Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!amdcad!ames!rutgers!clyde!cbosgd!osu-cis!tut!tut.cis.ohio-state.edu!ptero@zwicky
From: zwicky@ptero.cis.ohio-state.edu (Elizabeth D. Zwicky)
Newsgroups: comp.ai
Subject: Re: Language Learning
Message-ID: <1399@tut.cis.ohio-state.edu>
Date: Tue, 10-Nov-87 10:40:32 EST
Article-I.D.: tut.1399
Posted: Tue Nov 10 10:40:32 1987
Date-Received: Fri, 13-Nov-87 21:39:39 EST
References: <8986@shemp.UCLA.EDU> <1125@tut.cis.ohio-state.edu> <1966@uwmacc.UUCP>
Sender: news@tut.cis.ohio-state.edu
Organization: The Ohio State University Dept of Computer and Information Science
Lines: 52
Keywords: Goal Based

In article <1966@uwmacc.UUCP> edwards@unix.macc.wisc.edu (mark edwards) writes:
>In article <1125@tut.cis.ohio-state.edu> zwicky@dormouse.cis.ohio-state.edu (Elizabeth D. Zwicky) writes:
>:Give me a break here. You CANNOT test hypotheses about whether or not
>:there is a crystallization period after which language cannot be learned
>:without dealing with degenerate cases. 

> Why can't you test hypotheses about whether or not there is a crystal-
> lization period ..... ? It seems that if we can not test it then, we
> can not test it at all. How would you test it. Use your children?
> Somebody elses ?

My very point is that the question was about first language learning. In
order to test for a crystallization period for first languages, it is
necessary to deprive a child of language to see whether s/he can still
learn it later. This will inevitably produce a "degenerate case" as
human society is centralized on language. It is also a cruel and
inhuman thing to do to a child, and so known cases are usually even
more degenerate, since they involve children with parents who do
cruel and inhuman things to them - and rarely stop at depriving them
of language. I am not suggesting that we should create any such
children, but merely pointing out that we cannot reject the data out
of hand because of the complicating factors. 

You must admit that it seems likely that first language learning
is different from second, even in children.

>:The idea of a crystallization period is supported by the data about
>:second language learning in normal humans.

> I don't see why second language learning in adults supports the
> "crystallization period". For one, when a child starts learning
> to speak they do it naturally, they have plenty of time, words
> are repeated until they absorb the correct pronunciation. The
> key words here are plenty of time, they are corrected many times,
> and it takes them many "years" to progress to fluency. 

It takes them not many but several years. In any case, you are putting
words in my mouth - I never said I was talking about adults, just
"normal humans" by which I meant people of any age who had not been
abused, were not aphasic, were of normal intelligence ... Children
do learn second languages; the interesting question is whether or not
there is a specific point at which children stop being good at learning
second languages, and the data shows that there is, and it is puberty.
In the environments in which these data have been collected, there
is no drastic change in environment at puberty, which would tend to
suggest that the diffference is not due to factors like lack of time
and patience, but has some deeper cause.


> mark

	Elizabeth