Jeffrey Goldberg (J.Goldberg@cranfield.ac.uk)
Mon, 17 Jul 1995 16:54:54 +0100 (BST)
On Mon, 17 Jul 1995, Gert Webelhuth wrote: > On Mon, 17 Jul 1995, Jeffrey Goldberg wrote: > > > It is standardly assumed (and has been made explicit in a number of places > > as well) that non-finite verbal froms do NOT specify case information about > > their subjects. > > The solution described above cannot be the whole story, however, because > the SUBJ of the infinitive does behave as if it bears NOMINATIVE CASE when > it is not embedded under a raising verb, as is shown when there is a > secondary predicate in the sentence which agrees in case with the > SUBJECT, e.g. in German: > > [Ein Held zu sein] macht Spass > a hero to be is fun > nom I will try (as I have tried to do often in my life) to avoid much disussion of the internal stucture of German NPs, but one could easily imagine that it is something like [np [det Ein] [n1 Held] [vp zu sein]] The head "Held" has the case of the entire noun phrase (NOMINATIVE) which is comes from the lexical entry for the non-finite verb "macht". The exact bar levels and relationship between the head noun and the determiner is not really important. The only important assumption is that "Held zu sein" is not an infinitival S, but is an N followed by and infinitival VP, and the the N, "Held", is the head of its phrase. > Sie liessen ihn Vorsitzenden/*Vorsitzender werden > they let him chair person become > acc /*nom Same thing. the "ihn" gets its case from "liessen", and "werden" in a non-finite from doesn not assign case, but it does force case agreement between its subject (which gets its case from "liessen") and its complement "Vorsitzenden". This is actually a really nice example of how unification is supposed to work. > I imagine you find the same effect with predicative adjectives that > inflect for case, e.g. in Icelandic. There thus does have to be SOME > mechanism to make the subjects of certain infinitives NOMINATIVE. Does > your theory cover these cases, Jeff? An extra mechanism is needed for Icelandic (see the Sag, Kartunnen, Goldberg paper mentioned earlier. Its title is "A lexical analysis of Icelandic case") However, I do not see that the "trick" we used for icelandic is needed here at all. Best wishes, jeff Jeffrey Goldberg Email: J.Goldberg@Cranfield.ac.uk WWW: <http://WWW.Cranfield.ac.uk/public/cc/cc047/>
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:33:03 PST