The learning and use of suprasegmental information in second-language speech perception and spoken word recognition: Where are we now?
The Learning and Use of Prosodic Information in
Second-Language Speech Perception and Spoken Word Recognition:
Where Are We Now?
Spoken word recognition in the native language (L1) is rapid and efficient. One factor that contributes to the efficiency of this process is native listeners’ use of word-level prosodic information. For example, in languages with lexical stress contrasts, native listeners more strongly activate word candidates if their stress pattern matches the acoustic input than if it does not (e.g., Cooper, Cutler, & Wales, 2002; Donselaar, Koster, & Cutler, 2005; Reinisch, Jesse, & McQueen, 2010; Soto-Faraco, Sebastián-Gallés, & Cutler, 2001). To illustrate, upon hearing the stressed syllable CAR-, English listeners more strongly activate words such as CARcass, CARpet, and CARtridge than words such as carTEL and carTOON (e.g., Cooper et al., 2002). This reduces the number of word candidates that strongly compete for activation and speeds up word recognition, making it a far more efficient process than if lexical stress were ignored.
It is well established that spoken word recognition in a second language (L2) is slower and less efficient. Previous research has effectively demonstrated that this lack of efficiency can stem in part from L2 learners’ non-native-like use of word-level prosodic information (e.g., Connell et al., 2018; Dupoux, Sebastián-Gallés, Navarrete, & Peperkamp, 2008; Lin, Wang, Idsardi, & Xu, 2014; Qin, Chien, & Tremblay, 2017; Tremblay, 2008). While researchers generally agree that L2 learners’ difficulty in the perception and processing of lexical stress can be traced back to the L1, they disagree on the nature of L1 effects. Phonological approaches stipulate that L2 learners’ ability to learn lexical stress and use it in spoken word recognition is determined by whether or not the L1 has lexical stress contrasts (e.g., Dupoux et al., 2008; Lin et al., 2014; Peperkamp & Dupoux, 2002; Tremblay, 2008). Phonetic (cue-weighting) approaches instead propose that L2 learners’ ability to learn lexical stress and use it in spoken word recognition is determined by the degree to which the acoustic cues to stress in the L2 signal lexical contrasts in the L1 (e.g., Chrabaszcz, Winn, Lin, & Idsardi, 2014; Connell et al., 2018; Qin et al., 2017; Zhang & Francis, 2010).
In this presentation, I will provide evidence that a phonetic (cue-weighting) approach has more explanatory power than a phonological approach for understanding L2 learners’ use of lexical stress in spoken word recognition. I will do so by reporting the results of cue-weighting, sequence-recall, and eye-tracking experiments testing the perception and/or processing of lexical stress by Dutch and Mandarin L2 learners of English (and native English listeners). The presentation will conclude with a discussion of the theoretical and pedagogical implications of this research.