Re: [tied] 'Hard-Wired' Grammar Rules

From: Piotr Gasiorowski
Message: 12092
Date: 2002-01-18

Gerry,
 
This is not really a Cybalist question, but if you want to know my opinion as a linguist, Larry Trask is dead right about Chomsky and the Chomskyan paradigm. The linguistic ability is of course innate; neurally hard-wired "parameters" are scientific fiction.
 
Piotr
 
 
----- Original Message -----
From: Gerry Reinhart-Waller
To: language-origins@yahoogroups.com ; GlobalBrain ; cybalist@yahoogroups.com ; palanthsci@yahoogroups.com
Sent: Friday, January 18, 2002 5:57 PM
Subject: [tied] 'Hard-Wired' Grammar Rules

To Jose Luis and all,
Since our leading syntax specialist decided to retire for a while, I didn't post this.  However, now that he has returned (welcome back Jose Luis) I would like to hear his comments on this.
 
The initial article appeared in the NYTimes by Brenda Fowler.  A copy was sent out by evol.-psych and the following with comments by Larry Trask is from Anne's Palanthsci.  Thank you all.
 
Questions (from me):
1)  Do most linguists agree that language is innate?
2)  What about the grammars of all languages?  Are they similar?
3)  Does the article (Larry Trask in particular) present an accurate explanation of Chomskyian grammar?
4)  Have we finally reached a time when both sides to the argument can be melded?
5)  If melding can occur, then who would like to take a stab at the meld?
 
I shall post this also to Cybalist and ask Piotr, John, Glen, Marc and all to comment as well. 
 
Best wishes,
Gerry
 
=============================================================
New York Times
 January 15, 2002

'Hard-Wired' Grammar Rules Found for All Languages
 By BRENDA FOWLER

 In 1981 the linguist Noam Chomsky, who had already proposed that
language
 was not learned but innate, made an even bolder claim.

The grammars of all languages, he said, can be described by
 a set of universal rules or principles, and the differences
among those grammars
are due to a finite set of options that are also innate.

While most linguists would now agree that language is innate, Dr.
Chomsky's ideas about principles and parameters have remained bitterly
controversial. Even his supporters could not claim to have tested his
theory with the really tough cases, the languages considered most
different from those the linguists typically know well.

 TRASK: Exactly.  Chomsky has worked on no language other than English, and
 his
 followers have worked on only a handful of other languages, almost all of
 them European languages which demonstrably share a common ancestry with
 English.  The 6000 or so other languages of the world have received no
more
 than the occasional passing mention.  Accordingly, sweeping claims about
 universals of grammar are out of order.  It's rather as though I were to
 look carefully at six animal species here in Sussex and then announce my
 "principles of universal zoology" to a startled world.

 But in a new book, Dr. Mark C. Baker, a linguist at Rutgers University
 whose dissertation was supervised by Dr. Chomsky, says he has discerned
 the parameters for a remarkably diverse set of languages, especially
 American-Indian and African tongues.

TRASK: He has, has he?  I'm afraid that working out the grammar of a
single
 language is an enormous task requiring years of dedicated work.  I really
 do not believe it is possible to read through other people's publications
 on a number of American and African languages that you don't know yourself
 and thus "discern" anything of substance.

 Now, what about those hypothetical "parameters"?  In the New York Times
 article, Baker suggests that we might need something like thirty
parameters
 in all, though I don't know how he arrives at this figure.  But this
 "parameter" business is deeply suspect.  Here's what seems to happen in
 practice.

 The Chomskyans look at a new language.  Its grammar appears to be
different
 from the grammars of other languages already examined.  So, they invent a
 new parameter P, and they declare that the new language has a grammar with
 this parameter assigned to setting P2, while all the other languages have
 grammars assigned instead to setting P1.  Ergative syntax?  A new
 parameter.  Object incorporation into verbs?  A new parameter.  No
 COMP-trace effects?  A new parameter.  Obligatory passivization?  A new
 parameter.  Verb serialization?  A new parameter.  And so it goes.

 You can see where this is leading.  Every detectable difference between
one
 grammar and another is squeezed into a "parameter" setting invented for
the
 purpose.  In this way, the hypothesis of universal grammar is rendered
 invulnerable to falsification.

 But there's more.  Languages being the tiresome beasts that they are, they
 sometimes refuse to obey their own parameters.  Some people might consider
 such cases to be falsifications of Chomskyan theory.  But not the
 Chomskyans.  Their solution?  They have invented a distinction between
"core grammar", which is subject to the requirements of universal grammar,
 and "peripheral grammar", which is not.  So, if some phenomenon in some
 language fails to behave as required, it is declared to be merely a part
of
 the periphery, and therefore not subject to universal grammar, and
 therefore not  a problem.

 Folks, I am not making this up.  This is an honest and accurate account of
 the way the Chomskyans do syntax.  And I hope it explains why so many of
us
in linguistics prefer to pursue other approaches.

Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.