Linguistics in Göttingen - A platform for empirical and theoretical linguistics

Krzysztof Migdalski (Wrocław): Cliticization patterns in Slavic and their relation to tense marking

Slavic languages exhibit a remarkable variation concerning patterns of pronominal cliticization. Czech, Slovak, Serbo-Croatian, and Slovene have second position (Wackernagel) clitics; in Bulgarian and Macedonian there are verb-adjacent clitics like in Romance languages, Polish has weak pronouns, while contemporary Russian and Ukrainian exclusively display strong pronominal forms. Historically, the distribution of pronominal clitics was more uniform, as in Old Church Slavonic relics pronominal clitics were predominantly verb-adjacent (see Radanović-Kocić 1988; Pancheva 2005), the way they still are in Bulgarian and Macedonian.

In this talk I will present an analysis of diachronic changes in the syntactic position of pronominal clitics in Slavic. On the basis of Old Serbian and Old Polish data, I will show that the loss of verb-adjacent clitics was contemporaneous with the loss of tense morphology of the aorist and imperfect and the adaptation of aspect marking as the exclusive means of characterizing temporal relations in these languages. In syntactic terms, I analyze the change as the loss of the TP projection, as a result of which clitics may not adjoin to T⁰ any more, as they still do in verb-adjacent clitic languages, but rather they target specifiers in the extended functional VP projections as XP-elements.

This talk will show that the loss of TP may have two potential repercussions: (i) the shift of verb-adjacent clitics to second position, attested in Serbian, Czech, and Slovene, and (ii) possibly at a later stage, the reanalysis of second position clitics as weak pronouns in Old Russian and Old Polish. I will interpret the process as a case of degrammaticalization: in comparison to clitics, weak pronouns are prosodically more independent; they exhibit more robust scrambling possibilities, and they do not need to be adjacent to each other or to a verb. Thus, these data provide empirical evidence against the idea of the irreversibility of grammaticalization (Haspelmath 1999) and universal directionality of language change.