The Evolution of Verbal Agreement in Iranian Languages

International audienceVerbal agreement in Iranian languages is usually onefold agreement with the subject. It is canonically realized by a verbal ending to the stem. However, in a number of languages so called clitic “pronouns” fulfill the function. In this talk, we investigate various historical sources of agreement markers such as verbal endings, auxiliaries, and pronouns, with a focus on enclitic pronouns. We will explain the reanalysis of the latter on the basis of their discourse functions in Middle Persian texts. Enclitic pronouns encoding the subject in an ergative construction are comparable with verbal endings in that they both are the “lowest degree” of cross-referencing the subject. In hanging topic constructions where the enclitic pronoun resumes the topic, a bridging context occurs in the way that the clitic encoding the subject resembles the agreement marker with a null (omitted or pro-drop) subject. This paves the way for the grammaticalization of the clitic as an agreement marker. It also explains why other oblique enclitic pronouns (e.g. encoding objects) “do not make it” (Siewierska 1999), because such a bridging context does not exist for them. The use of enclitic pronouns as agreement markers is most famously illustrated by Sorani (Jügel 2009, Haig 2008, Samvelian 2007, Bynon 1979), but also frequent among languages of the central Iranian plateau and the Fars province (Jügel and Samvelian, in press), which opens up to various hypotheses for the dialectal development of Persian.We will apply our account to the use of enclitic pronouns in so called “compound verbs of experience” (Barjasteh 1983) in Modern Persian. Despite evidence for a cline of cross-reference ranging from pronominal realization to agreement marking in the history of Iranian languages, we will nevertheless argue in favor of maintaining both concepts as different phenomena