Corpora: Chomsky and corpus linguistics

[Mills, Carl (MILLSCR)]

	As I said in an earlier post, I have followed this thread with
interest.  Having made my own small contribution, I intended to go back to
lurking and let things play themselves out.  However, the interesting
remarks by Ramesh, Tony, and Mike Barlow have led me to one more
observation.  Variants of this discussion have recurred over the years with
slight variations in emphasis depending on the forum involved.

	Criticisms that Chomskyan linguistics--which is, incidentally, a
better term than generative linguistics, a term that is both too broad and
too narrow and which is out of date, anyway--is philosophy, not science, go
back at least to the writings of the late C.F. Hockett in the 1960s.  And
Thomas Bever, no enemy of Chomskyan linguistics, pointed out the gaping
holes in Chomskyans' reliance on linguistic intuitions as early as 1970.
Since the 1970s, a research group at Edmonton, centered on the work of Bruce
Derwing and Gary Prideaux, has attempted serious non-Chomskyan work.

	In the mid-to-late 1990s, the late lamented LBRAINLIST developed a
long thread of exchanges between Chomskyan and non-Chomskyan linguists.  And
FUNKNET, a couple of years ago, had a spirited exchange between "formalists"
(i.e., Chomskyans) and functionalists.

	And of course, one major theme in the panels that Vic Yngve has been
leading in various forums has been the unscientific nature of Chomskyan
linguistics.

	None of this is to suggest that members of Corpora are re-inventing
the wheel:  the discussions have varied depending on the interests and
backgrounds of participants.  But I find it interesting that such
discussions have been going on sub rosa for a long time and seem to be
increasing, both in frequency and visibility in recent years.

	Carl Mills