Marta Ruda

In this short contribution I suggest that Polish personal pronouns have two available representations: first and second person pronouns are PersPs, whereas third person pronouns are either PersPs or NumPs. This structural difference is responsible for the availability of not only definite, but also indefinite (including unspecific) readings of personal pronouns in Polish, regardless of their morphological complexity (i.e., both full and reduced forms can have different types of interpretations). This follows on the assumption that NumPs can be interpreted as property anaphora.

Pronominal clitics in South Slavic languages have been shown to manifest the strict/sloppy reading ambiguity effect. In this paper I examine Polish object pronouns from this perspective, observing that even though they are not clitics, they can still be compatible with the sloppy interpretation if the right type of context is provided. The data speak against an ellipsis-based approach, aligning with the view that the sloppy reading is not a viable test for ellipsis. I thus pursue an alternative analysis on which the strict and sloppy readings are associated with a structural difference in the composition of the pronoun (PersP vs. NumP respectively), offering along the way additional evidence pointing to the importance of pragmatic distinctions in investigations of the interpretive properties of different types of nominal elements. From a more general point of view, the discussion suggests that the empirical picture related to the sloppy interpretation is highly complex, making an investigation of a broader spectrum of languages and contexts indispensable for disentangling all the relevant factors and developing an optimal theoretical approach.

The paper offers a novel analysis of the impersonal construction marked with -no/to in Polish. Contra previous accounts, the -no/to verbal morphology is decomposed into two morphemes, -n/t, realizing an impersonal active Voice head and –o, the default spell-out of unvalued agreement features of finite T. The analysis is embedded within a wider set of assumptions about the composition of the extended verbal projection in Polish, including a second active Voice head in addition to Voice found in personal structures. This suggests that the inventory of Voice heads in natural languages includes not only two non-active heads (i.e. passive and middle), but also two active Voice heads (i.e. personal and impersonal). The distributional and interpretational properties of the construction, including Case-related behaviour in secondary-predication contexts, suggest that the impersonal subject is best analysed as a minimal pronoun, whose Case feature is unvalued/absent in the narrow syntax.