:: CIIL Grammars ::

Theory of binding Book

TOWARDS AN ALTERNATIVE
THEORY OF BINDING

Abhilasha Jain
and
B.N. Patnaik

PRO can occur in any argument position, governed or ungoverned and its illegitimate occurrences can, then, be filtered out by (6) above or by its failure to receive antecedent etc., a matter we will turn to later in this chapter.

Now, why should the PRO Filter make a mention of the CM ka and thereby assign it a special status? We argue below that the distinction between ka and the other lexical CMs is justified. We show below that although ka has phonological substance like other lexical CMs, it differs from the latter in that it lacks semantic substance and is an empty CM in this sense. As such it is only proper to make a distinction between ka and the other lexical CMs.

In the discussion that follows we refer to the ka of the ka-na construction as the nominalizer ka and distinct from the ka in the genitive phrases such as [ram ka ghar] (Ram's house) which is an NP. The nominalizer ka differs from the other lexical CMs in at least three ways:

a)		it does not reflect AGR
b)		it can be optionally deleted
c)		it has no semantic substance

In Hindi, only those CMs which are a-ending manifest AGR. Of the CMs in Hindi (ne, ko, se, me, etc.) only the ka of the ka-na construction and the genitive CM ka are a-ending. Therefore in order to test the correctness of (a) one has to contrast the ka under reference with the genitive ka. Whereas the genitive ka exhibits gender and number agreement, the ka of the ka-na NP does not do as:

(7)	(i)	ram ki kitab acchi hai ram CM book nice be+PRES (Ram's book is nice)
	(ii)	ram ka ghar accha hai ram CM house nice be+PRES (Ram's house is nice)
	(iii)	ram ke kutte acche hai ram CM dogs nice be+PRES (Ram's dogs are nice)