One of the main puzzles that drives linguistic research is the following: why is the set of attested languages only a small subset of the set of languages that could exist, given a naïve theory of the potential for linguistic variation?  My research explores this question in the domain of phonology, seeking to understand what generalizations unite the set of attested phonological systems, and what these generalizations reveal about the nature of the phonological grammar.  I am interested in questions regarding the structure of the phonological grammar (e.g., what are the right constraints?) as well as how the grammar interacts with the outside world (e.g., how do conditions of learnability constrain the attested typology?).  My work draws from a range of empirical domains and employs a number of diverse techniques, from large-scale typological surveys to quantitative analysis of data patterns in individual languages.

You can learn more about my research by accessing a list of publications and a list of presentations.  Descriptions of topics I have worked on, with links to representative publications and presentations, are below. For a more in-depth summary of several current projects and their relationships to one another, see my research statement.


Constraints on the distribution of nasal-stop sequences

My dissertation research explores the role of contrast in explaining the cross-linguistic distributional properties of nasal-stop sequences. The dissertation is based on a combination of original large-scale survey work and more detailed analysis of language-internal phenomena, taking into account evidence from the distribution of phonemic nasal-stop sequences (with respect to different positions in the word, different nasal vs. oral vowel qualities, and other nasal-stop sequences) as well as the distribution of allophonic nasal-stop sequences.  Based on the results of these investigations, I argue that constraints on contrast (Flemming 2002) provide the only accurately restrictive characterization of the resulting set of generalizations. Overall, the findings form a unified, substantial argument that constraints on contrast are an essential part of speakers’ phonological grammars.

Learnability and typology

In Optimality Theory, typological predictions of a constraint set are typically assessed by exploring its factorial typology, equivalent to the set of systems generated by each possible ranking of constraints. Much recent work in phonological theory, however, has begun to investigate the hypothesis that learnability also plays a role in shaping typology: the range of patterns that can be accurately and reliably learned. To date, my work in this area has argued that the absence of systems exhibiting Kager’s (2012) midpoint pathology is not due to the absence of constraints that could generate them, but rather due to difficulties that they pose to the learner.

(Non-)myopia and its Implications for the Grammar

Unbounded spreading processes are often claimed to be myopic: the ability of some feature [F] to spread from some segment z to some segment y does not depend on its ability to further spread from y to x. Recent work has however cast doubt on the uni- versality of this claim, by showing that non-myopic processes exist. My contribution to this body of work argues that a particular kind of non-myopic process, trigger deletion, is attested in Gurindji (Pama-Nyungan, McConvell 1988): when the spreading domain contains a certain kind of blocking segment, the spreading trigger deletes. I argue that in order to capture the Gurindji pattern, as well as the extant typology of non-myopic processes, any successful analysis of unbounded spreading must allow surface candidates to be globally evaluated. In the future, I am interested in figuring out why myopic processes are far more common than non-myopic ones.

Theories of Dissimilation

Recently, I have become interested in understanding (i) how typological predictions of different theories of long-distance dissimilation differ from one another, (ii) what properties of the theories lead to these differences, and (iii) which set of predictions best characterizes existing generalizations in the typology of long-distance dissimilation. I have begun initial investigations into these questions within the domains of blocking and similarity in dissimilation, and plan to both further these investigations and extend them to new domains (e.g. locality) in the future.

Prosodic Misapplication in Copy Epenthesis

The term copy epenthesis is used to describe a class of patterns in which the properties of an epenthetic vowel (the copy) are dependent on the properties of one of its vocalic neighbors (its host). With Sam Zukoff, I am investigating the nature of the constraints that regulate the relationship between the copy vowel and its host. We argue that a class of prosodic misapplication effects in copy epenthesis necessitate the introduction of a correspondence relation between copy vowels and their hosts. In future work we plan to explore the implications of our proposal regarding certain parallels between the typologies of copy epenthesis and reduplication.

Stress-Morphology Interactions

In many languages, the stress pattern of a complex word resembles the stress of its morphological base, or another related word. A classic example comes from the English word orìginálity, which resembles the accentual profile of its immediate subconstituent oríginal (cf. mono morphemic Mèditerránean). My work has addressed the following question: what kinds of constraints promote this identity? Based on a typology of stress-morphology interactions in Australian languages, I have argued that output-to-output constraints provide the most accurately restrictive characterization of the overall typology. With Donca Steriade, I am investigating the extent to which extensions of this theory can provide us with a unified account of stress in English suffixal derivatives.

A’-Movement in English

It is well-known that A’-extractions in English share a large number of basic properties (e.g. ability to license parasitic gaps, ability to strand prepositions, etc.). My work in syntax has explored lesser-known properties that are not shared by all A’-extractions (e.g. potential for massive pied piping, weak crossover effects, antipronominality effects) and asks why these divergences exist. I have argued that limited application of Wholesale Late Merger is responsible for differing sensitive to antipronominality effects among different classes of A’-extractions.