Data-driven theory development

My research is empirically driven, making use of diverse methodologies, including corpus work, traditional field elicitation, and psycholinguistic experimentation. This work proceeds from a guiding principle that new data should always be a motivating factor in theoretical development.

My typical workflow starts from noticing something that seems to bump against existing theory, leading to good old-fashioned introspection, followed by corpus searching (if possible), and then experiments or fieldwork to test the predictions of existing theory. Usually, this is all in the realm of generative syntax, and the syntax/semantics interface.

Below, I elaborate on some areas of my research.

Research Areas

BerlinPotsdamer.jpg

Binding and Reference Resolution

What are the processes that go into resolving the reference of a pronoun or anaphor, beyond the ABCs of Binding Theory? Using methods ranging from contextual felicity and truth value judgement tasks, through to visual world paradigm eye-tracking, sorting out the various influences on determining the meaning of a pronoun or reflexive is a recurring theme in my work. Languages included in this work are English, Korean, Shona, Persian, and Turkish.

GrouseSkyride.jpg

Raising (and movement in general)

Since being a graduate student, I have always returned to questions around raising verbs, mainly in English. This has included the characterization of the tense and scope properties of English raising via Tree Adjoining Grammar, to an examination of whether Persian raising is A or A’ movement. This work has led me to thinking more about the far-less studied Copy Raising construction, as a potential explanation for some facts in Persian, inspired by a project I supervised on Turkish. I have also supervised work on the deletion of sentence-initial pronouns in English raising sentences.

Viaduct.jpg

Scope

In languages known to be scope-rigid, the investigation of scope can be used as a diagnostic for the syntactic positions of scope-taking elements like quantified arguments or negation. Using mainly truth value judgement task studies, I have worked on Japanese, Korean, and Persian, examining the position of negation, and parameters of verb movement in those languages. Most recently, I have been looking at the scope relationships between arguments and temporal expressions in Korean.

DresdenWallTAGTrip.jpg

Tree Adjoining Grammar

Aside from working in “mainstream” generative syntax, I have maintained an interest in Tree Adjoining Grammar, and specifically Synchronous Tree Adjoining Grammar, which assumes a parallel and connected derivation of what ultimately become the traditional PF and LF outputs. This allows for constraints on a syntactic derivation to be formalized in the semantics, where they may be less arbitrary than if they were expressed in terms of feature checking. This has covered analyses of raising and control, as well as Wh-movement and parasitic gaps in English. This work has also attempted to capture binding constraints in Korean and Mandarin, and a small fragment of the interaction between tense and demonstratives in Blackfoot.

HimselfMug.jpg

Variation in English Reflexives

Around the world, English shows considerable variation in its paradigm of reflexive (self) pronouns, both in terms of their forms and functions. My work in this area largely derives from corpus searches (including geo-located twitter data), examining issues such as the variation between himself and hisself in the USA, and changing patterns of use for themselves and themself, as referential singular they becomes more widespread. This work has also spread out into the neopronoun paradigms of English (e.g. emself, hirself, etc…), showing that different forms are used by different groups, usually as a mark of solidarity, but sometimes as a tool for scorn.

SeoulMetro.jpg

Gender Bias and Experiment Methodology

While it has been well-established for many years that speakers use gender biases as part of their toolkit for resolving pronoun ambiguities, it is only recently that laboratory studies in this area are examining whether the gender of participants also plays a role in their responses. Preliminary results from our lab suggest that participant gender is actually quite influential, even in areas where it is generally not expected to be, such as the computation of sloppy readings of pronouns under ellipsis. We have also replicated results that just the presence of gender stereotyped nouns in a trial item moderates the reading time at an Island in English. Given the known effects of gender stereotypes, is increasingly important to consider whether participant gender, and the maintenance or subversion of gender roles in experimental stimuli impacts results. This is quite timely, as Linguistics comes to grips with its history of very gender-normative (sometimes to the point of offensive) example sentences.