Section
V: a |
We now turn to the issue of the domain in
which an anaphor must find its antecedent. It may be recalled
that we have argued that the notion of "governing
category" cannot account for the occurrence of the
anaphor in the so-called dative subject construction.
Thus, there is a need for an alternative notion of the
binding domain. |
To start with, the binding domain of an anaphor
is the minimal proposition (henceforth P) where proposition
is a construction containing the anaphor/pronoun, the
anaphor's or pronoun's predicate and the mukhya
of that predicate. A certain subset of Ss, namely, those
containing agentive etc. predicates, the ka-na
NPs in which the predicate is always agentive and some
instance of [X ka Y], where Y is a simple or a derived
noun, qualify as P under this formulation. Consider the
following: |
(132) |
|
ram ka moh
ram CM attachment
(Ram's attachment.) |
(133) |
|
pani ka ghara
water CM pot
( Water's pot = Pot of water) |
|
|
(132) is ambiguous in the well-known way:
(a) ram is the target of someone's moh and
(b) ram is the one who is in the state of moh
for something or somebody. As such, it is only in the
latter interpretation that ram is the mukhya.
Hence, it is only in this interpretation that (132) qualifies
to be a P.
(133) does not qualify to be a propostion since it does
not contain a mukhya.
The notion of P is reminiscent of the notion of "Complete
Functional Complex" (CFC) of Chomsky; 1986a; 169).
However, it is the clearly different form notion of CFC
in that CFC involves grammatical function while the notion
of P does not. |
We now propose (134) as the statement of
the binding of an anaphor in Hindi: |
(134) |
|
an anaphor must be bound
in P where bound
means be assigned a mukhya as antecedent.
|
|
|
However, as already pointed out the theory
does not prevent an anaphor from being a mukhya
or part of the mukhya phrase. The binding principle
(134) needs to be relaxed in the light of this since it
will undergenerate. Therefore (135). |
(135) |
|
an anaphor must be bound
in P if P contains a mukhya distinct
from the anaphor. |
|
|
The following logically follows from the
requirement that anaphors have to be bound in the smallest
possible domain: if the mukhya is not distinct from the
anaphor or contains the anaphor then the anaphor must
be bound in the P , where P is the proposition immediately
containing P. Needless to say that P is the smallest possible
domain for such an anaphor. Thus, (135) must be supplemented
by (136). |
(136) |
|
an anaphor must be bound
in P such that is
does not violate (135). |
|
|
However, the following of data show that
(135) is inadequate to account for the anaphor-antecedent
relationship in Hindi: |
(137) |
|
*[ram aur mohan] ne socha ki
ek dusre ghar i
i
ram an mohan CM think+PAST COMP each other home
jayenge
go+FUTURE |
(138) |
|
mohan ne [(e) khana khate
hue apne dost] ko
mohan CM food eat PART self's friend CM dekha
see+PAST
(Mohan saw his friend eating food.) |
|
|
In (137), the P of ek dusre is ek
dusre ghar jayenge. This is an instance of an anaphor
being the mukhya. In terms of (136) it will look
for an antecedent in P which is root S. It does have a
possible antecedent in ram aur mohan. Thus, (136)
is satisfied. Yet, the sentence is ungrammatical. |
Consider (138). It has an embedded relative
clause khana khate hue apne dost which has the
structure given below (ignoring many details): |
(139) |
|
 |
|
|
The NP apna dost is the P for the
anaphor apna. The P for the anaphor is the root
S where mohan is the mukhya. NP* is not
P for apna because it does not contain a distinct
mukhya because (e) being coindexed with apna
dost. The embedded S in (137) is tensed while in (138)
it is non. The ungrammaticality of (137) shows that an
anaphor cannot look for an antecedent beyond the tensed
S. In view of this, we revise (136) to (140). |