Now not only are these contemporary usages quite different each from the other, they are also all different from the category of 'discursive formation' which is developed by Foucault. Foucault's concern is not to produce a general theory of discourse (whatever that might mean). His use of the term discourse may be taken to be tactical. It may be thought of as an attempt to avoid treating knowledge in terms of 'ideas'. The reason for avoiding the term 'ideas' is that it brings in its train a series of presuppositions which Foucault hopes to abandon. We will mention only three. The first is that an 'idea' is knowledge by virtue of being a proposition, a proposition being the logical form of an idea. Knowledge viewed in this logical sense may be thought of as a tissue of 'ideas'. Knowledge consists of ideas as they present themselves for validation. The second pre-supposition is that an idea' is a mental representation and is thus tied to the apparatus of production of thought by a human subject. Although these two presuppositions do not have to go together with any logical necessity, they frequently do so in historical investigations, especially in the sense of ideas being treated as propositions and at the same time having an 'author'. The third pre-supposition is that 'ideas' are expressed or have their existence in language. In this case the identity of an idea is its meaning and its basic units are sentences. As we shall see this trinity of proposition-subject-meaning which hovers over the idea is one from which Foucault tries to turn away in his analysis of knowledge. The use of the term 6discursive' should be taken as no more (but no less) than an index of this attempt. It may be, in fact, that Foucault does not fully succeed in this. But that should not undermine the strength of his arguments for such a move (Brown and Cousins, 1980). His analysis of what makes up knowledge is not reducible to propositions which appear in meaningful sentences and which have been produced by subjects. This attempt to skirt the category of 'ideas' in the analysis of knowledge requires Foucault to question a whole battery of other terms which are at work within the history of ideas. The argument of The Archaeology of Knowledge opens with the negative work of discarding (or at least suspending) conventional categories. The common denominator of the categories against which he argues is that they all unify disparate elements too easily or too early. They avoid the specification of differences through a facile synthesis. The first of such categories is that of tradition. In the history of ideas it is all too easy to simplify the problem of successive phenomena through the levelling agency of tradition. Ideas are given a life-span by persisting, by being continuously accepted, a life which is summarised as a 'tradition'. Foucault proposes to suspend this category of tradition not because the problems of the transmission and communication of knowledge are not important. It is that they are too important to be reduced to the undifferentiated category of tradition. The conditions of appearance and reappearance of forms of knowledge must be identified by reference to specific means. By the same token any reference to a 3pirit of the age' must be eschewed since such references establish links between phenomena through the dogmatic axiom that whatever is contemporaneous is necessarily related. Foucault rejects this axiom and insists that links which are made be established non-deductively. Lest it be thought that Foucault is engaging in an eccentric francophone purism it is worth noting that his argument exactly matches that of Gombrich in 'In Search of Cultural History' (in Ideals and Idols).
Nor is Foucault alone in rejecting easy syntheses of what we may call genres of ideas. The way in which we spontaneously distribute and name discourses as science, literature, philosophy or politics is a division of discourses of recent origin, and to apply them to medieval culture or antiquity is to risk the retrospective projection which is usually called anachronism. These suspensions of synthetic categories move in harmony with most contemporary work in the history of ideas. However Foucault pursues these suspensions with unusua1 rigour. In particular he argues for the suspension of the categories the oeuvre and the book. Neither, he argues, are unquestionable givers; they are themselves constructions in discourse. The category of an oeuvre serves to support the idea that it is the collected texts designated by a proper name, yet it does not contain everything written by the bearer of that proper name. The status of Nietzsche's laundry-lists has now spawned a considerable theoretical literature. For Foucault what is at stake in the category of the oeuvre is that it supports and protects the category of an author. If that category is accepted uncritically it introduces into the analysis of discourse the notion of an oeuvre as the expression of a human subject. In suspending the category of oeuvre and author, Foucault is not, as some critics allege, abolishing the categories but insisting that authorship is an historical and variable construction and thus cannot be uncritically used in analyses in its post-Romantic sense; likewise the term book. If we think of this as unproblematic we need only to consider medieval literate culture. There is no sense in which the medieval compendium corresponds to what we automatically consider as a book (What is an author, LGMP: 113-38). The argument of The Archaeology of Knowledge then opens with a series of suspensions of categories - ideas, tradition, period, oeuvre, author, book. Their suspension is undertaken in order to force analysis from the grip of the categories which work to unify the history of knowledge in terms of the human subject, consciousness and the march of reason.
The generic name for knowledge that is (nothing but) language is discourse. Discourse expresses, and is, the inherently transgressive quality of poststrucruralist intellectual politics, as one can see in Hayden White's definition:
A discourse moves "to and fro" between received encodations of experience and the clutter of phenomena which refuses incorporation into conventionalised notions of "reality," "truth," or "possibility."...Discourse, in a word, is quintessentially a mediatise enterprise. As such it is both interpretive and preinterpretive; it is always about the nature of interpretation itself as it is about the subject matter which is the manifest occasion of its own elaboration.
A postmodern social theory, whether avowedly sociological or not, is discursive in this sense of transgressing the subject-matter it interprets by constantly reflecting on the necessary and nature of interpretation itself.
Of course, there are problems with a proposal to make discourse both the subject-matter and the medium of sociological analysis. A discursive sociology would require an uprooting of deeply ingrained convictions -- belief in the subject-object dichotomy and the other classical dualities; loyalty to the ideal of sociology as a well-founded, scientific source of knowledge; expectations mat good work will produce identifiably worth- while political and intellectual outcomes.
But the far more serious problem with a discursive sociology in the post-structuralist or postmodern tradition is that posed by taking discourse as an object of study. It is one thing to accept a discursive, transgressive method as the condition of sociological practice, another to deal with evident dilemmas in the discursive analysis of discourse. Sociologists and other intellectual practitioners can be discursive in the sense of appropriating the attitude of constant, as White puts it, to-ing and fro-ing with the real world. Social theory as reflective, intransitive action is thinkable even if objectionable to some. But what are the limits-of discourse as an "object" of study? This question demonstrates the severity of the challenges posed by poststructuralism. One must bracket even the term "object." But what do the brackets mean? Does a discursive social theory mean there are no "objects," that is to say, no contents to intellectual pram tices Is such a practice forever doomed to a world of talk about talk itself-, of the interpretation of interpretation, of a program without performances? The problem is acute when one considers the question, Is there, in the "'real" world, nondiscursive social action? It is one thing for a discursive intellectual work re treat other discursive materials of the same sort. This is what the poststructuralists mean by intertextuality in the strictest sense of the concept.
The success of poststructuralism in literary studies may rely considerably on the fact that, in this area, other texts are the proper subject-matter. The most compelling successes, in my opinion, of applied post matter. structuralism have been among feminist, third world, and Afro-American critics who uncover the discursive power of hitherto silent, oppressed women, black, or third-world writers. In a case like Henry Louts Gates' analysis of the confluence between the African Esu-Elegbara and the Afro- American signifying monkey figures in two separated but historically bound cultural systems, the analyst is applying a discursive method to texts that are found to be surprisingly discursive themselves. Both figures served to contain and express the doubled cultural experience of those who are simultaneously in some fractured way both African and American. The figures are discursive in that they mediate the divided social reality of people for whom colonial oppression and slavery was the decisive social attribute. This discovery of the discursive and political consciousness of so-called nonliterate or otherwise excluded people is parallel to similar discoveries of the study of oppressed women, the working class, and other victims of colonial domination, and this literature- of which E.P. Thompson's The Making of the English Working Class is a locus classics -is familiar and assimilate to even normal sociological thought.
The greater difficulty concerns the hint strong within poststructuralist thought that everything social is discourse. Are there no events in the "real" world that lack this transgressive, mediatise quality? This, of course, is a very familiar question, arrived at by a different route. What are we to make of the irregular silence of oppressed people Is their silence merely a latent discursively, covered by false consciousness? It is one thing to say that certain slave narratives are discursive, and another to suggest that all which is said by, or inscribed on behalf of, slaves is discursive, and still another, by extension, to suggest that slavery is nothing but discourse. This is the question that separates a prospective sociological postmodernism from poststructuralist literary criticism. Sociologists should have little difficulty accepting the idea that there are hidden or underlying variables behind surface appearances. But they will have trouble with the suggestion that those variables are exclusively discursive. Is there nothing in the "real" world but texts and discursive talk? Literary theorists and others, including social historians, can plausibly study nothing but texts. Can sociologists? Or, better put, what does it mean to propose that sociology be the discursive study of discursive texts ?
In a different guise this is the familiar problem of the presumption of a necessary difference between theory and concrete empirical data. Most sociologists could, if pressed, consider the proposition that theory is the discursive property of any sociological work. This would amount to little more than granting that in theory, whatever else we do, we state and describe both a statement about the "real" world and the rules by which we arrive at that interpretation. Usually, however, even in a radical version of this conviction, sociologists hold to the existence of a "real" world outside of the discursive sway of theory. The world's "reality" is taken, normally, as the source of concrete empirical data. This conviction, we can now see, would be treated with great scepticism by poststructuralism and postmodernism. The idea of a free-standing reality as the source of empirical data partakes of the modernist distinction between the knowing subject and the world of objects, and relies on a belief in attainable knowledge as the arbiter of that distinction. We might grant, therefore, that postmodernism would have this particular philosophical attitude toward the division of theory and data. But, can we grant that sociology can get along without free-standing data, that is, without data from the world as the resource of theory? Viewed through the lens of a postmodern critique, we can see mat the question need not be posed so narrowly. We can agree that data are necessary to even a postmodern sociology and still accept the proposition that those data are neither necessarily of an order different from theory nor nondiscursive.
This line of questioning requires a reconsideration of the status of our concept of reality; clearly postmodernism would abandon the notion altogether. So, it seems possible, even if only for tactical purposes, that one can avoid the threats of such a course. Here is where the poststructuralist ideas of discourse and textuality offer considerable leverage even with their terrible philosophical troubles.
A poststrucruralist or postmodernist approach to the concept of "reality" would be pragmatic. What do we intend by it? And can we get around it in order to enhance our ability re know and discuss? Can, therefore, the theory of Texts, including discursive texts, get us around the problems sociology, and other sciences, usually solve with reference to ideas like "empirical reality"?
The prospect of such an alternative depends on the plausibility of four assumptions already presented, explicitly or implicitly:
1 that theory is an inherently discursive activity; 2 that the empirical reality in relation to which theoretical texts are discursive is without exception textual; 3 that empirical texts depend on this relationship to theoretical texts for their intellectual or scientific value; and 4 that in certain, if not all, cases a discursive interpretation yields more, not less, adequate understanding.
Assumption 1 was stipulated in the above discussion. Assumptions 2 and 3 require further discussion. Assumption 4 is best considered with reference to a case study.
Theoretical statements mediate the "reality" contained in empirical texts - answers to questionnaires, performed rituals and observed behaviours (usually inscribed on film or rape or in notebooks), letters, corporate reports, transcripts, interviews, archives, census tracts. It is far from clear mat there are any data "purer" (that is, "more real") than these. And none of these is anything but textual in the two senses post- structuralism employs. First, they are literally inscribed on one medium or another and are never used for analysis without being thus written. Secondly, they are useful for knowledge only to the extent that they exist in an intertextual field - with other empirical texts of the same sort, with other empirical texts of a different kind, and, most of all, with the theoretical texts out of which sense is made of them. It hardly need be said mat raw data, in whatever form, are useless until they are situated with respect to theoretical statements. Theoretical statements, regardless of the "school" or methodological style in which they are expressed (scientific, humanistic, qualitative, ethnographic, etc.), are never made without a relationship to empirical data or an empirical reference, how ever abstract. Parsons' most abstract theory of the AGIL paradigm requires a great number of assumptions about the reality of the social world, such as a willingness to believe that societies are patterned, that culture is an effective control over society, that societies need integrative mechanisms like laws. None of these beliefs, however arguable, is held without reference re a wealth of empirical references. These references when held by a reader are necessary to the sense of Parsons' theory. They arise from the many empirical texts - ranging from survey results to everyday life conversations and everything in between - that inform a reader's ability to read. Similarly, such texts are also written, whether consciously or nor, as an intervention in the field of existing texts sociologists variously consider germane to their work. It is not at all clear why one needs the idea of an empirical foundation existing beyond such an intertextual field.
Of the four assumptions, 4 is the sternest test of the prospects of a post modern sociology. In the end, it is hardly worth the while to my something with so many inherent difficulties if there are no anticipated advantages over what we have now. So, then, what are the advantages? A question I propose to answer with reference to a case of undeniable, but still uncertain, reality.
Important as it is to American, and global, history the reality of the war in Vietnam is far from certain. For the majority of those who attempt to interpret it, their most vivid impressions come not from direct experience but from a strange conglomeration of texts - the memorial on the Mall in Washington, films, first-hand accounts of speakers, friends, or relatives, novels, Neil Sheehan's New Yorker articles and prize-winning book, college and high school courses, rhetorical allusions by politicians, archives, microfilm and microfiche, and so on. Is it an accident mat the most searing film account, if nor the roulette scene in The Deer Hunter, is Apocalypse Now, a montage of craziness and dream-like irreality in which the viewer is made to feel that nothing real was there? Was Vietnam after all nothing more than a repetition of a classic Conradian narrative -- a crazed voyage through an exotic jungle in search of an unamenable insane kingdom in the heart of darkness One wanes to argue that this is a fiction and mat the reality is still there. Reviews of each serious Vietnam film center on the question: Did this one, Platoon perhaps, finally capture the reality of the war?
It is possible mat the search for the reality of social things is the true Conradian search. Where would one look for the reality of Vietnam? Are recollections of veterans or POWs more real than Apocalypse Now? Are the Pentagon Papers? Are Neil Sheehan's articles? Are Stanley Karnow's history and PBS documentary? Is that finer reality still buried in an archive somewhere? And can these questions be asked of most complex social-historical events?
In pursuance of a postmodern sociology, what can then be said about the empirical reality of a series of events like the war in Vietnam? I propose that we ignore, for the moment, our sociological thirst for reality, and consider it simply and straightforwardly as though it were, for all intents and purposes, a monstrous but plausibly discursive text. In this respect, we should have to entertain the proposition that the war itself was discursive, a global inscription in which the United States sought to mediate its own sense of the irreality of world history.