prepositional adjuncts

[Jesse Tseng]

Stefan Mueller writes:

> In addition to that I decided not to treat them as markers in my grammar
> of German since there are differences between NPs and PPs as far as
> extraposition is concerned. NP extraposition is more marked than PP
> extraposition. If semantically vacous prepositions are (marked) NPs they
> should be linearized as NPs but they aren't. Of course one could
> distinguish between marked and unmarked NPs but then one had three
> classes instead of two.

Well, there will always be three classes, the choice is whether
you want two classes of PPs or two classes of NPs.

Like Stefan, I stick with the head analysis instead of a marker
approach in my thesis.  Broadly speaking, semantically vacuous
Ps and other Ps have the same syntactic properties, and the PPs
they project show the same syntactic behavior.  So I see no
motivation for grouping prepositions into two different
syntactic categories or making some of them heads and others
non-heads.  The syntactic analysis for all PPs ought to be quite
uniform.

Vacuously-headed PPs do behave differently with respect to
pronoun and quantifier binding, semantic role assignment, and so
on.  But these differences can and should be traced back to the
empty semantic content of the preposition (not to its syntactic
features).  In my analysis, prepositions are always treated as
heads, and they get the first chance to contribute the phrasal
content.  If the head is semantically empty, however, the
non-head's content appears on the mother instead.

Jesse