BOOKS
479
phy is not a special case; "'literature' is the effort to salvage language
from the unconscious constraints of (differential) structure." In fact, he
goes so far as to suggest that structuralism and literature are essentially
incompatible operations because structuralism questions the exclusive–
ness of literature's appropriation of language.
While Mehlman challenges " literature" (our notions of the insti–
tution of literature) more than structuralism, Jonathan Culler's
Struc–
turalist Poetics
challenges structuralism to discover how it can be
serviceable to literature. Culler's wide-ranging and closely argued
survey of structuralism opens with an examination of its underlying
linguistic models, which have derived from Saussure's principle that
" in language there are nothing but differences
without positive terms."
Although he recognizes the importance of linguistic models in struc–
turalism 's achieving its "rejection of the notion of the 'subject,'"
Culler also finds the systems of convention which replace the subject
inadequate to account for our experience of literature. And he is
particularly strong in debunking attempts like those of Jakobson to
work out " discovery procedures" for reading. While he is more
impressed by Greimas 's work in semantic structures (because "the
structural and the semiological cannot be dissociated," ) Greimas too
falls prey to a fundamental oversight: the structures are curiously
divorced from any account of the "strength of the expectations which
lead readers to look for certain forms of organization in a text and to
find them." Yet despite his quarrels with individual structuralists,
Culler understands their basic enterprise quite well: "The task is ... to
construct a theory of literary discourse which would account for the
possibilities of interpretation, the 'empty meanings' which support a
variety of full meanings, but which do not permit the work to be given
just any meaning."
Reading, rather than structuralism per se, is really the focus of
Culler's attention, and he is extremely convincing in talking about the
ways in which our expectations govern our reading strategies: we read
a sentence from a philosophical essay or a news story differently when
the words are arranged as poems. But even his efforts to demonstrate
the inadequacies of current linguistic and semiological models reveal a
curious ambiguity at the heart of his notion of reading. On the one
hand, reading remains a (perhaps inevitably) mysterious process; on
the other, he wants "empirical data," "a set of facts" about reading
which would become the basis for a theory, distinctions between the
" functional " and the "nonfunctional"-in short, as he so often says,
"evidence." A concomitant penchant toward breaking down novels
