[Corpora-List] RE:[Copora-List] Chomsky

[Mike Maxwell]

Otto Lassen wrote:
> Chomsky liberated linguistics from behaviourism
> but degenerated in metaphysics e.g. his move-theories:
> every practical sentence should be formed by
> changing an ideal sentence.

The notion of movement is logically independent from the question of
using corpora as evidence vs. using one's intuitions about the language
of which one is a native speaker.  The theory that many sentences are
derived by transformation from other sentences was originated by Zellig
Harris, Chomsky's linguistic teacher--and Harris worked with corpora.
(AFAIK--Harris might have had a predecessor, too.  And Chomsky's theory
of transformations has always been different from Harris's.  But both
relied on a sort of syntactic movement.)

The independence goes both ways: there are plenty of
non-transformational linguists, in the HPSG and LFG schools for example,
who rely largely on intuition, rather than (or in addition to) corpora.

I don't want to get into a long interchange on this; the debate has come
up before in this list, and the answer that many of us (well, at least
me :-)) are happy with is that there is room for both types of evidence,
corpora and intuitive.  I'd humbly suggest that people look at the
archives for this list to see that past discussion (around a year ago, I
think) before penning a reply.  Besides, my boss (who probably reads
this list) has other things for me to do :-!.

> Judging malformed
> sentences from children or not-educated people
> (and understood by everyone) as being skrewed
> is beyond linguistics. May the disciples of
> Chomsky change their views. The cognitive functioning of
> language is too complicated to fit in Chomsky's
> metaphysics.

These are straw men.  Chomsky has never had anything against evidence
from children (his wife has worked in child language acquisition) or
non-standard dialects (which in English at least is what "not-educated
people" often speak).  He just doesn't happen to be a native speaker of
such a dialect, so that isn't the evidence he uses.  Other generativists
who are native speakers of e.g. Ozark English have used such dialects as
evidence, without anyone complaining.

As for the second straw man--language being too complicated to "fit in
Chomsky's metaphysics"--he never said that his theory was supposed to
account for everything that goes on in language.  Quite the contrary.
It's just supposed to account for the results of some sorts of
"experiments" (his word, but for the record I think it's the right
word), and presumably those experiments say something about how language
works in the brain.  At least that's the claim; one can argue (and
people have) about whether the experiments represent any kind of
reality.  But if you argue that there's more to language than that, I
think Chomsky would be the penultimate one to disagree (with myself as
the last one :-)).
--
	Mike Maxwell
	Linguistic Data Consortium
	maxwell at ldc.upenn.edu