|From:||Marcus Smith <smithma@...>|
|Date:||Wednesday, July 18, 2001, 18:05|
Jesse S. Bangs wrote:
>This is exactly what the Principles and Parameters (PP) theory of
>Universal Grammar claims to do, though with varying degrees of success.
>PP theory gives a set of possible forms for phrase structure rules
>(actually one highly abstract form),
You mean X-Bar Theory, which Chomsky has eliminated in favor of Bare Phrase
> and then a series of parameters which
>determine how those phrase structure rules are actually interpreted.
>Under PP theory languages differ only in their lexicons and the settings
>for the various parameters.
>I personally think that PP is a wonderful theory, at least in its
>underlying assumptions. The form that the theory has taken so far isn't
>yet satisfactory, but they seem to be moving in the right direction
As a matter of fact, the theory is moving away from the idea of a universal
parameter set. Mark Baker summarized the reason very clearly in his book
_The Polysynthesis Parameter_. One parameter is supposed to account for a
bundle of traits. For example, the Pro-Drop Parameter was supposed to
accout for 7 traits that distinguished Italian and Spanish from French.
Detailed investigation of these parameters usually reveals that the traits
don't actually bundle together as tightly as was originally proposed, so
the parameter has to be broken up into many parameters. The Pro-Drop
Parameter was broken down into 6 or 7 parameters based on the evidence from
dialects in Northern Italy and Southern France. This textbook case of a
parameter was destroyed based on evidence from the languages it was
originally proposed for. Examining additional languages makes its validity
even more suspect. What ends up happening much of the time is that there is
a one-to-one correlation between parameters and traits.
So what's the point of a parameter that fixes only a single trait? How is
setting the value of a parameter different from learning the trait?
>some really good ideas have been proposed.
> The theory does exactly
>what a good theory should--it accounts for (almost) all of the relevant
>data, and predicts that all languages will share some features and no
>languages will have other features.
I'll stop with paraphrases (simply because I can't remember the exact
words) from two of my professors in the past two years who worked under the
- I can't believe I used to believe that [P&P Theory].
- As the theory has advanced from Aspects to Minimalism, our ability to
account for the data has decreased.
Unfortunately, or luckily,
no language is tyrannically consistent.
All grammars leak.
-- Edward Sapir