AW: Increasing interest in the HPSG conference

[John Nerbonne]

Sorry to have been offline for so long.

And my apologies at the very least to Anne Abeille, who certainly should
be mentioned in any acount of linguistically interesting HPSG (and TAG!)
work.  I know I owe similar apologies to others who I didn't mention,
but I couldn't even try to be exahustive.   Detmar and Anne gave examples
of people who've already written papers in the way I suggested, which
means that the advice was (for them) superfluous, but I feel I still see
papers taking the risk of losing a big chunk of their potential readership
by couching most of what they have to say in HPSG formalism, e.g.,
making their general linguistic points via AVM's.  I agree that Detmar's
and Anne's papers are nice examples of how to separate the general
linguistic interest from the nuts and bolts in AVM logics, and the
publication in Language is a guarantee that Anne's paper (with
Danielle Godard) was regarded as sufficiently accessible to the general
linguist.

Tibor interpreted me to advocate ignoring logical issues, but my hope
would be that many HPSGers would continue that work, only that it
wouldn't dominate the aggregate of presentations so much.  This is what
I intended originally, too.  Sorry, Tibor, if the posting was obscure.

Carl asked what I meant by my second point, "What's new?", to which
I should like to reply that the way others have reformulated it seems right,
namely that GB dictates the research agenda too much. Too many
papers simply show that the results of a GB analysis can be recast into
HPSG, and too many fairly novel analyses get buried in--very valuable,
but nonetheless limited--detailed internal criticisms.  There is a lot of
interesting syntax outside the 20 or 30 common topics that dominate
theoretical linguistics.

Ken Church suggested in one of his ELSNET columns that program
chairs in computational linguistics might ask whether papers submitted
to conferences involved substantial innovations in technique or
application area, and that program chairs should make allowances for
lower technical marks in the case of such papers.  Including more
such papers would be a way of preventing your conference from
becoming a collection of nicely detailed, closely argued papers with
too little innovation and, ultimately, too little import.

I realize that I'm not responding to everything that was said, but I wanted
to focus on the point of increasing interest in HPSG.  I read all the
responses, and I appreciate the fact that so many people took the
time to reply.

This is obviously a list that can take some criticism.  Thanks to all
of you for listening critically but not responding emotionally. Two people
wrote privately saying that they agreed but didn't want to get flamed.   It
is clear that their fears were unfounded!

John
--
John Nerbonne, Information Science              +31 50 363 58 15
P.O.Box 716, University of Groningen        FAX +31 50 363 68 55
NL 9700 AS Groningen, The Netherlands	www.let.rug.nl/~nerbonne