Date of Award

Spring 5-20-2019

Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Claire Bowern


Nepali presents with a complex case marking pattern in which ergative case is obligatory in perfective transitive clauses, disallowed in unaccusative intransitive clauses and copular clauses, and varies with the nominative elsewhere. Where ergative marking is variable, its usage correlates with a variety of semantic and pragmatic factors. The purpose of this investigation is to precisely delineate the grammatical domains for which ergative marking is variable and to provide a unified analysis of the semantic and pragmatic factors that correlate with its expression.

The study of pragmatic phenomena requires the implementation of multiple strategies for collecting language data. The data for this investigation come from four con- verging lines of inquiry: descriptions of the Nepali pattern in the literature, targeted elicitations with thirteen native speakers, the implementation of a grammaticality judgment survey in Kathmandu in 2016, and the analysis of a published corpus of spoken Nepali.

The analysis found ergative marking to be obligatory in perfective main clauses and variable in subordinate clauses. What appears to be active marking in intransitive clauses is analyzed as ergative marking in transitive clauses with covert objects. The only categorical split is the distinction between perfective and non-perfective verb forms. Every other association was found to be non-categorical.

These non-categorical associations include a positive correlation between subjects with inanimate reference and the expression of ergativity in common nouns, and a

negative correlation between first person pronouns and ergativity in the pronominal domain. This follows expected patterns of marking based on the types which are most frequent in discourse. Ergative marking is somewhat associated with highly individuated objects, but not with affected objects.

Ergative marking is positively associated with characterizing or individual-level predicates, kind readings, categorical propositions, and strong construals of quantifiers. There was no correlation found between ergative marking and agency or volitionality.

The unified analysis of these associations contributes to theories of Optional Ergative Marking and to optional case marking systems in general. The main claim is that the Nepali ergative marks an effector of the event described by the clause. This term refers to a participant which is implicated in enacting and effecting the event, but is not necessarily its main controller or instigator. As a component of the ergative case marking system, it has a pragmatic usage, implicating the subject as a participant in a prototypically transitive event. Aspects of this analysis contribute to the general theory of Optional Ergative Marking and its relation to argument proto-roles. Associations between the ergative and prototypical properties of a transitive event arise from the meaning of the ergative marker as an effector. This analysis also provides a straightforward explanation for the lack of volitional correlations in Nepali that we find in other languages with variable ergativity.

Other semantic and pragmatic features are associated with discourse prominence. These include the correlation with categorical propositions and characterizing predi- cates. Here the associations are attributable to general principles of semantic marked- ness. Variable ergativity represents the presence of pragmatic implicatures of various strengths. Gradient markedness oppositions can lead to the conventionalization of these associations into semantic entailments. This is demonstrated for English gender marking, the association between ergative marking and semantic properties of the transitive subject in Nepali, and the association between ergative marking and Nepali perfective verb forms.

Open Access

This Article is Open Access