| |
Abstract:
Sauerland and Gibson (1998) claimed that the choice of an
attachment site for the relative clause is influenced by a case
matching preference. Under this view, relative clauses attach to a
noun phrase that matches the case of their relative pronoun. I
tested this prediction in Polish, since it is a language with
richly marked case system. Written questionnaires were used to
measure interpretation preferences for six conditions: complex NPs
of the type NP1Acc NP2Gen with accusative, genitive, or nominative
relative clauses and the same NPs in reverse order (NP1Gen NP2Acc)
with accusative, genitive, or nominative relative clauses.
The results show that the head of the NP (Acc) is the preferred
host in both orders (1-3 and 4-6) and there is no systematic effect
of the case on the pronoun. They are illustrated in the table
below:
-
Percentage of head
(NP-Acc) preferences :
|
NP1-Acc
NP2-Gen
|
NP1-Gen
NP2-Acc
|
RC
|
|
76%
|
93%
|
Nom
|
|
76%
|
87%
|
Gen
|
|
83%
|
78%
|
Acc
|
|
