|
|
RECIPIENTS BETWEEN GOALS AND POSSESSORS: EVIDENCE FROM HEBREW
Itamar Francez
Stanford University
Wednesday, September 22, 2 PM MJH Rm 126
Sponsored by the Stanford Humanities Center/Mellon Foundation
Graduate Research Program
This paper presents new data showing that the distribution of the two
Hebrew prepositions 'le' and 'el', both roughly equivalent to English
'to', is more complex than previous analyses suggest. Previous analyses
have assumed that 'le' can occur wherever 'el' can, but not vice versa.
I show that this assumption is empirically inadequate. The combination
of these prepositions with pronouns reveals three distributional
patterns: in some instances 'le' can exclude 'el', in others 'el' can
exclude 'le', and in still others they are interchangeable.
I argue that the distribution of the two prepositions is determined by
the interaction of their function with the lexical semantics of the
verbs with which they occur. Recent studies by Levin and Rappaport Hovav
(2001; Rappaport Hovav and Levin 2005) argue for a three-way semantic
distinction among ditransitive verbs:
(a) Verbs that encode transfer of possession
associated meaning: x cause y to have z (y = possessor)
(b) Verbs that encode change of location:
associated meaning: x cause z to be at y (y = goal)
(c) Verbs that compatible with both (a) and (b).
I argue that in Hebrew, 'le' functions as a marker of possessors, while
'el' functions as a marker of goals. Each preposition is therefore expected
to occur with verbs that involve the relevant semantic role.
I show that the three distributional patterns described
above for 'el' and 'le' correspond exactly to the three verb classes
identified by Levin and Rappaport Hovav. Specifically, 'le' excludes
'el' with verbs associated with meaning (a); 'el' excludes 'le' with verbs
associated with meaning (b); the two are interchangeable with verbs
that are compatible with both meanings.
The proposed analysis reveals the lexical semantic underpinnings of the
morphosyntax of Hebrew ditransitives. It also provides strong evidence
that the same semantic patterns underlie the morphosyntax of ditransitive
constructions cross-linguistically.
|
|