prepositional adjuncts

Stefan Mueller Stefan.Mueller at dfki.de
Fri Mar 30 20:35:45 UTC 2001


Hi everybody,

Ulrich Koch wrote:
> for languages like German, there is a further argument
> in favor of the head-marker analysis of some "PPs":
>
> It is commonly assumed that in a PP, the preposition selects
> the case of its NP sister.  This actually works for many
> prepositions, but there are counter-examples:
>
> (1) Hans verlaesst sich auf einen Freund.
>     Jack relies    self on  a-ACC friend
>     `Jack relies on a friend.'
> (2) Die Loesung  beruht auf einem neuen Ansatz.
>     The solution rests  on  a-DAT new   approach
>     `The solution is based on a new approach.'
>
> Both "verlaesst" and "beruht" select "auf" (i.e. you can't
> use any other preposition instead), but in (1), the sister
> of "auf" must be in the accusative case, whereas in (2),
> it must be in the dative.
>
> With the head-complement analysis, we have to posit two
> lexical entries for "auf" that differ in a feature called,
> say, OBJ-CASE. Then each verb selects the preposition with
> the correct OBJ-CASE value. This solution seems quite ad hoc.

There is a simpler solution to that: The preposition has a case feature
the value of which is shared with the case value of the NP. So there is
just one lexical entry for the complement preposition `auf'. The problem
of Ps having a case feature remains.

Greetings

	Stefan



--
Fon (Jena): (+49) (+3641) 9 44 320     (Berlin): (+49) (+30) 34 50 77 56
Fax (Jena): (+49) (+3641) 9 44 312     (Berlin): (+49) (+30) 34 50 77 77

http://www.dfki.de/~stefan/

http://www.dfki.de/~stefan/Babel/Interaktiv/Babajava/



More information about the HPSG-L mailing list