LINGUIST List 2.883

Sat 28 Dec 1991

Disc: Nominative with Infinitive

Editor for this issue: <>

Directory

1. , Re: 2.868 GB
2. "Geoffrey S. Nathan", Nominative with infinitive

Message 1: Re: 2.868 GB

HASPELMATHphilologie.fu-berlin.dbp.de writes:
"A language I am working on, Lezgian (Nakho-Daghestanian), also shows
Nominative case in non-finite clauses, e.g.
(2) Didedi-z gada-0 agaq'-na k'an-zawa. 'Mother wants the boy to arrive.'
 mother-DAT boy-NOM arr.-NONFIN want-PRES"
Nominative in non-finite clauses is familiar from Icelandic (e.g.
Andrews paper in the LFG volume edited by Bresnan and other recent
treatments; Sigurdsson's dissertation, perhaps). However, what
makes it clear that the nominative here is in the lower clause, as
opposed to the higher? DAT/NOM configurations with 'want' are common
(e.g. Russian), where the nominative is an object of 'want', and the
non-finite verb would then be part of a second object, its subject
controlled by 'boy-NOM'. Cf. 'What I want of him is that he arrive' for
a double-object use of 'want'.
-David Pesetsky

=========================================================================4000001
 B0C
Although in general I agree with the Indo-European bias of most
current formal linguistic theories, you don't have to go to
Dravidian to find languages where embedded infinitive clauses
take nominative subjects. In Portuguese, first of all the
infinitive is inflected for person, and second, infinitive
clauses optionally allow subjects (Ptg. is a pro-drop language).
Thus the following is well-formed in Ptg:
Na~o e necessidade de (tu fazer-es tanto barulho)
not is necessary of you-nom make-inf-2sg so-much noise
`It is not necessary for you to make so much noise.'
A graduate student of mine just handed in a term paper on this
construction, which is why the data is handy. His name is
Aaron Smith and he is not on e-mail. The problem is not unknown
to GB theorists, but their approach to dealing with it seems
very unprincipled and ad hoc. His solution to it is about as
ad hoc but waddayagonnado? Most IE languages mark these
subjects with accusatives and so the theory was designed to
work that way.
 Geoff Nathan <ga3662siucvmb>
 Southern Illinois University at Carbondale