Newsgroups: comp.ai.philosophy,sci.lang
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!udel!gatech!swrinde!sdd.hp.com!hplabs!hplntx!curry
From: curry@hpl.hp.com (Bo Curry)
Subject: Re: What's innate? (Was Re: Artificial Neural Networks and Cognition
Sender: news@hpl.hp.com (HPLabs Usenet Login)
Message-ID: <D3nA3L.GKD@hpl.hp.com>
Date: Tue, 7 Feb 1995 19:28:32 GMT
References: <3esaig$6h5@mp.cs.niu.edu> <3f6ep0$r5f@oahu.cs.ucla.edu> <3f6il0$io6@mp.cs.niu.edu> <D2IDC2.12J@cogsci.ed.ac.uk> <3gshff$e5k@tardis.trl.OZ.AU> <D3LF78.C8n@hpl.hp.com> <3h67n3$d4r@mp.cs.niu.edu>
Nntp-Posting-Host: saiph.hpl.hp.com
Organization: Hewlett-Packard Laboratories, Palo Alto, CA
X-Newsreader: TIN [version 1.2 PL2]
Followup-To: comp.ai.philosophy,sci.lang
Lines: 58
Xref: glinda.oz.cs.cmu.edu comp.ai.philosophy:25273 sci.lang:35279

: In <D3LF78.C8n@hpl.hp.com> curry@hpl.hp.com (me) writes:
: >Now really, you guys. You may disagree with Chomsky, but to assert
: >that he has *no* evidence for his claims is simply ludicrous.

Neil Rickert (rickert@cs.niu.edu) wrote:
: I just saw a drop of water on the floor.  That is evidence that a
: winged unicorn flew through the house and spat on the floor.  Sure,
: if you want to interpret "evidence" in a very broad sense, there is
: evidence.  But we don't normally talk that way.  You would be quite
: justified in saying that there was no evidence for the winged
: unicorn.  When we say that there is no evidence, we usually mean that
: there is no evidence that would persuade us.

: There might be evidence that would persuade Chomsky.  I don't know
: whether there is or not.  On my reading of Chomsky, he has some very
: strong ideas that he brings to his studies, so I cannot be sure
: whether he got his ideas of UG from his interpretation of the
: evidence, or from his prior strong committment to cartesian
: rationalism.

I find the work on Creoles, both spoken and signed, to be very
strong suggestive evidence, at the very least. The "poverty"
observations follow directly from the UG hypothesis, whereas
they require considerable handwaving to explain away without
it. In this sense, they are strongly corroborative. It is all
very well to make claims about general learning algorithms,
but I have seen no alternative to Chomsky's UG hypothesis
which makes anywhere near as specific, testable claims. In
the absense of a viable alternative, strongly suggestive
evidence is persuasive, at least to me.

: >Why, Neil, do children's "approximations" follow such a constant
: >course?

: They are probably not as constant as you think.  I suspect there is
: significant variability from child to child during the period of
: language acquisition.

Well, sure there's variation. But within fairly rigid bounds.

: >We might not all have the same "approximate grammar", but
: >the approximations we make are all of a kind.

: This may have a lot to do with commonality in our environment.  It is
: my distinct impression that a child raised in Japan makes very
: different approximations, and finishes up with a very different
: language capability.  Presumably the sounds in that child's
: environment had something to do with it.

Chomsky's work includes Germanic, Asian, and many other languages.
It is just those approximations, and ultimate capabilities,
which are *common to all known languages* which provide evidence
for universal grammar. There are many such commonalities, and they
are very difficult to explain under any alternative hypothesis
I am aware of (precisely because of the many cultural and
environmental differences you point out).

Bo
