3.307 Potter's book: a preliminary report (51)

Mon, 31 Jul 89 20:21:58 EDT

Humanist Discussion Group, Vol. 3, No. 307. Monday, 31 Jul 1989.

Date: MON 31 JUL 1989 18:10:00 CDT
From: Jim McSwain <F0A8@USOUTHAL>

This is a preliminary report on Rosanne G. Potter, ed., LITERARY
COMPUTING and LITERARY CRITICISM: Theoretical and Practical Essays
on Theme and Rhetoric (Univ. of Pa., 1989). It contains twelve
essays, three of which were published in the 1970s in several
journals, in addition to a bibliography assembled by the editor
and an index. The work is addressed to two sorts of researchers:
literary critics who hope to use a computer's computational and
search skills to analyze texts for factual information (occurrence
of particular words, occurrence in connection with other words, etc.)
and "computer critics," which I take to be literary scholars who have
embraced computational analysis of texts to the extent that they have
"lost" sight of the original "critical aims" of their hardware and
software procedures. The editor's preface begins with an interesting
assertion: "computer science and literary criticism differ
considerably . . . one rooted in facts, the other rooted in ideas; one
focusing on the replicable, the other on the unique." I suppose this
means that critical judgments about literature, consisting of ideas,
falls in the realm of subjectivity . . . the non-verifiable, and not
the world of "facts." There is also an interesting observation on p.
xvii that "verification" in literary criticism, counting occurrences
of words, shifts the discipline from "brillance of insight and assertion
toward the detailed testing of scientific experimentation." This
procedure concerns "inductive proofs based on example" rather than the
"more typical . . . traditional deductive proofs from authority . . .
earlier critics or one's own responses . . ." This is only a taste of
the provocative things addressed in this book, although a brief survey
of the essays shows that despite serious effort to make the statistical
material comprehensible, one still needs some sense of what statistics
is about. The editor notes, however, that one must not veer off into
number juggling, or the critical aims of using the computer are lost
by reductionist techniques. As I work through some of the original
essays, I will try to throw out more remarks which some participants
may wish to debate. Although my report is limited, I think this book
addresses many of the fundamental issues regarding texts and statistical
analysis which go to the heart of the uncertainties, theoretical
ambiguities and unspoken questions many may have about the still uneasy
relationship between the IBM logo and your diminishing sense of selfhood
and humanity. Regards, JMcSwain...