empty elements

[Mike Maxwell]

At the risk of getting into s.t. I know zip about (empty categories in
current versions of HPSG etc.), I'll just say what seems to me to be the
major difference between empty categories in syntax, and empty inflectional
morphemes in morphology.

In syntax, the empty categories are getting their properties
syntagmatically, that is, from the environment in which they appear (if an
empty category can be said to "appear").  So either in the context of a
construction, or (at least in old-style GB--that should date me!) from their
syntactic environment ("governed" or not) and from their antecedents.

On the other hand, in inflectional morphology (and I think this is what
Berthold Crysmann was referring to), a zero morpheme gets its properties
paradigmatically, that is, in opposition to the various overt morphemes
which could have appeared, but did not.  Or in a word-and-paradigm
morphology (as Farrel Ackerman alluded to), the _absence_ of any morpheme in
a particular "slot" (rather than the presence of a zero morpheme) indicates
morphosyntactic properties by virtue of the morphemes which could have
appeared in that slot, but did not.

Whether the latter explanation extends to derivational morphology or not, is
another question.

     Mike Maxwell
     Linguistic Data Consortium
     maxwell at ldc.upenn.edu