Research

Enhancing linguistic research through critical use of race and ethnicity information

The discipline of linguistics has been rightly criticized for its undertheorized application of race and ethnicity. A survey of methods textbooks indicates that scant attention is paid to appropriately collecting and using race and ethnicity data in training linguistic researchers. This piece provides researchers with tools to reflect on their own study designs, reflect on their own responsibility to participants and communities, and design study prompts that allow more nuanced representation of race or ethnicity. In this article, my colleagues and I define race and ethnicity and explain their relevance for linguistic research. We discuss study design considerations and offer specific recommendations for the solicitation and use of race and ethnicity data for avoiding social harm, providing appropriate levels of generalization for theory building, and improving descriptive accuracy, especially for undersampled groups. We consider issues germane to collecting self-identified ethnicity and race information in a range of study types spanning multiple subfields of linguistics. We give concrete examples of questions that may arise in planning studies in computational and corpus-based linguistics, formal linguistics, experimental linguistics, and qualitative linguistics. We speak to ethical considerations, including the importance of using locally-constructed labels, analyst positionality, and respect for communities.

References: Squizzero, Robert; Horst, Martin; Wassink, Alicia Beckford; Panicacci, Alex; Moroz, Anna Kristina; Conrod, Kirby; & Bender, Emily M. (Under review). Enhancing linguistic research through critical use of race and ethnicity information. Language.

Squizzero, R., Horst, M., Wassink, A.B., Panicacci, A., Jensen, M., Moroz, A.K., Conrod, K., and Bender, E.M. (2021). Collecting and using race and ethnicity information in linguistic studies. University of Washington Working Papers in Linguistics. http://hdl.handle.net/1773/48570.

Perceived ethnicity, gender, and attitudes towards second language Mandarin speakers

Research has shown that when a listener believes that a second language (L2) speaker is a member of a different ethnic group from the listener, sometimes the listener will perceive that speaker to be more accented or of inferior character than a L2 speaker of the listener’s ethnic group. Most existing research demonstrating this has been done on majority languages of North American and European countries, raising questions about the generalizability of the research. In a pair of studies, ethnically Chinese first-language Mandarin listeners rated the personalities and language abilities of highly proficient L2 Mandarin speakers of Chinese and non-Chinese ethnicity. Results show an effect of ethnic ingroup favoritism on several personality traits but no difference in perceived language proficiency. Results also indicate that this effect is more pronounced for men listening to female voices. A follow up study showed that endorsement of hostile sexist viewpoints is associated with greater ethnic ingroup favoritism.

Reference: Squizzero, R. (2020). Attitudes toward L2 Mandarin Speakers of Chinese and non-Chinese Ethnicity. Proceedings of the 32nd meeting of the North American Conference on Chinese Linguistics. (link)

Vowels and consonants in second language Mandarin Chinese

For first-language (L1) speakers of non-tone languages, learning the four tones that distinguish words in Mandarin Chinese is often the hardest part of its pronunciation, if not the hardest part of learning the language itself. This may be the reason why the majority of existing research on acquisition of second language (L2) Mandarin pronunciation has focused on tone. Pronunciation of vowels and consonants, however, is highly important for communication, even in tonal languages. This study found statistically significant differences in the production and perception of Mandarin vowels and consonants between L1 Mandarin speakers and L2 Mandarin speakers with English as their L1. This paper was be presented at ISAPH 2022 in Lund, Sweden. Slides.

CLOx transcription

CLOx is a web-based service that uses Microsoft’s Speech Service API Client Libraries, previously known as Project Oxford, to orthographically transcribe sociolinguistic or other audio-recorded interviews in a format amenable to linguistic analysis. CLOx outputs transcriptions in a .csv format with timestamps indicating the start and end time of each turn of speech contained in an audiofile. It is estimated that CLOx enables accurate transcription of a sociolinguistic interview to be completed in one-fifth or less of the time it would take to produce a manual transcription. CLOx currently supports over 100 languages and language varieties. CLOx is also available as a desktop application for macOS. (with Alicia Beckford Wassink, Campion Fellin, and David Nichols)

Reference:

Wassink, A.B., Squizzero, R., Fellin, C. & Nichols, D. (2018, October). CLOx: Time-saving automated orthographic transcription tool for sociolinguistic interviews. Workshop given at New Ways of Analyzing Variation 47, New York, NY.

Improving the quality of dynamic acoustic measurements

Phoneticians measure formants, or acoustic resonances of the vocal tract, as one way to quantify the characteristics of vowels. These measurements rely on an imperfect signal processing procedure called linear predictive coding. A common approach used to identify measurements that are likely to be erroneous is to look for measurements that fall outside of two standard deviations from the mean for a given speaker and vowel. Preliminary indicate that robust Mahalanobis distance, a multivariate method, is more accurate at identifying erroneous measurements than using either a standard deviation or median absolute deviation approach. (with Alicia Beckford Wassink)

Ethnic minorities’ participation in a Pacific Northwest English sound change: An articulatory and acoustic investigation

This is a sociolinguistically-informed phonetic study of words with certain vowels before /l/ in the varieties of English spoken in Washington State. Some Washington English speakers do not differentiate certain words that speakers differentiate in other parts of the United States. Ultrasound imaging is used to investigate the timing and magnitude of tongue gestures and their effects on the acoustics of speech. The study samples from African American, Mexican American, and Yakama populations, ethnic groups that are minorities in Washington State and traditionally underrepresented in linguistic studies.