from a review...

[Shravan Vasishth]

I tend to follow Paul Postal's lead on this. In March 2000 he gave a talk
on negative polarity here at OSU. A colleague and I raised some questions
about his judgements, and provided some corpus evidence against his
empirical claims (on which his theory rests-- though to this date I
don't really know what the theory was but an assembly of observed facts
renamed and dressed up with words like "Principle of" preceding them).

Paul Postal's response to the counterexamples was that no-one could
possibly provide a counterexample to his theory, since the only
counterexamples that would count would be from among sentences that he
himself considered grammatical (or not).  So, now, if I ever I have a
judgement problem regarding NPIs, I plan to email Postal and get
his judgement on it.

For areas other than negative polarity, I am building theories of my own
private grammars. I find it's more convenient to first build the theory
and then convince myself that the data fit.

On Tue, 9 Apr 2002, Ivan A. Sag wrote:

> FYI.
>
> This is from Mario Montalbetti's (LINGUIST List) review
> of Zagona (2002) The Syntax of Spanish:
>
> [...]
>
> A word to pre-empt inevitable attacks. It is very likely that the native
> speaker of Spanish will object to some grammaticality judgments.  This is only
> natural, given the ample dialectal diversity in Spanish. To get stuck here
> would be a pity. First, because as Chomsky (1995) has pointed out in his
> (in)famous fn7 on p203, there is no grammaticality. And second, because it
> would be yet another unfortunate case of not seeing the woods because of the
> trees. In any case, I expect judgment differences throughout the book to be
> minimal.
>
> [...]
>
> Dang! So how do I evaluate it?...
>
> -Ivan
>

--
Shravan Vasishth
http://www.ling.ohio-state.edu/~vasishth