ИСТИНА

It is widely acknowledged that infinitival clauses in Russian exhibit intermediate behavior with respect to locality restrictions. In some cases, they pattern with finite embedded clauses in being opaque for syntactic interactions — e.g. infinitives constitute the binding domian for the reciprocal anaphor drug druga ‘each other’ and the disjoint reference domain for the pronominal (Rappaport 1986). In other cases, infinitives are transparent for syntactic processes. Thus, the binding domain of the reflexive sebja ‘oneself’ is larger than the reflexive’s governing category (i.e. the minimal noun phrase or clause containing an accessible subject) and corresponds to the minimal finite clause containing the reflexive. Infinitival complement clauses easily allow wh-extraction, which is much more restricted in finite embedded clauses (Khomitsevich 2007, Antonenko 2010, Bailyn 2020). At the same time, there are significant discrepancies among the infinitives themselves. Putting aside aspectual and modal constructions with infinitives (which exhibit multiple properties of functional restructuring), we observe that subject and object control infinitives behave differently with respect to many diagnostics. Franks (1995, ch.6) and Babby (1998) report that in subject control infinitives, floating quantifiers (or PROs which they are construed with) agree with the matrix subject, whereas in object control infinitives, floating quantifiers (or PROs) are invariably dative. The same pattern is found with primary and secondary predicate adjectives (Nichols 1981, Franks 1995, Madariaga 2007): the agreement options available in subject control infinitives lack in object control infinitives. The aim of my talk is to examine differences between subject and object control infinitives in Russian for yet another phenomenon usually restricted by clause-level locality — licensing of negative concord items. It turns out that negative concord is licit across the infinitive’s boundary but illicit across the finite clause’s boundary. Interestingly, negative concord in subject control infinitives is rated much higher than in object control infinitives; yet, the latter is still acceptable (Lyutikova & Gerasimova 2021). To account for these oppositions in respondents’ ratings we need a three-way distinction between subject control infinitives, object control infinitives and finite embedding. In my talk, I put forward the hypothesis about the structural difference between subject and object control infinitives explaining their behavior with respect to negative concord; I also show that this hypothesis can be successfully applied to account for other differences between subject and object control infinitives outlined above.