In keeping with this distinction, we can understand why it isinevitable that practitioners of such discourses must 'return to theorigin*.

The initiation of a discursive practice,unlike the founding of a science, overshadows and is necessarilydetached from its later developments and transformations. As aconsequence, we define the theoretical validity of a statement withrespect to the work of the initiator, whereas in the case of Galileoor Newton, it is based on the structural and intrinsic norms estab-lished in cosmology or physics. Stated schematically, the work ofthese initiators is not situated in relation to a science or in thespace it defines; rather, it is science or discursive practice thatrelate to their works as the primary points of reference.

a scientific programme, the founding act is on an equal footingI with its future transformations: it is merely one among the manymodifications that it makes possible. This interdependence cantake several forms. In the future development of a science, thefounding act may appear as little more than a single instance of amore general phenomenon that has been discovered. It might bequestioned, in retrospect, for being too intuitive or empirical andsubmitted to the rigours of new theoretical operations in order tosituate it in a formal domain. Finally, it might be thought a hastygeneralization whose validity should be restricted. In other words,the founding act of a science can always be rechannelled through' the machinery of transformations it has instituted.

other hand, Marx and Freud, as'initiators of discursive practices', not only made possible a certainnumber of analogies that could be adopted by future texts, but,as importantly, they also made possible a certain number of dif-ferences. They cleared a space for the introduction of elementsother than their own, which, nevertheless, remain within the fieldof discourse they initiated. In saying that Freud founded psycho-analysis, we do not simply mean that the concept of libido or thetechniques of dream analysis reappear in the writings of KarlAbraham or Melanie Klein, but that he made possible a certainnumber of differences with respect to his books, concepts, andhypotheses, which all arise out of psychoanalytic discourse.

The distinctive contribution of these authors is that they pro-duced not only their own work, but the possibility and the rulesof formation of other texts. In this sense, their role differs entirelyfrom that of a novelist, for example, who is basically never morethan the author of his own text. Freud is not simply the author ofThe Interpretation of Dreams or of Wit and its Relation to theUnconscious and Marx is not simply the author of the CommunistManifesto or Capital: they both established the endless possibilityof discourse. Obviously, an easy objection can be made. The authorof a novel may be responsible for more than his own text; if heacquires some 'importance' in the literary world, his influence canhave significant ramifications. To take a very simple example, onecould say that Ann Radcliffe did not simply write The Mysteriesof Udolpho and a few other novels, but also made possible theappearance of Gothic Romances at the beginning of the nine-teenth century. To this extent, her function as an author exceedsthe limits of her work. However, this objection can be answeredby the fact that the possibilities disclosed by the initiators of dis-cursive practices (using the examples of Marx and Freud, whomI believe to be the first and the most important) are significantlydifferent from those suggested by novelists. The novels of AnnRadcliffe put into circulation a certain number of resemblances andanalogies patterned on her work - various characteristic signs,figures, relationships, and structures that could be integrated intoother books. In short, to say that Ann Radcliffe created the GothicRomance means that there are certain elements common to herworks and to the nineteenth-century Gothic romance: the heroineruined by her own innocence, the secret fortress that functions as

I believe that the nineteenth century in Europe produced asingular type of author who should not be confused with 'great'literary authors, or the authors of canonical religious texts, andthe founders of sciences. Somewhat arbitrarily, we might call them'initiators of discursive practices'.