UC Santa Cruz

The central claim of this dissertation is that there is more variation than previously recognized in the types of plural events that verbal predicates can denote. To make this argument the dissertation presents a detailed description and analysis of a series of pluractional suffixes in the Mayan language Kaqchikel that derive verbal predicates that cannot be satisfied in single-event scenarios. The guiding theoretical question is to determine the relationship between pluractionality and better understood semantic phenomena such as nominal plurality and quantification. Based on original fieldwork on Kaqchikel, I argue for a broad three-way distinction between pluractional affixes, where: (i) the first class generates pluractional predicates with denotations similar to a formally distinct subclass of group nouns like grove, bouquet, horde, etc., (ii) the second class generates pluractional predicates similar to bare plurals, and (iii) the third class generates pluralities similar to those that are introduced when interpreting a quantifier like every. I further argue for a morphosyntactic spilt between the first type of pluractional affix and the latter two. The former composes with verb roots directly before they undergo cumulative closure, while the latter two apply to verb stems that have been cumulatively closed.

While pluractional verbs have types of plural reference familiar from the nominal domain, verbs in general establish plural reference in different ways than nouns, which follows from the fact that events, unlike individuals, are individuated in terms of their participants and spatiotemporal location. I argue that to individuate the events that constitute an event plurality, pluractionals structure the way that a spatiotemporal trace function or theta-role function maps an event to its domain. In the case of the first two pluractional affixes, the temporal trace is crucial. In the last case, the pluractional targets a verb's thematic role. By structuring the relations between events and functionally related domains, the Kaqchikel pluractionals are semantically similar to Krifka's (1986,1992) incremental theme role or certain distributive adverbial modifiers in languages like English and Japanese (Henderson 2009 and Nakanishi 2007, respectively). By separating out the trace-based individuation requirements, not only is it is easier to see that there are fine-grained similarities between types of plural nominal and verbal reference, but differences between pluractional verbs and their nominal counterparts are accounted for.

Along the way, the dissertation develops novel analyses of a series of unexplained phenomena in Kaqchikel and English. In support of the analysis of the third class of pluractionals, I develop an account of dependent indefinites in Kaqchikel and other languages that explains why they are licensed by being interpreted both in the scope of a quantifier and as an argument of certain pluractionals. The analysis not only permits a better understanding of dependent indefinites in Kaqchikel, but it clarifies their place in a crosslinguistic typology of similar expressions (Balusu 2006; Choe 1987; Farkas 1997; Farkas 2002; Yanovich 2005, among others). Finally, I argue that insights from the analysis of pluractional affixes in the first part of the dissertation help explain the behavior of so-called pluractional adverbials in English, such as one by one or house by house (Beck 2007; Henderson 2009). While pluractional adverbials are different than Kaqchikel pluractionals due to their close connection to verbs of scalar change (Hay 1999; Kennedy 2008), they are similar in that they derive plural event predicates by structuring the relationship between the event argument and a functionally related domain, in this case, a domain of degrees. The proposal is that these adverbials fix the unit along which the progress of a verb of scalar change is measured, and by requiring at least two such increments, the modified predicates can only be satisfied by plural events.