Martha McGinnis: Gender in DM (response to Bobaljik posting)

[Martha McGinnis]

I think there's evidence from Betsy Ritter's work on the structure of
DP that gender isn't necessarily a property of the root, but is
something determined within an nP domain (perhaps by n itself, as
Alec suggests).

Ritter's 1995 NLLT paper argues that number affixes in Hebrew can be
attached syntactically (in which case they determine number
compositionally, but have no effect on gender) or lexically (in which
case they can affect gender, and don't have a compositional effect on
number).

In a DM framework, the "lexical"/"syntactic" split would be handled
in terms of domains.  For example, n heads a domain of special
meaning, just as agentive v heads a domain of special meaning in the
vP (see Marantz's 1997 PWPL paper). Thus number affixes attached
outside nP have compositional meaning, while those attached inside nP
(or, perhaps, as the n head itself) can have special meaning.  Under
this analysis, gender would be determined within nP, but not
necessarily by the root.

Cheers,
Martha