A critical (re-)assessment of the effect of speaker ethnicity on speech processing and evaluation by Noortje de Weers M.A., Leiden University, 2013 M.A., Leiden University, 2012 B.A., (Hons.), University College Roosevelt, 2011 Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Department of Linguistics Faculty of Arts and Social Sciences © Noortje de Weers 2020 SIMON FRASER UNIVERSITY Spring 2020 Copyright in this work rests with the author. Please ensure that any reproduction or re-use is done in accordance with the relevant national copyright legislation. Approval Name: Noortje de Weers Degree: Doctor of Philosophy (Linguistics) Title: A critical (re-)assessment of the effect of speaker ethnicity on speech processing and evaluation Examining Committee: Chair: Nancy Hedberg Professor Murray Munro Senior Supervisor Professor Tracey Derwing Supervisor Adjunct Professor Yue Wang Supervisor Professor Elina Birmingham Internal Examiner Associate Professor Department of Education Molly Babel External Examiner Associate Professor Department of Linguistics University of British Columbia Date Defended/Approved: February 7, 2020 ii Ethics Statement iii Abstract In recent years, there has been a growing interest in the bidirectional relationship between speech and social processes, as increased attention is given to how speakers’ physical appearance, in combination with their accent, can influence the perception of their spoken language. Two competing theoretical frameworks have been proposed to explain conflicting findings in the existing literature: supporters of the reverse linguistic stereotyping hypothesis argue that listeners’ inherent racial biases against certain groups and their speakers negatively influence their speech evaluations (e.g., Rubin, 1992; Yi, Phelps, Smiljanic, & Chandrasekaran, 2013), while proponents of exemplar-based models of perception maintain that such negative judgments reflect the cognitive consequences of incongruent face–accent pairings (e.g. Babel & Russell, 2015; McGowan, 2015). Using this debate as a point of departure, this cross-cultural and cross-linguistic investigation was designed to determine whether reported effects of speaker ethnicity also extend to online processing speeds. Two response time studies (one using photographs and one using dubbed videos of Asian and White speakers of English) were conducted in Canada, while a third study using dubbed videos of Moroccan and White speakers of Dutch was conducted in the Netherlands. Additional offline dependent measures included sentence verification scores, accentedness ratings, verbal repetition accuracy, and credibility scores. Results from the three experiments showed (1) a processing cost associated with foreign-accented and non-standard speech, but (2) no effect of ethnicity on processing speeds or on the other dependent measures. These outcomes do not support the predictions of either theoretical framework, given that both presuppose an effect of speaker ethnicity on speech evaluation. The fact that the observed null findings are consistent with some previous studies highlights the potential influence of methodological choices underlying the seemingly contradictory findings in the literature. In view of this possibility, the findings are discussed in relation to the distinction between perception and interpretation. Further research will be needed to determine the true nature and magnitude of the effect of visually based social information on speech processing and evaluation. Keywords: audio-visual speech processing; ethnicity; accents; speech evaluation; sociophonetics; racial bias iv Dedication For John and Milo. Home is wherever you are. v Acknowledgements I am fully aware that I would not have been able to succeed in completing this project without a battalion of people who steadfastly encouraged and supported me throughout the years. The list of people who helped me in some shape or form is very long, so I would like to use this section to specifically mention a few: my supervisor Murray Munro, who wrote me countless reference letters, gave me access to expensive equipment, provided feedback on my writing, and helped me complete this monstrosity of a thesis. Tracey Derwing, whose life advice and expert guidance helped shape this dissertation. John, the best partner I could’ve hoped for. You must’ve heard the rundown of my research so many times, you could probably have written the abstract for me! My older brother Koen, for getting a haircut he definitely did not need, just to get me a voice donor. My sister Sietske and her partner Erik, who singlehandedly recruited most of my Dutch participants in Amsterdam. My younger brother Thomas, for enthusiastically suggesting his partner as a voice donor for me (thanks J!). My parents Hettie and Paul, for teaching me to be adventurous, critical, confident, and principled. Cliff, for making me feel at home in the first months after moving to Canada, and for the many talks in his office. Molly Babel, for sharing her multi-talker-babble file with me. Stefan Grondelaers, for helping me with my data collection in the Netherlands. Saya Kawase, for volunteering to record some of her Japanese students for my first study. My ‘stats guys’ Neil Faught, Ian Bercovitz, and Benjamin Taft, without whom this sophisticated statistical analysis would not have been possible. Finally, I owe the success of this research to those who graciously lent me their voices and faces, and of course to those who volunteered their time to participate. After all, without them I would have had no data to analyze! This research was also financially supported by many organizations. The Association for Canada Studies in the Netherlands was the first to provide funding, followed by Dr. Hendrik Muller’s Vaderlandsch Vonds, and het Prins Bernhard Cultuurfonds. These three not-for-profit organizations financially supported my proposal before I had even begun my studies, which helped convince others that my research had merit. I am also very grateful to Simon Fraser University, the Province of British Columbia, the David See-Chai Lam Centre for International Communication, Mitacs, the Dr. Tai Whan Kim endowment, and the Social Sciences and Humanities Research Council of Canada for all awarding me scholarships to support my livelihood and this research. vi Table of Contents Approval .......................................................................................................................... ii Ethics Statement ............................................................................................................ iii Abstract .......................................................................................................................... iv Dedication ....................................................................................................................... v Acknowledgements ........................................................................................................ vi Table of Contents .......................................................................................................... vii List of Tables ................................................................................................................... x List of Figures................................................................................................................ xii List of Acronyms ............................................................................................................ xiii Inspirational Quote ....................................................................................................... xiv Chapter 1. Introduction 1 Chapter 2. Visual information and speech processing 4 2.1. Speaker information through labeling .................................................................... 5 2.2. Visual speaker information .................................................................................... 8 Reverse linguistic stereotyping 8 The exemplar theory 13 Bias and stereotypical expectations 17 2.3. Perception vs. interpretation ................................................................................ 27 2.4. Interim summary .................................................................................................. 32 2.5. Measuring comprehensibility through reaction times ........................................... 33 Chapter 3. Experiment 1 with photographs in Vancouver 35 3.1. Methods Experiment 1A ...................................................................................... 37 Stimulus sentences 37 Speakers 37 Task Design 38 Listeners 40 Procedure 41 3.2. Methods Experiment 1B ...................................................................................... 42 Stimuli 43 Listeners 43 Procedure 43 3.3. Predictions ........................................................................................................... 44 3.4. Results ................................................................................................................ 45 Analysis I: Response times 45 Analysis II: Probability of correct sentence identification 51 Analysis III: Predicted RTs based on comprehensibility and accentedness 54 Analysis IV: Individual voice differences 58 3.5. Discussion of Experiment 1 ................................................................................. 59 The effect of Voice 59 vii The effect of Face 60 Speed/accuracy trade-off 60 Limitations 60 Chapter 4. Experiment