I am an Assistant Professor of Instruction in the Department of Linguistics & Cognitive Science at the University of Delaware. My research interests lie in the area of argument structure, which I explore using a variety of theoretical and experimental methodologies. In particular, I focus on questions about how verbs’ and their arguments’ meanings relate to the syntactic structures they occur in. I address these questions with both generative models and experimental evidence. You can find my most recent CV here.

With Bob Frank, I am investigating the performance of computational language models on a variety of psycholinguistic tasks. Computational language models have achieved a high level of success on a variety of linguistic tasks, but it is still unclear how similar the knowledge they demonstrate is to humans’ knowledge of language. We find that although such models can often get the “right” answer, deeper probing reveals that they make different errors from those people characteristically make in psycholinguistic experiments. This shows that despite surface-level similarities between people and computational language models’ linguistic behavior, there are still deep differences in how this knowledge is represented that remain to be addressed if we aim to use such models to understand how people represent language. Some of our work was recently accepted for publication in Transactions of the Association for Computational Linguistics, as well as the BlackboxNLP and GenBench 2023 workshops.

In my dissertation, I investigated the syntax and semantics of verbal argument structure, focusing primarily on spray/load verbs and related verbs. These verbs are defined by their ability to appear in two distinct syntactic frames, as demonstrated by pairs of sentences like I sprayed the wall with paint and I sprayed paint onto the wall. I find evidence from readings of again with these verbs that supports a syntax that allows phrases to belong to more than one mother node (a syntax that permits multidominance). I have been following up on this research in different ways with Bob Frank and Tom Roeper. I also investigated interesting patterns found with non-agentive uses of these verbs, which I argue can be derived by P-conflation with the verb root.

Somewhat less recently, I have done experimental work on the English dative alternation, and formal work on the syntax and semantics of adjectival passives, as well as on an aspectual (eventive/stative) alternation shared by object experiencer, location, and govern-type verbs, which I argue is a kind of double-object causative-inchoative alternation. I have also collaborated with Rong Yin on the syntax and semantics of argument and adjunct coordination in Spanish.

I have also worked on the acquisition of recursion with a large research group headed by Tom Roeper, and have begun collaborating with him and Shota Momma to examine priming of recursion in adult speakers. For the acquisition side of this project, I developed an experiment to test the acquisition of possessive recursion in English, which was adapted and run in Mandarin by Daoxin Li and colleagues.

Email: mawilson@udel.edu

GitHub: github.com/mawilson1234