PARTISAN REVIEW
543
assumed but elusive reader - is often the poem's central significance.
The alternative to a so-called scientific method is not necessarily a
vaguely impressionistic criticism, although Jean Cohen suggests that
it
is
when, in
Structure du langage poetique,
he tendentiously opposes
"feeling" a poem to "knowing" a poem. Thc relation of poetry to
itself (if that exists) includes an experience which is not merely
spatial, an experience which must also be described in temporal,
linear terms. The performing voice is, if you like, an objective enough
fact about the poem, but it is unavailable to the objectifying modes
of analysis proposed by structuralism. The spatial diagramming prac–
ticed by the structuralists necessarily misses the poem's movement in
and through the time it takes for the voice of the poem to express
itself and to define its audience. This time, and the numerous verbal
choices made along the temporal line, are what Poirier asks us to
follow and study, and there is no reason to believe that a study of
the moment-to-moment tactics in a poem will provide us with less
information about the specificity of poetry than a general formula
about paradigm imposed on syntagm.
As a final example of the deadening effect of structuralist ter–
minology used on individual works,
I
mention with some reluctance
Gerard Genette's essay on Proust, "Proust et Ie langage indirect," in
Figures II.
Genette has written perhaps the best essay we have on
the promises and limitations of structuralism for literary criticism
(reprinted in the first volume of
Figures) ,
and he has generally dis–
cussed literature with marvelous incisiveness and clarity. But the ad–
mirable penetration which, for example, characterized Genette's study
of
L'Astree
is replaced in the Proust essay by an irritating, and
habitual, structuralist technique which can best be called verbal es–
calation. The essay is full of plot summary and "rich" in quotations.
The only argument
I
was able to glean from it is the universally
acknowledged point that Marcel is educated into making a distinction
between words and things, between language and reality; he learns,
more particularly, that what people say cannot be trusted as a literal
indicator of what they mean or are. This familiar insight is "digni–
fied" by an array of fashionable technical terms: "illusion re£erentielle"
and "illusion semantique;' "enonce deforme" and "enonce deplace,"
"asteisme," "synecdoque descendante" and "synecdoque ascendante,"
"shifters," "glissement mbonymique," etc. These technical terms ex-
