Agenda

Relatore: Steven Franks (Indiana University)

Abstract

This talk examines variation in so-called “Person Case Constraint” (PCC) effects, primarily in the South Slavic languages. After presenting core data (drawn from Runić 2013, Harizanov 2014, Stegovec 2016, and others) about which combinations of pronominal clitics are better than others in the various languages, I sketch an analysis which adapts Béjar and Řezáč’s 2003 “Person Licensing Condition” and Halle’s 1997 [±Participant] and [±Author] feature system. Clitics are defective lexical items in various ways—prosodically, syntactically, and semantically. I argue that clitic pronouns may be deficient in PERS features, the possibilities being for them to lack PERS, PART, AUTH, or both PART and AUTH. Each of these corresponds to one of the Slavic patterns—Strong, Weak, Me-First, and Strictly Descending, respectively. The clitics then receive these features by spreading from a higher functional category that establishes the sentence’s Point of View (e.g., Agr, Appl, or even PoV), in the spirit of Charnavel and Mateu 2015 or Pancheva and Zubizarreta 2017. I show how the different patterns straightforwardly derive from top-down association of these features, assuming the particular clitic deficiencies and standard intervention principles. Next I consider repairs to illicit combinations, where all repair strategies in some way involve elimination of the intervener. Under certain conditions reordering is possible, but alternatively one of the clitics can sometimes be replaced by its full/tonic counterpart. Interestingly, this latter strategy gives rise to pairs of solutions which differ in terms of functional sentence perspective. It is shown that the pragmatically neutral repair operates in a way which maximizes the sentence’s point of view specification, and that this follows from the general system developed to handle the observed variation in Slavic PCC systems.