Advances in the Syntax of DPs - Structure, agreement, and case

(ff) #1

56 Steven Franks


Nadja in TP2, even though the fact the Nadja is nominative only becomes available
when Tfin is merged in TP1. Note however that the feature sharing model raises famil-
iar issues about timing and how predicate adjectives get case. Nadja in (68) only has
its nominative features valued upon introduction of TFIN into the structure, so the
semipredicative sama can agree with Nadja locally but cannot be sent to PF until it has
its case features set. What this means is that Spell–Out must be delayed until the end
of the derivation. Given the nature of co-valuation this seems inevitable in a bottom-
up model of grammar, and in Franks (2014) I argue for late Spell–Out on completely
different grounds.
In an SD situation, such as (7d) repeated in (69), samim again shares features
with the subject, this time PRO. I assume that PRO comes specified with phi fea-
tures and that NPs must match in phi features in order to be construed as corefer-
ential. In this instance, PRO is plural in order to be interpreted as ‘we’. PRO, unlike
a lexical subject, lacks case features. Instead, these are directly valued on samim by
C, as in (70).
(69) Dlja nas utomitel’no [CP [PRO delat’ èto samim]].
for us exhausting do.inf this self.pl.dat
‘It is exhausting for us to do this on our own.’
(70) [CP C [TP2 TINF got ovit’ ]]

[SAM]

[PHI=] [CASE=]

[PRO]

Here the SD agrees in phi features with PRO but has its case valued only upon merger
of C into the structure.
This study of predicate adjective agreement and the second dative construction
in Russian is a work in progress. I have considered various alternatives and weighed
their assets and liabilities. The overgeneration problem – the need to prevent spuri-
ous assignment of dative to OC semipredicatives and to ordinary adjectives – led
to an approach which rejected agreement with a dative PRO in favour of direct
assignment from C to the semipredicative. This raises more general questions about
the features of PRO and whether the theory of grammar should countenance case-
marked PRO at all.
Free download pdf