Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!ix.netcom.com!netcom.com!vlsi_lib
From: vlsi_lib@netcom.com (Gerard Malecki)
Subject: Re: What's innate? (Was Re: Artificial Neural Networks and Cognition
Message-ID: <vlsi_libD33461.36s@netcom.com>
Organization: VLSI Libraries Incorporated
References: <3g9ab6$es@mp.cs.niu.edu> <3gbiam$7ja@ixnews1.ix.netcom.com> <3gbknj$7tj@mp.cs.niu.edu>
Date: Fri, 27 Jan 1995 22:08:25 GMT
Lines: 46

In article <3gbknj$7tj@mp.cs.niu.edu> rickert@cs.niu.edu (Neil Rickert) writes:
>In <3gbiam$7ja@ixnews1.ix.netcom.com> petry@ix.netcom.com (david petry) writes:
>>In <3g9ab6$es@mp.cs.niu.edu> rickert@cs.niu.edu (Neil Rickert) writes: 
>
>>>  If
>>>a group of children are not exposed to language, they will probably
>>>invent one.
>
>>Is there any evidence that that is actually true?
>
>As far as I know, the following are true:
>
>  (i)	Deaf children, not provided with any exposure to a sign
>	language, spontaneously invent one.  This does not occur
>	with an isolated deaf child in a community of normal
>	people.  But in a community of deaf people, this apparently
>	happens.  You will find some mention of this in "Seeing
>	Voices", by Oliver Sacks.

How does sign language compare with linguistic ability as far as UG goes?
Is it the same part of the brain (for eg., Broca's area or whatever) that
controls sign language coherency? In general, how are both significantly
different from learning a skill like juggling or playing a video game? In
the latter we have the syntactic rules replaced by the rules of the game.

My guess is that even speech is produced by templates rather than construction
from scratch. For example, we tend to punctuate our conversations
with often meaningless phrases like "well", "you know", "sort of", etc.  
Even larger phrases like "all said and done" are treated as a single word
and no subconscious effort is made in atomizing them. the situation is 
very similar to compilers. In most of the cases, the machine language
instructions are not directly produced by the grammar engine, but are
actually templates (written by humans, of course), patched up together. 
This relieves the work of the compiler. In the case of humans, the templates
are what he/she learns in early life.
>
>  (ii)  Identical twin children apparently tend to invent some of
>	their own private language modifications.  (This based on
>	second hand information from friends who had identical
>	twins.)
>
Does this happen when two children who are not identical twins grow up
together? 

Shankar Ramakrishnan
shankar@vlibs.com
