THE DISTRIBUTION OF BARE NPS IN RUSSIAN: A DYNAMIC ACCOUNT

Abstract

Bare nominals in many articleless languages present an interesting case within the typology of (in)definiteness. On one hand, they pattern with indefinites in that they can be used to talk about a discourse-novel individual that non-uniquely satisfies their descriptive content. At the same time, they pattern with definites in that one cannot felicitously use two identical bare nominals in the same context to talk about different individuals. This behavior of bare nominals is challenging for the theory of nominal meaning, as the null hypothesis is that such nominals, lacking any (in)definite counterparts, should freely be used in definite and indefinite contexts. In this dissertation, I provide a novel account for this phenomenon. Developing my analysis from a case study on Russian, I argue that the null hypothesis is in fact right, and semantically, bare nominals are neutral in terms of (in)definiteness (or, alternatively, they are indefinites that do not convey any “anti-definite” content); the observed definite-like behavior of bare nominals is due to an independent discourse-level principle that I call Unambiguity Constraint: a nominal cannot be used to introduce or pick out a discourse referent i if there is a familiar discourse referent j that maps onto another satisfier of that nominal. I show that this constraint not only manages to explain the observed “definiteness effects” but also makes independent predictions leading to novel generalizations about the distribution of introductory and anaphoric uses of bare nominals. I also compare my analysis with previous approaches to bare nominals, arguing that it manages to balance out between different extremes that make the latter either underpredict or overpredict the distribution of bare nominals. Finally, I show that my analysis has a cross-linguistic potential: judging from preliminary data, bare nominals in at least five more articleless languages are compatible with it. I also discuss challenging cases of indefinites that can be used twice in the same sentence to talk about different individuals, and argue that with certain assumptions made, the UC-based approach can account for them as well.