Academic journal article Journal of Slavic Linguistics

5. Argument Small Clauses

Academic journal article Journal of Slavic Linguistics

5. Argument Small Clauses

Article excerpt

In this section I turn to the question of why in argument small clauses with null predicators such as (41) the predicate must receive instrumental case, regardless of the matrix verb's tense.

(41) Dima scitaet Mis-udoktorom/ *doktor/ *doktora. Dima considers [Misha.sub.ACC] [doctor.sub.INST]/ [doctor.sub.NOM]/ [doctor.sub.ACC] 'Dima considers Misha a doctor.'

We ask: Why is instrumental case possible in (41)? Why is nominative or accusative impossible there? The answer to the first question falls out quite naturally from the current proposah in small clauses, the run-time for the argument introduced by PredEv is provided by the aspect of the matrix verb (42), making PredEv possible.


Since the run-time of e is given by the matrix Asp, it is the same as the run-time of 'consider', which corresponds to the intuitions of native speakers.

Let us now see why nominative and accusative predicates are disallowed in argument small clauses. Recall that in Nom-Nom constructions the nominative case on the predicate appears as a result of Multiple Agree. However, T is too far: it is separated by two phase-marking heads--v in "consider' and the Pred (43).


Crucially, the v in 'consider' cannot check accusative case on the small clause predicate via Multiple Agree either. Recall that T can enter in Multiple Agree with two goals because its phi-features, though valued, are not erased until the spell-out of the phase (CP) (Pesetsky and Torrego 2001). Now, v, unlike T, is a phase-marking head (Chomsky 2001). Once v values its phi-features by agreeing with the small clause subject, the phi-features are immediately erased, rendering v incapable of performing another deletion operation.

Importantly, v can delete several case features in a single casechecking operation if all of the case features are contained within one complex DP/NP. That is, once a relation between the probe v and a goal G is established, v can delete several active case features within G (in parallel, following Hiraiwa 2001), unless there is some goal-internal probe to check these features. This kind of Multiple Agree is independently needed in order to ensure successful case checking on complex DPs, conjoined DPs as in (44a), or on DPs with AP modifiers as in (44b). (24)

(44) a. Ja znaju doktora Ditou/ Dimu i Misu. I know [doctor.sub.ACC] [Dima.sub.ACC]/[Dima.sub.ACC] and [Misha.sub.ACC] "I know Doctor Dima/Dima and Misha.'

b. Dima videl tolstuju kosku. Dima saw [fat.sub.ACC] [cat.sub.ACC] 'Dima saw a fat cat.'

With this in mind, let us return to argument small clauses. In (45), once Agree between v and the DP Misa is established, v's phi-features are valued and erased. It cannot delete case on the predicate 'doctor' because 'doctor' is nota part of the same DP as 'Misha'. Hence, accusative predicates are impossible in small clauses with null predicators.


Since neither the matrix v nor the matrix T can license case on the small clause predicate, PredEv (or some other case-licenser) must be present.

A reviewer rightly points out that the semantic interpretation of the small clause predicate is not easily construed as eventive. However, since an event can be arbitrarily long or short (cf. the discussion in section 3), the non-eventive feel of these sentences is not surprising. The meaning of the instrumental predicate in the small clause in (42) is not very different from that of the instrumental predicate in the copular construction in (1). However, the eventiveness of the latter construction is clear when we compare it with the corresponding NomNom construction (2) in the past tense. Since such a comparison is not possible in the argument small clause--accusative and nominative predicates are out in these situations--we are left without a comparison. (Small clauses with an overt predicator za or kak have a somewhat different meaning, by virtue of having the overt predicator, which makes them bad minimal pairs. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.