The only argument which could be the antecedent of PRO
here is ram, but, if PRO receives ram
as its antecedent, this would violate the semantic property
of the verb lag as used here. Recall the discussion
on the semantic property of pasand in the chapter
on Binding. We observed that pasand is not a
self-oriented predicate in the sense that the perceiver
of the predicate and the theme have to be distinct.
lag is a predicate of the same class. It is clear
from the following: |
The
occurrence of PRO in (30) and (31) would be legitimate.
PRO does not violate the PRO Filter and in each case
it receives the proper antecedent within the sentence.
In (30), the antecedent is kitab in (31) it is
larka. The PRO in the VP in the dative subject
construction would be illegitimate because the assignment
of the dative subject as antecedent to PRO in the absence
of any other possible antecedent would violate the semantic
property of the verb, as discussed earlier. In sum,
all the legitimate occurrences or PRO in Hindi are correctly
accounted for in our theory and all the illegitimate
occurrences of PRO are also correctly ruled out by the
theory. We have argued that the theory that PRO must
occur in the ungoverned position alone cannot account
for the legitimate and illegitimate occurrences of PRO
in Hindi. In addition, the proposed theory has no conceptual
problems of the sort mentioned above that the traditional
theory of PRO has Of course, the correctness of the
claims in favour of the proposed theory can be properly
assessed only when we discuss the question of antecedent
of PRO. We address ourselves to this question below. |
