of the knowing and believing subject and the epistemic role of subjective positions. Part of what makes for the coherence of a discursive formation, Foucault suggests, is the relations between certain groups of statements (along with the objects formed by those statements) and the "subjects" from which those statements come.

Statements achieve meaningfulness, significance, acceptance, and are taken to possess a positive epistemic status within a discourse due, in part, to the subject(s) from whom the statements have come. Various authorities, Foucault writes, are sanctioned (by legality, tradition, or otherwise) to produce particular bodies of knowledge and the extent and functioning of one authority is defined by its relations to and differences from the extent and functioning of other authorities. Certain relations also exist between authorities and the rest of society and they differ in respect to various relations (e.g., the shape of a doctor's authority depends upon whether he is privately consulted, intervening on behalf of an incapacitated patient, paid by the state, forms of contract, etc.). (1972:50-51)

There are also important institutions associated with authorities. For example, medical doctors may be variously associated with hospitals, private practices, laboratories, or the academy. Different institutions authorize different sorts of knowledge, information gathering, levels of trust, and the like. (1972:51-52)

There are important relations that a subject can possess in respect to a particular realm of objects. These relations by which "knowledge" is (authorized to be) procured may include forms of questioning or interrogation, kinds of listening and seeing, and the use of instruments. Thus the psychiatrist may ask certain questions the average agent may not under normal conditions; the doctor can hear and interpret noises the body makes in ways others cannot; the artist can see the underlying geometric structures of a still life which are invisible to the casual observer; and the physicist can use various detectors to track and measure subatomic entities. (1972:52-53)

It is also important to note that in discussing the subject of a statement Foucault is not primarily thinking about the author who emitted or produced a particular series of signs. Rather it is "a particular, vacant place that may in fact be filled by different individuals" and it is a place that is "variable enough to be able either to persevere, unchanging, through several sentences, or to alter with each one." Discursive entities count as statements not because they have an author, but because we can determine "what position can and must be occupied by any individual if he is to be the subject" of that particular statement. (1972:95-96)

So much for subjects. Let us briefly consider the final two items in Foucault's analysis. First, there is the formation of concepts and an associated field. Sets of concepts do not necessarily form fully coherent wholes, obeying rigorous deductive norms. Nor do they come to be at once, full and complete or constructed piece-by-piece like steel bridges. Thus describing the formation of concepts must not presuppose those notions. Rather, Foucault's description focuses on an analysis of the organization of actually existent statements as they appear and circulate. This organization, as Foucault describes it, involves the following (from 1972:56-59):

It is these elements in the formation of concepts that gives content to what is said (believed, known, etc.) and thus the propositional content of a particular statement P is integrally related to other statements that coexist with P, are referred to by P, by which P is made possible, which P itself makes possible or gives occasion for, and which stand and fall with P in terms of P's status as true, important, valuable, sacred, prophetic, and so on. (19 72:98-99)

Now for the final item in Foucault's analysis of discursive formations: strategic choices and a materiality. By this Foucault includes how the various discursive elements we have described come together in (more or less) coherent wholes as opposed to some other way they might have come together and, further, what various possible alternatives for future theory formation and choice are created. He also considers the role of a particular discursive formation in the overall discursive economy at a given time (e.g., does it serve as a concrete model for more abstract discourses, does it complement or stand analogously to other discourses, and so on; 1972:64-66).

The role of non-discursive practices is also of great importance for Foucault here: particularly, rules and processes by which a particular discourse is appropriated and to what various ends (1972:64-68). Thus the statement "Species evolve" has changed from before to after Darwin, not in the meaning of its words, but in its role in relation to other statements--in terms of what is required for verification, what further research is possible, what problems it is thought to solve, its role in stabilizing discourse, its use politically (e.g., social Darwinism), socially (e.g., the proliferation of birth technologies), theologically (e.g., the thought of Teilhard de Chardin), or otherwise, and what positions desire may take in relation to it. (1972:100-1 04)

It seems, then, that for Foucault belief-formation essentially involves the "conceptual-reading practices" that Millar has already outlined in terms of forms of inference and quasi-inference. But Foucault goes far beyond Millar in several respects, employing a wider range of elements in his epistemology and more heterogeneous ones at that. First, Foucault develops the notion of a concept. For him a concept is not merely a functional entity with associated patterns of inference and quasi-inference that allows subjects to move from certain inputs (propositions, experiences, and so on) to certain outputs (beliefs, conclusions, and so on). Concepts are integrally placed within a larger system in which they are made determinate by what they do and how they can function differently from other elements within that system in terms of significance, value, coexistence with other concepts, relationships with authorities, and so on. This in turn helps us see how broad mastery of a concept includes much more than just patterns of inference. It also helps explain how the concepts involved in, for example, trust and testimony, are able to be utilized so discriminately.

Second, Foucault develops the relation between concepts and the world of objects. For an entity to be an object of knowledge or of belief, that object must be constituted as such by the discursive concepts that are employed. Believing a part of the world to be water (or, if one is a Hopi, pahi or keyi), is, for Foucault, a complex interaction of factors: a subject must move doxastically within a certain discursive formation in such a way that the subject is authorized to appropriate a certain concept which then functions (is applied, used, connected to other concepts, and so on) in a particular way within a particular practice (e.g., the practice of believing "That's water" when having particular sorts of experiences--a form of "quasi-inference"). Depending on the discursive formation an instance of this (of the concept of water functioning in a certain way) may or may not be mistaken. What counts as water within that discursive formation, at least, is constituted by which ways in which the concept can be used are authorized as correct.

Foucault never fully analyzes the relation between the world and the statements within a discursive formation which are taken to be true descriptions of the world. He certainly allows room for error or falsehood in respect to those statements which are taken to be true and believes such error to be detectable by means of immanent critique and other methods (e.g., massive and persistent inconsistencies between doxastic outputs can indicate, for Foucault, a discursive system that produces falsehood; cf. his methodology in Discipline and Punish 1979 and his critiques of abusive power 1980). These points primarily concern criteria for the conditions under which a certain proposition is to be taken as true or false (and shown as false or proven true and so on). Foucault, however, never analyzes what truth is as such or what it is for the world to be such that it can sustain the truth of the outputs of a particular discursive formation.

Third, Foucault develops the relation between concepts and the subjects of belief. For Foucault the apparent epistemic status of a certain belief or the demonstrable veracity of that belief depends in part upon the authorized competence of the believing subject in utilizing the concepts and practices required for the formation of the belief. That is why, on his view, various belief-forming practices must be taught by properly trained authorities (e.g., teachers who instill "appreciation" and "good taste" in their pupils). It is also why forms of trust are extended to certain classes of people differently than to others (e.g., priests and doctors as opposed to strangers, store clerks, friends, accountants, politicians, or used car sales men). Again, it is why various institutions and practices are formed and maintained (e.g., academic journals, review boards, experimental documentation and reproduction, the Hippocratic oath, the seal of confession, tenure, quality control, merit-based promotion, and so on).

Finally, Foucault underscores the goal-directed nature of belief. Discursive formations--including the objects they "create," the subject positions they authorize, the concepts they construct, and the belief-practices they engender--are formed to achieve particular ends, fulfill certain desires, meet specific needs, and, as we shall see, channel power into unique forms. It is these considerations that determine what questions are asked, which ones make sense, what kinds of answers are satisfactory, what knowledge is important, and which beliefs can possibly be taken to have justification.

That is among the reasons why the question posed in French criminal court shifted from the 17th century to the present day: from "Who committed it?" to "How can we assign the causal process that produced it? Where did it originate in the author himself? Instinct, unconscious, environment, heredity?" (1979:19). As Alcoff points out, "in seventeenth century France the statement 'The accused committed the deed but is not guilty' would not have had a truth-value whereas today it does" (1993:106-7). The goals of criminal investigation, prosecution, defense, and punishment have shifted; that is to say, the discursive formation (in respect to criminality and guilt) has developed (though development, Foucault will insist, should not be taken to entail positive progress).

Thus the elements that make up a coherent conceptual scheme or discursive formation are, for Foucault, more in number and diversity than usually suggested. The picture he paints is also, I think, more realistic because of its heterogeneity. Despite diversity, some degree of logical consistency is supposed by Foucault, at least normatively speaking, within a discursive formation. And though Foucault's notion of coherence falls far short of mutual entailment (a lofty goal of some theorists), he does sketch the number of forms of mutual support already outlined. How consistency and these forms of mutual support play into the more normative concerns of justification, truth, and knowledge, will concern us further below. Foucault's account also helps us see how we might go about exploring what it means to properly engage in a particular doxastic practice, at least descriptively, in terms of what is socially established. Let us now turn from Foucault's analysis of discursive formations to his discussion of power.