home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!europa.asd.contel.com!paladin.american.edu!auvm!FAC.ANU.EDU.AU!ANDALING
- Message-ID: <9211100517.AA00618@fac.anu.edu.au>
- Newsgroups: bit.listserv.csg-l
- Date: Tue, 10 Nov 1992 16:17:02 EST
- Sender: "Control Systems Group Network (CSGnet)" <CSG-L@UIUCVMD.BITNET>
- From: "(Avery Andrews)" <andaling@FAC.ANU.EDU.AU>
- Subject: language & modelling
- Lines: 25
-
- [Avery Andrews 9211414]
-
- (Dag Forssell (921109))
-
- I don't think the lack of good modelling in linguistics by people
- interested in PCT due to ignorance of PCT--it's due to ignorance of
- how to build `essentially connectionist' models (e.g. ones that are
- consistent with what we know about how neural circuitry works) that
- actually do the sorts of things that language use obviously involves.
- I can't do it, Penni can't do it, Bruce can't do it. I suspect that
- the person who figures it out will be neither a linguist nor a
- neuroscientist, but somebody with a substantial but maybe not
- fully professional grade knowledge of both subjects, plus lots of
- some kind of math. & this person may come along next Thursday or
- 100 years from now, or never.
-
- At any rate, great sophistication in linguistics is certainly not
- required to work on this subject - any PCT modeller could start
- tackling the problem of, say, how to build a device that would
- hear or read a description of where a place was supposed to be,
- and then go there. In fact I think that would be an excellent
- problem to work on for people who think that the linguists on
- CSGNet are perverse.
-
- Avery.Andrews@anu.edu.au
-