COMPLEMENTATION AND CONSTRAINTS ON PRONOMINALS IN
RUSSIAN
Konstantin Kazenin
The paper deals with a very intricate case of non-complementarity of
anaphors and pronominals in Russian, previously not reported in
literature. Though it is well known that anaphors within infinitive
clauses can corefer both with the PRO and with the matrix subject
(1), it is surprising that the same two possibilities exist for
possessive pronominals (2):
(1) Otec (i) poprosil Ivana(j) [PRO (j) zabrat_ svoi (i,j) veshchi so
staroj kvartiry].
Father asked Ivan to take his(anaph.) things from the old flat.
(2) Otec (i) poprosil Ivana(j) [PRO (j) zabrat_ ego (i,j,k) veshchi so
staroj kvartiry].
Father asked Ivan to take his(pron.) things from the old flat.
The coreferentiality between _ego_(his) and _Ivan_ in (2) apparently
violates Principle B, because the pronominal is coindexed with PRO, thus
having an antecedent in its minimal clause (which is the local domain
for Russian pronominals). The purpose of the paper is to delimit the
scope of this phenomenon and to suggest a preliminary account for
it. The coindexing of a pronominal with PRO is precluded with subject-
controlled infinitives:
(3) Ivan(i) xochet [PRO (i) pogovorit_ s ego (j, *i) otcom]
Ivan wants
to talk to his
father
We follow the analysis of Babby (FASL VI), arguing for structural
difference between subject-controlled and not subject-controlled
infinitives in Russian: the former are bare VPs lacking the subject
position, whereas the latter subcategorize for a subject and assign it
the dative case (the central argument for this analysis comes from case
agreement of adjuncts). It turns out that the possibility for possessive
pronominals to be coindexed with the dative subject in their minimal
clause is not limited to infinitive complements. It is also attested in
a number of Russian constructions with overt dative subjects, e.g. with
predicates assigning _quirky_ dative case:
(4) Ivani (i) stalo holodno v ego (i,j) mashine.
Ivan-DAT became cold in his(pron.) car
Several other constructions with dative subjects are discussed in the
paper. The descriptive generalization is that it is a general property
of the Russian possessive pronominals to be able to corefer with local
dative subject. (This parallelism between not subject-controlled
infinitives and other constructions with dative subjects confirms
Babby_s analysis of the former in a crucial way.) The main questions is
how these data can be reconciled with Principle B. It is suggested in
the paper that Principle B in its proper formulation for Russian
should rule out not any antecedent c-commanding a pronominal in its
local domain, but only a subject antecedent which triggers agreement (by
rising in SpecAgrSP). Obviously the dative subject does not trigger
agreement. The same holds true for subjects of nominalizations, which
can be coindexed with a possessive pronominal e.g. within the object NP,
thus confirming our hypothesis:
(5) [pobeda Petra (i) nad ego (i,j) novym sopernikom]
victory of.Peter upon his(Pron.) new
rival
The revised version of Principle B for Russian predicts every instance
of the coindexing in question, and not any ungrammatical coindexing.
Some typological and theoretical problems with this solution are
discussed in the paper.
|BACK TO PROGRAM|