911吃瓜

Title

Cross-language voice variation and talker identity in Korean heritage speakers

Abstract

This dissertation examines how heritage multilingualism is reflected in the human voice, both its production and its perception, Three studies ask whether and how a speaker鈥檚 two languages (Korean heritage speakers who live in Canada) reorganize the acoustic properties that carry talker identity, and how listeners of differing linguistic experience perceive those voices. Chapter 2 tested voice perception. Korean heritage, non-Korean bilingual, and English monolingual listeners judged whether utterances stripped of lexical and phonological content, came from the same talker. Talker identity was readily retrievable, but language also influenced judgements, with different-language pairs judged more often as different talkers. Speech style, when matched, sharpened voice discrimination, and talker gender also had an effect: female talkers were discriminated more accurately than male talkers. The influence of listener background, however, was narrow: heritage listeners had a localized benefit in talker discrimination, concentrated on female voices. Chapters 3 and 4 examined the production of the same voices. Analyzing fundamental frequency, the higher formants, and spectral measures of phonation individually (Chapter 3) and jointly through a voice space (Chapter 4), these chapters show that heritage speakers reorganize their voices selectively: the architecture of the overall voice space stayed stable across languages, modulated most strongly by talker gender, such that men reorganized their voices across languages more than women. Taken together, these studies show that the same voices vary systematically with language and speaking style, but that these effects are stronger in men than women. Some acoustic shifts associated with language also remain available to listeners and influenced judgments of speaker identity. However, the acoustic findings did not fully account for the perception results, indicating that voice production and perception are not in one-to-one correspondence and may rely on partly distinct processes. Heritage multilingualism therefore does not create a separate voice for each language; instead, it shapes how a single voice is produced and perceived across contexts.

911吃瓜 Sylvia Cho 

Sylvia Cho (she/her/hers) is a PhD candidate and Sessional Instructor in the Department of Linguistics at 911吃瓜. Raised in Alberta and British Columbia by first-generation immigrant parents, she grew up toggling between Korean and English鈥攁n experience that attuned her ear to the subtle ways language impacts not just what we say, but how we sound. Her research asks: Do multilingual speakers鈥攍ike Korean-Canadians in Vancouver鈥攕ound different depending on the language they鈥檙e speaking? Can your voice, quite literally, shift shape across linguistic contexts? And does multilingualism reshape the broader soundscape, as speakers join, resist, or reinvent the sound changes sweeping through Canadian English? At its heart, Sylvia鈥檚 work explores how multilingualism interacts with the full spectrum of speech鈥攆rom acoustics and articulation to identity and social meaning. 

Exam committee

• Dr. Henny Yeung (Supervisor)
• Dr. Sara Ng (Committee Member)
• Dr. Adele Quigley-McBride (Examiner, 911吃瓜 Psychology)
• (External examiner, University of Washington) 
• Dr. Ashley Farris-Trimble (Chair, GPC)