Current and ongoing research projects:
Analysis of stylization through extreme variation
My dissertation uncovers the nature of "everyday" speech. In it, I analyze two weeks worth of self-recordings conducted by one young California female speaker as she goes about her everyday life. Focusing in the sociophonetic domain for now, I assess the entirety of this individual's speech production over this time, looking for evidence of systematicity and stylization. Labov (1972) has said that "the grammar of a speech community is more regular than the behavior of the individual" (p. 109). For several generations now, I and other variationists have taken this assumption to heart, conducting large-scale community studies aimed at understanding and documenting the systematicity of language in broad strokes. While the field has benefitted tremendously from these efforts, I think we are now at a point in time technologically and analytically where we can and should test the assumption that individuals are less systematic than communities. To what extent is the speech of the individual systematic? Is there actual evidence of a "vernacular" speech style if we have all the data from a speaker's everyday production? And, if a vernacular can be documented, what about deviations from it? Can stylistic variation be traced to extreme tokens outside of the vernacular "baseline"? How systematic is stylization anyway? My dissertation continues what my second qualifying paper started. Based on the observation that stylistic variation often occurs on the extreme edges of a speaker's normal repertoire, I am developing quantitative methodologies for identifying moments of stylization in the everyday idiolects of speakers. A pilot phase of this study examined a prototypical jock and a burnout from Penny Eckert's famous Detroit-area ethnographic study (presented at NWAV 42). In my qualifying paper, I conducted distribution analysis on the vowel productions of speakers from the Voices of California corpus, which I presented at the 2015 LSA annual meeting. Now, I am using a large amount of data from one speaker to search for not only stylization but also the existence of a vernacular "baseline."
Acoustic analysis of Nisei style in the Japanese American community of Livingston, California:
Together with Annette D'Onofrio, I am investigating the speech of Japanese Americans from a long-standing central California community. In particular, we are looking at the vowel production of speakers who identify as Nisei, or second-generation Japanese Americans. As young adults, the Nisei were held in internment camps during World War II, and are purported to have developed a distinctive speaking style. The question of whether this was an intermittent style or a full-fledged ethnic variety has yet to be examined quantitatively. Importantly, the Nisei are native speakers of English, knowing little to no Japanese. Thus, whatever features may characterize this style cannot be attributed to Japanese interference. In this study, to be presented at NWAV 44, we acoustically analyze relative monophthongization of the FACE and GOAT vowels for these speakers, comparing them with White speakers from a neighboring community. We find significant differences between ethnicities but also between genders, suggesting these features are indeed socially conditioned for this generation. In future work, we will do cross-generational comparisons to see these features persist.
Acoustic analysis of the front vowels in inland California :
Together with Annette D'Onofrio and Teresa Pratt, I am investigating the character and dynamics of the front vowels FLEECE, KIT, FACE, and DRESS in inland, non-urban California English. The California Vowel Shift (CVS) has been characterized as a chain shift involving the lowering/retraction of the KIT and DRESS vowels to follow TRAP, but this pattern has not been documented in any large-scale dialectology study. In previous work, we documented the retraction of TRAP across the state. Now, we assess whether the front lax vowels KIT and DRESS are indeed lowering/retracting to follow TRAP. Indeed, we find significant retraction, but no lowering of the lax vowels. Moreover, this pattern emerges for the tense vowels as well. This is suggestive of an overall front-back compression of the vowel space, not necessarily a chain shift.
Acoustic analysis of sibilants in inland California:
Together with Rob Podesva, I am analyzing the sibilant /s/ in the speech of white speakers from Redding, California in particular, but also from Merced and Bakersfield, California. Not only are we attempting to refine acoustic measurement methodology for these consonants, but we are also seeking to ascertain how these features are used by speakers with "country" orientation vis-à-vis those with "town" orientation, as well as considering the other social factors of sex, age, and education. An important focus of our work is how these country/town identities intersect with sexuality and gender identity. In two recent papers, we lay out our findings thus far. We are currently preparing an additional manuscript that outlines the complex array of linguistic factors that structure this variation as well.
Acoustic analysis of the low-back vowels in inland California :
Together with Annette D'Onofrio, Penny Eckert, Rob Podesva, and Teresa Pratt, I have investigated the character and dynamics of the low/back vowels TRAP, LOT, and THOUGHT in inland, non-urban California English. While the low-back merger has long been considered to be a flagship feature of California speech, the exact character of the merger is up for debate. Additionally, with TRAP-retraction (backing and lowering) occurring for California speakers who participate in the California Vowel Shift (CVS), there is some question as to which vowel results as the lowest in the vowel space. In a recent panel at the 2015 American Dialect Society annual meeting, and in a forthcoming volume on Voices of the West (Publication of the American Dialect Society), we show how low vowel shifting and configuration patterns vary across California, between the coast and the Central Valley, and then again between the North Central Valley and the South Central Valley.
Longitudinal morphosyntactic analysis of African American English:
For several years now, much of my research focus has been analyzing the development of African American English (AAE) in the speech of children. The data come from a longitudinal corpus housed in the Frank Porter Graham Child Development Institute. Independently, and in partnership with Walt Wolfram, my work has focused largely on conducting morphosyntactic analysis of variation in the speech of children over the early lifespan. Primarily, I have used a Dialect Density Measure (DDM) to count the relative number of AAE features per utterance of speech, as a means of operationalizing "vernacularity" for these speakers at different points in the lifespan. Using the DDM as a variable in regression analysis has been useful in associating the use of AAE with other social measures such as academic achievement, mental health, ethnic identity, demographic information, and more. In addition, I've conducted dyadic analysis between the children and their peers' speech, to appear in a forthcoming article. Recently our team has successfully collected post-high school interviews with the 60+ speakers remaining in the sample, adding another time point to the longitudinal analysis. After presenting in the so-called "PANEL panel" at this past Methods in Dialectology XV conference, Walt Wolfram and I are contributing a chapter to a special volume on the topic, to be published in 2015.
Analysis of tense/aspect in African American English:
In addition to morphosyntactic analysis, the Frank Porter Graham corpus of longitudinal African American English lends itself nicely to other analyses as well. For Charlie Farrington and me, it's a great place to examine the development and use of tense/aspect features particular to AAE. At this point, we are working primarily with aspectual be, which is difficult to study in a traditional variationist paradigm. At the 2014 LSA meeting, we presented a construction- and semantic-based analysis of aspectual be using the FPG corpus. Our work was then picked up by Brit Peterson for a piece in The Boston Globe's The Word column ("Linguists are like, 'Get used to it!': Why a new way to quote people has taken English by storm"). Even better, that article was subsequently iconized by Randall Munroe in an xkcd comic!
Acoustic analysis of consonantal variation in ethnic American Englishes:
I've also long been concerned with how ethnic identity is manifest in phonetic production, particularly in lesser-studied features such as consonants. Particularly, I've examined the consonant /l/ in two ethnic communities: African Americans from central North Carolina and Mexican Americans from Southern Texas. I've been primarily concerned with refining measurement methodology for /l/ as well as examining what, if any, change this feature is exhibiting over apparent time in these communities, and examining neighboring European American communities as well to see the extent to which the feature is parallel, divergent, or convergent over time. Particularly with the African Americans my apparent time approach has been longitudinally extensive, including the analysis of ex-slave recordings in addition to living speakers. In addition to /l/, for the Mexican Americans, Erik Thomas and I have investigated these patterns in other relevant consonants, including /r/, VOT, and /ð/. Thus far, this work has shown how consonantal variation can be powerfully indicative of ethnic identity and changing perspectives in communities over time. To this end, Erik Thomas and I have written a chapter on consonantal variation in Chicano English, to appear in a forthcoming volume on Latino English (also edited by Erik Thomas).
Acoustic analysis of voice-onset time in Korean stops:
Together with Rob Podesva, Annette D'Onofrio, Jeremy Calder, Eric Acton, Hsin-Chang Chen, Sam Bowman and Benjamin Lokshin, we are investigating the effects of linguistic and social factors on phoneme categorization of Seoul Korean stops. We are seeking to see where and how acoustical manipulation of voice-onset time (VOT) for Korean stops affects the perceptual determination of either lenis or aspirated stops. So far, our experiments with 30 native speakers of Seoul Korean, all of whom listened 10 speakers' productions of /CVn/ tokens which had been manipulated incrementally for voice-onset time, show that the listeners are significantly likely to use social factors as well as phonetic cues to categorize stops as either lenis or aspirated, to differing degrees. Therefore, it so far appears as though social information is as important as phonetic information in perception for these features.
Syntactic structure of the Lithuanian and Kazakh noun phrases:
For my first qualifying paper, I focused on the syntactic structure of the Lithuanian noun phrase. Specifically, I investigated how the movement and extraction properties of prenominal elements in the noun phrase may or may not be direct extraction, or copy/movement with variable PF interpretation. Some linguists of Slavic languages have posited a link between a language's possession of definite articles and the inability to extract prenominal elements in that language. Lithuanian shares a lot of similar properties to Slavic languages, including the ability to extract prenominal elements. However, Lithuanian differs from most Slavic languages in that it has a definiteness marker. With this paper I was able to show that extraction phenomena in Lithuanian have little or nothing to do with the existence of a DP, but rather have everything to do with contrastive marking on elements of the noun phrase. My superivising committee for this project were Vera Gribanova, Tom Wasow, and Penny Eckert. Together with James Collins, I've added to this research by investigating similar phenomena in Kazakh noun phrases. This work has extended into an in-depth analysis of the Izafet construction and regular possessive constructions in Kazakh noun phrases.
English as a world language:
Not only is English's power, for better or for worse, clearly evident in the legacy of British colonization worldwide, but in the last centuries, we've seen English emerging as a lingua franca even in countries with no colonial connection to the language. And, in some cases, it is no longer a lingua franca as much as it has become an alternative national language. To what extent is koineization occuring in such places, where English is no longer spoken as a foreign language/lingua franca, but is instead exhibiting systematic features and variation characteristic of these speakers and their heritage languages? Particularly as it is manifest in the European context, this question has fascinated me for several years now. My work on the English spoken by Lithuanian learners is an attempt to see how and when this systematization occurs, if at all, and what underlying interactions with the native language affects the process.