Department of Linguistics
3125 Campbell Hall, UCLA
Los Angeles, CA
- Bryan Gick will be presenting our poster titled "Tonic activations in speech production" at the 178th Meeting of the Acoustical Society of America from December 2-6 in San Diego. You can see the abstract here.
- I'll be presenting a poster titled "Phonotactic learning with neural language models" with Max Nelson at the third annual meeting of the Society for Computation in Linguistics (SCiL) from January 2-5 in New Orleans. You can find the accompanying proceedings paper here.
- I'll be giving a talk titled "Conflicting trigger effects in Uyghur backness harmony" at the Workshop on Turkic and languages in contact with Turkic from February 8-9 at the University of Delaware. This work is in collaboration with Travis Major and Mahire Yakup.
I'm a fourth year PhD candidate in the Department of Linguistics at UCLA. I am a computational linguist, with a focus on the areas of phonetics and theoretical phonology. My research combines computation with a wide range of converging methodologies to explore questions of interest to the field. The questions that guide much of my research are (a) the extent to which linguistic structure is learned vs. innate; (b) how innate biases shape the learning process; and (c) how we can better understand the source and nature of these biases. I view computation as a tool that can be used to complement empirical and theoretical methodologies, testing claims and providing new predictions. I have used such methodology to probe the role of distributional information in phonological learning, to investigate the ways in which the biomechanics and motor control of the vocal tract shape our speech systems, and to explore the interaction between acoustic cue weighting and vocabulary acquisition, among other topics.
I also work extensively on the Uyghur language (Turkic: China). My dissertation provides a detailed empirical description of Uyghur backness harmony and investigates its learnability. Uyghur speakers asked to generalize backness harmony to wug forms do so in a way that is not consistent with lexical statistics. I show using tools from formal language theory that the surface pattern of Uyghur backness harmony exceeds the upper bound of complexity typically proposed for segmental phonology, and that the grammar learned by speakers is computationally simpler. I therefore suggest that a bias towards a certain level of computational complexity influences how phonological grammars are learned. This work attempts to bridge the gap between formal language phonology and constraint-based models. I have also performed descriptive research on this underdocumented language, including its intonation.
Please visit the Uyghur Human Rights Project to learn more about the deteriorating human rights sitatuation in East Turkestan/Xinjiang Uyghur Autonomous Region.