Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

Author ORCID Identifier


Open Access Dissertation

Document Type


Degree Name

Doctor of Philosophy (PhD)

Degree Program


Year Degree Awarded


Month Degree Awarded


First Advisor

Brian Dillon

Second Advisor

Rajesh Bhatt

Third Advisor

Adrian Staub

Subject Categories

Cognitive Psychology | Language Description and Documentation | Morphology | Psycholinguistics and Neurolinguistics | Syntax | Typological Linguistics and Linguistic Diversity


This dissertation develops a formal and psycholinguistic theory of person-based prominence effects, the finding that certain categories of person such as "first" and "second" (the "local" persons) are privileged by the grammar. The thesis takes on three questions: (i) What are the possible categories related to person? (ii) What are the possible prominence relationships between these categories? And (iii) how is prominence information used to parse and interpret linguistic input in real time?

The empirical through-line is understanding obviation — a “spotlighting” system, found most prominently in the Algonquian family of languages, that splits the (ani- mate) third persons into two categories: proximate, the person who is in the spotlight, and obviative, the persons who are introduced into the discourse, but are not in the spotlight. I provide a semantics for the feature [proximate], and detail a lattice-based theory of feature composition to derive the categories related to obviation in Border Lakes Ojibwe and beyond. This leads to insights about the syntactic and semantic relationships between person, animacy-based noun classification, number, and obviation.

The novel contribution to the theory of person-based prominence effects is to de- compose person features into sets of primitives. This proposal allows the stipulated entailment relationships between categories and features, as encoded in prominence hierarchies and feature geometries, to be derived from the first principles of set theory. I further motivate the account by showing that it has increased empirical coverage, and apply it to capture patterns of agreement and word order in Border Lakes Ojibwe.

Finally, I present a psycholinguistic study on how obviation is used to process filler- gap dependencies in Border Lakes Ojibwe. I show that obviation, and by extension, prominence information more generally, is used immediately to predictively encode movement chains, prior to bottom-up information from voice marking about the argument structure of the clause. I argue for a modular and syntax-first model of parsing, revising the Active Filler Strategy to be guided by pressures to minimize syntactic distance and maximize the expected well-formedness of each link in the chain. These pressures compete, accounting for effects of prediction, integration, and reanalysis in long-distance dependency formation.