paradigms

[Martha McGinnis]

>But you can rest assured that the points about what theories are for
>aren't really necessary at this level of discussion. Saying that you
>want to derive the facts rather than say that they are merely
>memorized adds no new information to anyone's processor, at least
>not anyone reading such a list.

Perhaps I misunderstood, but I don't think Jonathan was making a
metatheoretical statement.  I understood him to be making a
falsifiable empirical claim: that paradigms may be a useful
descriptive tool, but they have no theoretical status.

The example of do-support nicely illustrates how a paradigmatic
description of a synthetic / analytic alternation doesn't tell us
anything about the source of the alternation -- i.e. whether the two
structures are syntactically identical or not. There's still no firm
consensus on whether do-support is a morphological or syntactic
alternation.  Some have argued that it's purely morphological (for
example, see Jonathan's work on adjacency), while others have argued
that do-support adds a verbal head (e.g. see Embick & Noyer's LI
paper).  DM doesn't force one view or the other: both are possible,
and only empirical arguments can decide the matter.  If those
empirical arguments are telling us something meaningful, then a
paradigmatic analysis is inadequate.

Cheers,
Martha
--
mcginnis at ucalgary.ca