Search Abstracts | Symposia | Slide Sessions | Poster Sessions
“Cohort” and “Rhyme” in Lexical Tones in Spoken Word Recognition
Poster Session C, Friday, October 25, 4:30 - 6:00 pm, Great Hall 3 and 4
Xin Wang1, Bob McMurray2; 1Macquarie University, 2University of Iowa
One of the classic debates in Spoken Word Recognition (SWR) is whether lexical tones carry similar cue weighting as in segments in recognizing a spoken word (Malins & Joanisse, 2010). Tones and segments belong to different acoustic dimensions, and they are intertwined given a syllable. However, it is not clear how different tones, as critical phonetic cues in a tonal language, are integrated during spoken word recognition. Theories of speech perception and spoken word recognition must take into account the timing of tone integration (McMurray et al., 2008). Here, we ask a different question. Instead of comparing the relative cue weighting of tones and segments, we first establish the time course of tone integration during SWR. That is, do listeners wait till a unique point of pitch contours to kick off lexical competition or integrate tonal cues incrementally? We used the Visual World Paradigm to investigate this question (Tanenhaus, et al., 1995), specifically, for Tone 1, Tone 2 and Tone 4 in Mandarin Chinese. Mandarin Chinese has four distinct tonal categories (Wang, 2021). We showed tone timing of three tone contrasts (T1 vs. T2, T1 vs. T4, T2 vs. T4) as the Target and Competitor in VWP during word recognition (e.g., ‘/ba1/’ competed by ‘/ba2/’), given the same syllable in a particular trial. Eye-movement data analysis was based on non-linear curve fitting to show the time course of eye-movements on different objects (i.e., targets, competitors and distractors), and the competition effect was measured as the difference between eye-fixations on competitors and distractors (McMurray, 2023). In a within-participant and within-item design, we observed robust competition effects across all tone contrasts (all p <.05). These results indicate that lexical tones serve as independent cues in SWR. Importantly, these competition effects arising from different tone contrasts differ from each other, measured by curve parameters. That is, different tone contrasts generated different degrees of competition effects in terms of when the curves reached the peaks and their magnitudes. Our results showed earlier and stronger competition effects of Tone 1/4 contrast relative to Tone 1/2 contrast, indicating listeners were sensitive to F0 height when integrating tonal cues. In Mandarin, Tone 1 and Tone 4 start from similar F0 height but diverge from each other given time, and the competition effect for this tone contrast resembled the Cohort effect, where acoustic cues activate multiple lexical access (Marslen-Wilson, 1987). Tone 1 and Tone 2 start from different F0 height but gradually converge given time, and the competition effect of this contrast resembled the Rhyme effect (Luce & Pisoni, 1998). Finally, Tone 2 and Tone 4 contrast shares the least similarity in F0, and thus showed the least competition effect in their curves compared to the other two contrasts. To summarize, our results demonstrate incrementality in lexical tone processing and support word recognition theories of cue integration in which tones are used for lexical access as soon as they are available.
Topic Areas: Speech Perception, Phonology