MICHEL FOUCAULT
613
there to have been a departure, not an accidental loss, not a confusion through
some kind of misunderstanding, but an essential omission. The act of estab–
lishing a discourse is such, in effect, that in its own essence even, it can be
forgotten . That which makes it known, that which proceeds from it, is, at the
same time, that which establishes the variation and violates it. This nonacci–
dental omission must be surrounded by precise operations which can be lo–
cated and analyzed by a return to that very act of establishment. The omission
is part of the discursivity in question, it is what gives it its law. Further, this
return is addressed to what is present in the text; more precisely, one returns
to the text itself, to the text in its nakedness, and at the same time one returns
to what is marked in the blankness of the text. One returns to a void which the
omission concealed or obscured. The reexamination of all of Galileo's texts
can certainly change our knowledge of the history of mechanics, but it could
never change mechanics itself. On the contrary, reexamining Freud's text
modifies psychoanalysis itself, as the reexamination of Marx modifies
Marxism. Thus a final category must be added : a kind of enigmatic interplay
between the author and the work.
I regret not being able to conclude more positively, or at least with some
suggestions for future analysis.
Such an analysis, if it were developed, would perhaps permit the intro–
duction of a typology of discourses. In effect, such a typology would not be
constructed simply on the basis of the grammatical characteristics ofdiscourse,
on its formal structures , or even on the basis of its subject matter; without
doubt, there exist properties of relations which are properly discursive (irre–
ducible to the rules ofgrammar or logic, or to the laws of methodology) and it
is to these that one must address oneself in order to distinguish the large
categories of discourse . The relationship (or the lack of relationship) to the
author, and the different forms of this relationship constitute one of these
discursive properties. I also believe that one can find here an introduction to
the historical analysis of discourses . Perhaps it is time
to
study discourse not
only according
to
its expressive values, or in its formal transformations, but
also according to its modes of existence: the modes of circulation, evaluation,
attribution, and appropriation of discourse vary with each culture and are
modified within each ; the effect on social relationships can be more directly
seen, it seems to me, in the interplay ofauthorship and its modifications, than
in the themes or concepts contained in the works.
The author, orwhat I have tried to describe as the function of authorship,
is no doubt a function of the subject . But is it a possible or a necessary one?
