Informations and abstract
Keywords: verbal morphology, paradigm structure, pluractionality, distributive dependencies, semantic data collection methodology.
The present study examines verbal morphology in Seri. Seri verbs have dedicated verb stems marking multiple events for singular and plural subjects respectively. However, the morphology marking the stem-forms is not transparent: neither subject number nor multiple events are associated with consistent exponents. As the exponents on the verb stems do not provide any clues to the structure of the paradigm, the features structuring the paradigm have to be inferred from syntactic and semantic properties. Syntactic cooccurrence restrictions with singular and plural subjects clearly distinguish singular and plural subject stems. Within singular and plural subject stems, the further distinction between the neutral and the multiple event form (multform) is based on semantic dierences between the stems: mult-forms are only felicitous in contexts with multiple events. As multiple event marking on verbs is not a homogeneous class, it does not trivially follow that singular subject and plural subject mult-forms express the same type of event multiplicity. To establish the paradigm structure of Seri verbs we therefore need to examine whether the mult-forms express the same semantic feature value across singular and plural subject stems. We rst show that plural subject multforms pass the same diagnostics that show that singular subject mult-forms mark event plurality. As a second step we compare the meaning of singular and plural subject mult-forms. In the initial elicitations younger speakers uniformly interpreted the singular and plural subject mult-forms as iterative with events distributed in time; in contrast, older speakers interpreted singular subject mult-forms as iterative but allowed simultaneous events distributed over a participant plurality for plural subject mult-forms. Elicitation with dierent materials showed that the initial dierence in truth-conditional judgements re- ects a dierence in the preferred contexts assumed by the older and younger speakers, not a dierence between the semantic range of singular subject and plural subject mult-marking as such. We therefore conclude that the semantic evidence supports an analysis of singular subject mult-forms and plural subject mult-forms as expressing a single type of event plurality marking.