6.469 Rev: Text and Technology (Baker, Francis, Tognini-Bonelli ed.)

[The Linguist List]

----------------------------------------------------------------------
LINGUIST List:  Vol-6-469. Thu 30 Mar 1995. ISSN: 1068-4875. Lines: 171
 
Subject: 6.469 Rev: Text and Technology (Baker, Francis, Tognini-Bonelli ed.)
 
Moderators: Anthony Rodrigues Aristar: Texas A&M U. <aristar at tam2000.tamu.edu>
            Helen Dry: Eastern Michigan U. <hdry at emunix.emich.edu>
 
Review Editor: T. Daniel Seely: Eastern Michigan U.<eng_seely at emunix.emich.edu>
 
                            REVIEW EDITOR'S NOTE:
 
What follows is another discussion note contributed to our Book Discussion
Forum.  We expect these discussions to be informal and interactive; and
the author of the book discussed is cordially invited to join in.
 
If you are interested in leading a book discussion, look for books
announced on LINGUIST as "available for discussion."  (This means that
the publisher has sent us a review copy.)  Then contact Daniel Seely at
     eng_seely at emunix.emich.edu
 
-------------------------------Review-------------------------------------
 
Text and Technology: In Honour of John Sinclair. Edited by Mona Baker,
Gill Francis and Elena Tognini-Bonelli.
 
Philadelphia and Amsterdam: John Benjamins, 1993. Pp. xii +
361. ISBN 1-55619-494-3 (US)
 
Reviewed by Donald C. Freeman, Dept of Eng, U. of S. Cal.
 
About a third of the essays in this _Festschrift_
(hereafter, "T&T") report research conducted with the large
corpus of 18 million-plus words in the Cobuild (later Bank
of English) archive, and I will concentrate on these. Of the
remainder, one section on "computational tools" is beyond my
competence, save for Jeremy Clear's valuable corpus-based
study of the importance of phrases for lexicography. The
essays in the first section, "Spoken and Written Discourse,"
are standard neo-Firthian studies, of which the best
executed is Susan Hunston's "Professional Conflict --
Disagreement in Academic Discourse," a register-based study
of knowledge claims and their rhetoric. The essays in these
sections are applied in their focus and make few theoretical
claims.
 
Would that all the corpus-based studies had exhibited
similar restraint. Several authors assert that the "data-
driven" linguistic descriptions of corpus linguistics, all
based on attested utterances, make its theoretical insights
superior to those of Chomskian linguistics, purportedly
based on "introspective" data. I would find this claim at
least colorable if the essays in this section contained any
ground-breaking theoretical discoveries or falsifications of
major theoretical claims made by chalk-and-blackboard
linguistics. But they do not. As close as they get to a
theoretical claim is, in effect, a reshuffling of the
descriptive deck for the intensifiers _very_ and _utterly_
(Alan Partington) and the _actual-actually_ pair (Elena
Tognini-Bonelli). Gill Francis is more theoretically
ambitious, but manages to treat several varieties of noun-
phrase complements without even mentioning Rosenbaum 1967
and Postal 1974. Instead, we learn that 98% of sentences
with structures like "I made it my prime objective to settle
the matter" have verbs like "find" and "make," and that a
following adjective is likely to be one like "difficult" or
"easy," facts that "become obvious only when one
interrogates a [large] corpus." (p. 141)
 
But perhaps these facts are not obvious because they are not
particularly interesting. What seems most problematic in
these "corpus-driven" studies is their tendency to equate
statistical and theoretical importance.  Most major
theoretical discoveries in syntax, at least, have depended
on structures highly unlikely to occur in any statistically
significant frequency. Moreover, one of Chomsky's greatest
contributions to linguistic inquiry was to extend its focus
to _possible_ utterances rather than merely attested ones.
The value of many of these essays is diminished by
retrograde campaigns against the kind of evidence that
brought linguistics into the mainstream of contemporary intellectual life.
 
The importance of corpus linguistics for applied work is a
different matter. Bill Louw's study of "semantic prosodies"
shows that some words and phrases have a semantics that
infects their immediate context; so, for example, with _ set
in_ we are apt to find nouns with negative connotations:
decay, malaise, rot, and so on. In the study of irony, for
example, Louw argues that there must be a consistent
semantic prosody as a background against which irony can
work by violating reader expectations.
 
Louw demonstrates this claim in (among other illustrations)
an acute reading of Philip Larkin's "First Sight," a poem
ostensibly about lambing, focusing on the poem's last line:
"Utterly unlike the snow." He shows from databank evidence
that "utterly" has "an overwhelmingly 'bad' [semantic]
prosody": most immediately following adjectives in his
corpus have negative connotations ("arid," "confused,"
"meaningless," e. g.). So a reader will assign that the same
negative semantic prosody to what follows "utterly,"
interpreting "unlike the snow" as the "myriad cruelties of
the world the lambs have just entered." (p. 161).
 
While intriguing, this analysis only confirms a chalk-and-
blackboard demonstration of the same point that space limits
prevent my discussing here. Hence I find excessive Louw's
claim that "[s]emantic prosodies have, in large measure and
for thousands of years, remained hidden from our perception
and inaccessible to our intuition." (p. 173) Rather, Louw's
(quite sound) intuition found candidate semantic prosodies
only confirmed by the corpus data. This confirmation seems
the most appropriate use of corpus linguistics, which
Kirsten Malmkjaer uses in an interesting study of
translation.
 
Several authors insist that linguistics -- all linguistics
-- belongs to the social sciences. That insistence is at the
heart of the difficulties in T&T, for they are just the
problems that afflict contemporary social science: excessive
concern with the surface properties of science -- big labs,
big machines, impoverished notions of "scientific rigor" and
empirical evidence. The place of linguistics among the
disciplines always has been problematic, but in my view
linguistics is at once the most humanistic of the sciences
and the most scientific of the humanities. T&T's most
valuable essays are those in which this view informs the
uses authors make of technology and the findings of
theoretical linguistics to extend our knowledge of applied
linguistics. But I find unpersuasive the book's theoretical
and methodological claims for corpus linguistics.
 
_Works Cited_
 
Postal, Paul M. (1974) On raising: one rule of English
grammar and its theoretical implications. Cambridge, Mass.:
The M.I.T. Press.Rosenbaum, Peter S. (1967) The grammar of English predicate
complement constructions. Cambridge, Mass.: The M.I.T.
Press.
 ___
Donald C. Freeman is Professor of English and Law at the
University of Southern California. He has published
extensively in stylistics, metrics, and metaphor.
 
--------------------------------------------------------------------------
LINGUIST List: Vol-6-469.