Past Event

Yoichi Mukai
(University of Alberta)

Pupillometry as a processing measurement of morphologically complex words:

Using eye-tracking techniques a number of studies have measured gaze locations and durations to investigate how we comprehend language. Although pupil size data is acquired together with eye movement information, using them as a dependent measure is not widespread (Laeng et al., 2012). Pupillometry is beneficial because it has been shown to measure the degree of cognitive load (Beatty, 1982; Kahneman & Beatty, 1966). In the present study, I investigated what pupillometry, measured during a lexical decision task, can contribute to the advancement of lexical processing research. Employing both visual and auditory modalities, I specifically examined the time-course of the processing of pseudo complex words (e.g., corner), as compared to transparent complex words (e.g., cleaner) and monomorphemic words with an embedded stem (e.g., turnip). The results revealed interactions between the auditory and visual modality, as well as between the effect of stem frequency and word frequency. Our findings also provided supporting evidence for Rastle & Davis’s (2008) morphological decomposition model for visual word recognition and for the Shortlist model (Norris, 1994) for spoken word recognition. In short, this study showed the promise of pupillometry in both visual and spoken word recognition research, especially with the appropriate statistical analysis method.

How much do durational variations matter? Modeling the perception of fluency and accentedness in L2 speech using rhythm metrics:

Research suggests that contrastive rhythmic patterns between L1 and L2 affect the intelligibility of non-native speech (e.g., Adams, 1979). Dilley et al., (2012) found the positive correlation between judgments of rhythmicity and fluency. White & Mattys (2007) also revealed the relationship between rhythm metrics and the perception of L2 accent. Rhythm metrics measure the overall proportion and durational variability of vocalic and consonantal intervals (e.g., Ramus et al., 1999), as well as durational differences between consecutive vocalic and consonantal intervals (e.g., Grabe & Low, 2002). In this study, I examined rhythm metrics as perceptual correlates of fluency and accentedness in L2 speech using conversational speech of Japanese ESL speakers. The results revealed that rhythm metrics and speech rate can predict the native and non-native speech difference very well (0% error rate), as well as the ratings (i.e., High, Mid, Low) of fluency and accentedness relatively well (approximately 25% error rate). The ranking of variable importance also indicated that speech rate is the most important predictor for the perception of both fluency and accentedness. In short, rhythm metrics and speech rate may be a promising tool for the implementation of a computer-mediated fluency and accent rating.

About the Presenter

Yoichi Mukai is a PhD student in the Department of Linguistics at the University of Alberta. He received an MA in Applied Linguistics and Discourse Studies from Carleton University, and a post-graduate certificate in TESL from Algonquin College. He has taught Japanese at institutions including Carleton University, and Université du Québec en Outaouais

This event is sponsored by the School of Linguistics and Language Studies