from a review

[Shalom Lappin]

Determining Chomsky's views on what constitutes evidence for a
theory of grammar is an interesting problem in the history of ideas.
One way of settling it, at least for his current position, is simply to
ask him. Whatever the answer, it does not, in itself, entail a reply to the
question of what counts as relevant linguistic data. As various people
have pointed out, there are compelling reasons (independent of what
Chomsky or other linguists happen to think) for considering
psycholinguistic, computational, and other factors when evaluating alternative
grammatical theories. However, if a theory fails to account for the
widely held judgements of native speakers on the well formedness of a
signficant set of sentences in their language, then it is hard to see
how one could escape the conclusion that the theory has serious empirical
difficulties. If one rejects this view, then he/she is obliged to specify
alternative criteria for evaluating grammatical theories. This much would
seem to be uncontroversial to the point of triviality. Oddly, the quote that
Ivan cites appears to call this assumption into question.
It seems to do so by invoking Chomsky's views on the subject without
indicating an alternative criteria (Chomsky's or anyone else's) of
empirical adequacy. I took this to be the point of Ivan's well motivated,
if lighthearted expression of surprise. Regards.
                              Shalom