Past Event

“Two types of passive? Voice morphology and allomorphy in Classical Greek and Sanskrit”

Dr. Laura Grestenberger
(Concordia University)

Languages differ in whether or not their passives are compatible with > transitive input structures. While in some languages, the input to passives is the projection Voice that has merged an external argument (“high passive”, e.g., German, English, Modern Standard Arabic), it has been argued that other languages only have a lower (non-active or “middle”) Voice head that is incompatible with an external argument in the Spec of Voice and typically combines different non-active interpretations (anticausative, reflexive, passive), e.g., Modern Greek (Bruening 2013, Alexiadou & Doron 2012, Alexiadou 2013). This > classification leaves open the possibility that there are languages with both types of passives. It has long been known that there are two distinct types of passives in Classical Greek (CG) and Sanskrit: One in which non-active (“middle”) inflectional endings are used (1a, 2a), and one in which a specifically “passive” suffix appears close to the root and co-occurs with the inflectional voice morphology of the endings (1b, 2b):

Given that the passive stem-forming suffix co-occurs with the voice morphology of the endings, these languages could be interpreted as having two different Voice heads. I argue instead that only the voice morphology of the endings is an exponent of the head Voice, while the “lower” passive morphology is an exponent of the event projection v. This explains why “passive” -yá- in Sanskrit and -thē- in CG are 1) restricted to a particular tense/aspect stem (present in Sanskrit, aorist in CG), 2) in complementary distribution with other stem-forming morphology and 3) incompatible with transitive input structures. Building on this previous work, I assume that non-active morphology in Sanskrit and CG is spelled out when the Voice head does not merge an external argument in its specifier (e.g., Embick 2004, most recently Alexiadou, Anagnostopoulou & Schäfer 2015). If synthetic voice morphology in languages like CG and Sanskrit is assigned at Spell Out to the output of syntactic operations, we expect it to interact and co-occur with valency-changing operations like “passive formation” in CG and Sanskrit. I will discuss the implications of this analysis with respect to the voice allomorphy variation in the CG vs. Sanskrit “passive”: while the former triggers obligatory active endings (1b), the latter triggers obligatory non-active endings (2b). The data and analysis presented here provide further arguments for the view that passivization is parametrized and interacts with other (language-specific) morphosyntactic properties of the “event layer”.

About the Presenter

Dr. Laura Grestenberger received her PhD from Harvard University and now teaches at Concordia University. Her specialization is in morphology and syntax of classical Indo-European languages, both in synchrony and diachrony.

This event is sponsored by the School of Linguistics and Language Studies