prepositional adjuncts

Van Eynde fralau at iol.it
Mon Apr 2 07:56:41 UTC 2001


At 13.49 30/03/01 +0200, Jesse Tseng wrote:
>
>I wrote:
>>> The syntactic analysis for all PPs ought to be quite
>>> uniform.
>
>Frank Van Eynde wrote: "Why?"
>
>I agree with Frank, however, that among the set of PPs we can
>find some syntactically exceptional constructions where the P
>should not be treated as the syntactic head.  For English, I'm
>thinking of examples like "at war", "on sale", "in the know".
>In these cases I still wouldn't go for a marker analysis (or a
>minor preposition analysis) because the NP doesn't seem to be
>the syntactic head either.  Instead, I would say these are
>non-headed constructions, and I am tempted to approach some of
>Frank's Dutch examples the same way.  For some of the others
>(e.g. infinitival "te"), I would consider an affix treatment.

I have also considered an affix treatment for the infinitival 'te',
but the problem is that 'te' does not behave like an affix.
If you apply the Zwicky-Pullum criteria and if you compare the
infinitival 'te' with the prefix of the past participle 'ge-',
then it gets quite clear that the infinitival 'te' is a syntactic atom
rather than an affix, see my paper for the arguments.

It is also instructive to compare the infinitival 'te' with other
uses of the same word. In 'te huur' (for rent), 'te koop' (for sale),
'ten uitvoer' etc., it combines with a verbal stem (rather than with an
infinitive) and if you apply the Zwicky-Pullum criteria to THAT use,
then it can indeed be shown that 'te' behaves more like an affix than
like a syntactic atom in its own right.

In sum, I argue that there are three types of uses for prepositions:
(1) as head of PP; (2) as non-head sister of a nominal or verbal
projection; (3) as part of a morphologically complex lexical unit
(be it as an affix or as an incorporated formative).

Frank



More information about the HPSG-L mailing list