Section
VI |
In the first sentence of (157) - (159), P
is the root S, in which the pronoun does not have an antecedent
(henceforth antecedent-free) as expected and the sentences
are grammatical. In the second sentence of (157) - (159),
the pronoun has been assigned an antecedent within the
P and the sentence are ungrammatical, as predicted. |
In (160) and (161), P is the possessive phrase
uski kitab and P is the root S. In the first sentence
of (160) the pronoun is antecedent-free in the P, as expected.
In P, there is no argument other than the mukhya
for the pronoun to choose as antecedent. But a pronoun
cannot be bound in P. Thus the pronoun is antecedent-free
in the sentence and the sentence is grammatical. In 160
(ii), however, the pronoun is bound by the mukhya
in P . Consequently the sentence is ungrammatical. |
In (161) (i), the pronoun does shave an antecedent
in P but it is not bound i.e it doesnot have a mukhya
as its antecedent. Hence, the sentence is grammatical.
In 161 (ii) , however, the pronoun has been assigned the
mukhya in P as its antecedent and the sentence
is ungrammatical. |
In (162), P is the possessive phrase uska
ghar. P is the embedded ka-na NP. In, P the
pronoun is antecedent-free. The P does not contain an
argument other than the mukhya, which can be assigned
as antecedent to the pronoun. Thus the pronoun is antecedent-free
in P also. Outside the P the pronoun is free to choose
any argument as its antecedent and the pronoun is bound
by the mukhya, ram. |
In (163), P is the phrase uska santulan.
P is the root S. The pronoun is antecedent-free in P.
In P it has been assigned the mukhya, ram, as antecedent
and the sentence is ungrammatical as predicted by the
theory. Notice that in this sentence even if uska
had not been bound by ram, it would not have been
possible to use a pronoun in the relevant position. This
is because here the pronoun gets neither intrasentential
reference (for reason stated above) nor does it get contextual
reference, which is ruled out by the semantics of santulan
kho dena. If uska has contextual reference
the sentence, then (163) would have the following meaning: |
X lost Y' s balance, where X is ram
and Y is us. |
But this is bizarre. However, an anaphor
is possible in place of the pronoun in the (163) because
their even though the anaphor is free in P, it is bound
by ram in
and the sentence is grammatical, as predicted by our theory. |
In (164), uska dost is the P in which
us is antecedent-free as expected. In ,
the root S, us has to choose the non-mukhya as
its antecedent. The non-mukhya in ram.
Hence us is assigned ram as its antecedent and the sentence
is grammatical, as predicted. It must be borne in mind
that no matter whether ram is the mukhya or not
in ram ki bat (it indeed is the mukhya).
Ram is not the mukhya of the predicate of
the ,
that is manana (obey). Notice that apna
cannot be used in place of uska in the possessive phrase
in (164). This is beause apna does not get an antecedent
in the sentence. Apna dost is the mukhya
of the predicate manana, that is, it is the mukhya
of
Within the
there cannot be another mukhya by which apna
can be bound. Hence, apna remains free and the
sentence is ungrammatical. |
In (165), P is the possessive phrase uski
dekhbhal and
is the root S. In P, the pronoun is antecedent-free, as
expected. The mukhya in ,
is ram ki ma and the non-mukhya is ram.
In ,
the pronoun can choose the non-mukhya as its antecedent.
Thus, in (165) us refers to ram. When apna
is used in placeof the pronoun, it correctly chooses the
mukhya of the sentence ram ki ma as its
antecedent. |
In (166), P is the possessive phrase uska
pita.
is the embedded S. In
uska pita is the mukhya and the non-mukhya
distinct from the pronoun is shyam. The pronoun
chooses shyam as its antecedent as permitted. However,
in (166), the pronoun can choose antecedents outside the
.
As stated earlier, outside the
the pronoun can choose its antecedent freely. Thus, in
(166), it chooses both the mukhya and the non-mukhya
of the root S as antecedents. |
Now consider the following sentences which
demonstrate the effect of semantic considerations on pronoun-antecedent
relation. |
(167) |
|
ram ne mohan se kaha ki
voh parne me j
j
ram CM mohan CM say+PAST COMP he study+Nom CM
man lagaye
attention put
(Ram asked Mohan to study with diligence.) |
(168) |
|
*ram ne mohan se kaha ki voh parne
i i
ram CM mohan CM say+PAST COMP he study+Nom
me man
lagaye
CM attention put+PAST |
(169) |
|
ram ne mohan se kaha ki voh apni
j
j
ram CM mohan CM say+PAST COMP he self's
kitab pare
book read+PAST
(Ram asked Mohan t read his book) |
(170) |
|
*ram ne mohan se kaha ki
voh apni kitab pare i
i
ram CM mohan CM say COMP he self's book read+PAST
|
|
|
In (167) - (170) the pronoun occurs in an
exptraposed S which is the P. P is the root S. In P, the
pronoun is antecedent-free. Notice that the P is a CP*
also. Beyond the CP*, the pronoun can choose its antecedents
freely. Thus, in (167), it chooses the non-mukhya
as its antecedent and the sentence is grammatical. However,
when it is assigned the mukhya, ram, as
antecedent the sentence (168) is ungrammatical. It is
because of the semantics of the verb kah (in the
sense of "advise"). If voh is coindexed
with ram, it gives (168) the following interpretation: |
X ADVISED Y that X concentrate on his studies
where X is ram and Y is mohan. |
This is anomalous. Hence, (168) is ungrammatical. |
In (169) and (170), P is again the extraposed
S and P is the root S. However, P is also CP*. Thus, outside
the CP* the pronoun can choose antecedents freely. When
it is assigned the non-mukhya, mohan, as
antecedent (169) is grammatical. However, when it is assigned
the mukhya as antecedent, as in (170), the sentence
is ungrammatical. As in the case of (168), assigning the
mukhya as the antecedent to the pronoun violates
the semantics of the verb kah (used in the sense
of "suggest"). If voh is assigned ram
as the antecedent, as in (170), the sentence has the following
totally anomalous ram interpretation: |
X SUGGESTED to Y that X should read X's book,
where X is ram and Y is mohan. |
Our conclusions need not be modified in the
light of the above. The theory can assign both, the mukhya
and the non-mukhya as antecedent to pronouns thereby,
assgning two interpretations to each sentence. One of
these can be filtered out later in the propmatic component.
|
|
|
|