Typology, generative grammar, and statistics

[James Gair]

I have had only a chance to scan the postings on the statistics and
typology topic so far, and will get back to them more carefully.
However, I was really happy to see this exchange which gets to very
basic questions not  only about method, but also about the science of
language and ultimately what language is, (or perhaps more precisely
what various 'we's' think it is.). Speaking for myself,  since i have
been occupied with both typology and generative grammar for some
time, and I felt emboldened to add to volunteer some simple views at
this early point. I would consider all of the approaches as making
important and crucial contributions to the science. At the risk of
being self-serving, One small attempt to link these (except the
statistics) was the workshop we held in Delhi on anaphora in 14 South
Asian language,(and the resulting (expensive) volume Lexical Anaphors
and pronouns in selected South Asian Languages, ed. Barbara C. Lust
et al. , Mouton deGruyter 2000) The detailed  outline that the
authors were derived  directly from findings (or more accurately,
perhaps  questions and observations) from generative linguistics, but
the work was also intensively  descriptive and empirical, and authors
were invited to point out where the framework proved inadequate for
their specific languages. One result was the non-polemical
challenging of some claims that had been made made when faced with
data collected in a directed and organized fashion, but at the same
time some principles received vindication.



Though this is a small contribution, the  point here is that there is
much to be gained by considering each approach and its possible
contribution in linkage with the others. This includes, of course
statistical approaches of different kinds., and work like  that
furthering Greenberg' insights and findings along with such
wide-ranging work of Dyer's s, at the very least raise questions that
those of us who are formally inclined have to face.  Perhaps the
fundament question is 'what and why', and the more what the better,
especially if  we keep the 'why' goal in mind.

In any case, I hope that this really fascinating exchange continues.

James W. Gair
Professor of Linguistics Emeritus, Cornell University.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20030926/e0df96ba/attachment.htm>