A unifying approach to Russian impersonal morphology

Abstract

Despite the ever growing number of contributions on Russian impersonals, there is no unanimity about what constructions are to be labeled “impersonal”. While generatively inspired contributions tend to exclude infinitive phrases (type cto nam delat’) from the impersonal domain (e.g. Babby 2010), infinitive phrases seem to be part of the very heart of the impersonal domain in the view of others (e.g. Creissels 2007: 23). Similar controversy is observed with regard to numeral phrases or negated existentials yielding “genitive subjects”. One of the reasons why there is no consensus about how to define impersonals even within one language lies in the opposition between formal (subject-centered) and semantic (agent-centered) approaches to impersonality (Siewierska 2008). The paper attempts to combine both lines of reasoning. It suggests that the impersonal domain in Russian is represented by a network of constructions which are all related to one another by the formal marker of impersonality on the verb (i.e., third person singular neuter) and by semantic and pragmatic characteristics (namely, deviation from prototypical subjecthood; Malchukov/Ogawa 2011). At the core of the impersonal domain are constructions which indisputably lack canonical subject features such as agency, referentiality and topicality (e.g., weather impersonals). In the heterogeneous group of impersonals with experiencers of different kinds (e.g. меня тошнит; мне повезло), there is a referential, but clearly non-agentive NP, and topicality is at least questionable. Constructions peripheral to the impersonal domain (e.g., numeral constructions, negated existentials) are actually more instructive for the case in point: 1. Погибло/погибли несколько человек: lack of topicality, semantically unaccusative verb  impersonal preferred 2. Несколько человек обедали/?обедало вместе: topical numeral phrase, semantically unergative verb  impersonal hardly available 3. У них не было детей: lack of referentiality and topicality  impersonal only 4. Пять хулиганов били/*било припаркованные автомобили: topical numeral phrase, transitive, highly agentive verb semantics  impersonal usually not available While some of these findings have already been brought up, however sometimes implicitly, elsewhere (e.g. Corbett 1983 on agreement; Glushan 2013 on quantitative genitive; Kuznetsova 2005 on po-phrases; Buncic 2014 on definiteness and genitive of negation), they have not been given due attention in research on impersonality. The assumption that the likelihood of impersonal morphology increases when semantic and pragmatic subject properties decrease will be tested against corpus data and the results of a survey. If the data confirm the assumption, a unifying account to impersonality in Russian can be formulated which integrates prototypical and less-prototypical instances of impersonals on formal, semantic and pragmatic grounds.