Academic research

How does knowing and speaking two languages (bilingualism) contribute to variability in spoken language comprehension and production? 

How do individual experience (e.g. language proficiency, use, and history) and social factors (e.g. individual traits, attitudes, identity, conversational roles) shape language processing behavior? 

CV_Shen.pdf

Academic CV (last updated in 2020)

Recognizing spoken code-switched words

How do bilingual listeners manage switching between two languages? 

Does processing spoken code-switched words take longer than processing unilingual (single language) words? I.e., is there a switch cost in auditory language processing? 

Figure A. This graph shows results from an eye-tracking experiment on the perception of code-switches. The curves represent Mandarin-English bilingual listeners' looks to images corresponding to unilingual and code-switched target words, and the facets show listeners' language dominance, from more Mandarin-dominant to more English-dominant. The higher and further left the curve, the more listeners looked toward the image, indicating quicker recognition of the target word. While Mandarin-dominant listeners experienced a processing cost for recognizing code-switched English words, English-dominant listeners experienced a processing advantage. 

Mandarin-English code-switching is not necessarily costly. For bilingual speakers in the California Bay Area, infrequent insertions, such as Mandarin words in English sentences, can be costly. But the costliness of frequent code-switches, such as English words in Mandarin sentences, depends on the listener's dominant language (Figure A, left).

Shen_LSA2020.pptx

(Susanne Gahl, Keith Johnson, and I have also shown that phonetic patterns preceding a code-switched word can help listeners anticipate Mandarin-English code-switching. (See our paper and past presentations below.) 

CUNY18_Poster_Shenetal.pdf

CUNY 2018 Poster

CAMP2017_Slides_Shenetal_short.pdf
ASA2017_Poster_AShen.pdf

ASA 2017 Poster

ISB2017_Slides_AShen.pdf

ISB11 Talk

Pronunciation of code-switched speech

What are the phonetic consequences of code-switching? 

Unilingual Mandarin utterances are characterized by pitch patterns that depend on the tone of the following syllable. Mandarin-English code-switched utterances have similar pitch patterns as unilingual Mandarin utterances: the same tone-specific pitch trajectories occur before Mandarin code-switched words, although the pitch of the switched word itself may be reduced (Figure B, right).

PHREND2019_Poster_Shen.pdf
Figure B. This graph shows the pitch measurements 500 ms before and after unilingual and code-switched words in unilingual and code-switched sentences produced by a Mandarin-English bilingual speaker. While code-switched Mandarin words have amplified tones and preceding pitch coarticulation relative to corresponding unilingual Mandarin words, code-switched English pitch is amplified relative to unilingual English pitch. 

Acquiring second language (L2) sounds

How effective are different training methods for learning new sound contrasts?

This collaboration examined how different phonetic training methods improved monolingual English speakers' productions of a Marathi sound contrast. Ultrasound training, in which the speaker was able to receive visual feedback of their articulations, resulted in more improvement than training using static mid-sagittal diagrams of the vocal tract. 

ICPhS_948_Proc_Linetal2019.pdf

Other projects

LabPhon2018_Poster_BLutzrossetal.pdf

Power and phonetic accommodation (LabPhon 2018)

PhonLab2016_Report_ChengShen.pdf

Ultrasounding Tswefap back consonants (PhonLab 2016 Report)

ICPHS0887_Proc_Kingstonetal2015.pdf

Is perception personal? (ICPhS 2015 paper)

SROP_Shen.pdf

Phonotactic probability (UCB SROP 2013 paper)