JAMES D. McCAWLEYt
fying parts of speech "on a basis of general resemblance" to typical examples
is "the [technique] used, often unconsciously, by all who identify parts of
speech successfully" (Gleason 1965:120).
As the quote from Bloch & Trager suggested, the use of traditional syn-
tactic terms such as noun and verb by American descriptivist linguists was
for reasons of convenience: it made texts more readable, since it avoided
proliferation of unfamiliar terminology that served no significant purpose.
Early generative grammarians likewise used many traditional syntactic terms
for reasons of convenience, but a very different kind of convenience: conve-
nience of the writer who is impatient with the task of creating names for his
various constructs, rather than convenience of the reader who has to interpret
the resulting terminology. In early transformational grammar, labels corre-
sponding to putative syntactic categories proliferated wildly, in part because
early transformational theory required any difference in what things com-
bined with in deep structure to be reflected in a difference in their categoriza-
tion, and in part because category names were taken to have no internal struc-
ture, notwithstanding the typographical complexity of many of the names.
Since some verbs had idiosyncratic combinatoric roles in the deep structures
posited in early transformational grammar, its policies on categorization
implied that no category symbol could take in all 'verbs' in any traditional
sense. That, however, did not prevent transformational grammarians from
adopting "Verb" and various abbreviated and decorated versions of it as cate-
gory labels, and such labels were created with gay abandon as names for pu-
tative syntactic categories that at least had a verb as a constituent, as illus-
trated by the following structures for examples in Chomsky (1957) and Lees
(1960), constructed in conformity with their rules:
(7) a
Dostları ilə paylaş: |