Music and speech are distinct in lexical tone normalization processing

Music and speech are distinct in lexical tone normalization processing Ran Tao Department of Chinese and Bilingual Studies, The Hong Kong Polytechnic University rantao@polyu.edu.hk Gang Peng Department of Chinese and Bilingual Studies, The Hong Kong Polytechnic University gpeng@polyu.edu.hk Abstract This paper presents the results of a Cantonese lexical tone normalization experiment. Two non-linguistic (i.e., music and nonspeech) context conditions were used in addition to a speech context condition to elicit listen- ers’ normalization of level tones following contexts with fundamental frequency alter- ations. Participants showed clear tone normal- ization in the speech context condition. Par- ticipants’ normalization performances in the non-linguistic conditions were not significant but comparable to each other. The findings in- dicated that the non-linguistic music context is not sufficient to support lexical tone normal- ization. 1 Introduction In daily life, people speak with huge inter- and intra- talker variance (Peng et al., 2012). The task of normalizing variable vocal streams to identical lin- guistic codes seems challenging. When it comes to the tone languages such as Cantonese, the situ- ation is even more complicated (Wong and Diehl, 2003). Cantonese has three level tones convey- ing distinct meanings when superimposing to iden- tical syllables, leaving the pitch height the only clue for listeners’ successful judgments. The level tones’ pitch heights frequently overlap with each other among speakers and within a speaker, such as in different emotional states. However, the task of identifying words in a vocal stream, i.e., lexical tone normalization, can be accomplished by typical Cantonese without conscious awareness (Wong and Diehl, 2003; Francis et al., 2006; Zhang et al., 2016; Zhang et al., 2017). Research on lexical tone normalization has re- vealed the significant role of immediate context which provides a reference to normalize the tar- get pitch height. Researchers debated on whether such processes depend on a general-purpose audi- tory mechanism or a speech-specific mechanism. There are contradictory findings in the literature, providing evidence to support both sides. Huang and Holt (2009; 2011) found both nonspeech and speech context is sufficient to facilitate the listener’s tone normalization, with a restricted effect in the non- speech context. However, other researchers failed to replicate the facilitation effect of nonspeech con- text, and mainly support the speech specific mecha- nism (Zhang et al., 2012; Zhang et al., 2013; Zhang et al., 2016; Zhang et al., 2017). While the nonspeech context can replicate the pitch pattern in speech context, they differ in as- pects other than whether or not containing a linguis- tic meaning. The nonspeech context is novel to lis- teners because such stimuli are rare in daily lives. It is plausible that listeners cannot take advantage of the immediate context of nonspeech as they are not familiar with such auditory patterns. Music, as non-linguistic stimuli, frequently ap- pears in daily lives. More interestingly, previous re- search suggested that music could benefit linguistic abilities. Nan et al., (2018) found that piano training can enhance the neural processing of pitch and thus improves speech perception in Mandarin-speaking children. Other …

Content truncated to 500 words