An acoustic analysis of pause patterns in Chinese learners’ speech production

Pause has been defined operationally as the measurable silence between the words on either side of a syntactic boundary. While pausing is critical for intelligibility (Anderson-Hsieh, 1992), the literature contains few acoustic studies on the significance of pauses among ESL learners. The major goal of this study is to investigate the difficulties in English speech pause patterns encountered by Taiwanese learners through examining acoustic measures. An acoustic study designed to measure the three aspects of pausing was conducted. These patterns were: pause duration, pause frequency, and the distribution of pauses. They are defined and measured as follows (Riazantseva, 2001): (a) Pause frequency refers to the number of silent pauses that occur per sentence. (b) Pause distribution refers to the silent pauses that occur at- constituent boundaries (e.g. before or after NP constituent) and within-constituent boundaries (within NP constituent). (c) Pause duration refers to the length of silent pauses and was measured in milliseconds and reported in seconds. Four groups of subjects were recruited to produce a long English text, including 10 Americans from the US, 10 Chinese learners with low English proficiency (EFL-L) from Taiwan, 10 Chinese learners with high English proficiency (EFL-H) from Taiwan, and 10 ESL Chinese learners (ESL) from the US. The passage was recorded and analyzed with the Praat software. A new norm-referenced variability index (VI) was proposed and measured over the timing variables in this study. Major findings of this study are as follows. First, the three variables of the Taiwanese learners display specific patterns that deviate from those of native English speakers. They demonstrate more frequent and inappropriate pausing. Some word boundaries are maintained through the insertion of short pauses and glottal stops. The EFL-L, the EFL-H, and the ESL learners behave differently in acoustic timing patterns. The ESL learners do not significantly outperform the EFL-H learners on all the three variables. Language proficiency affects the way a person uses pauses when speaking an L2. In this sense, pausing could be a developmental phenomenon: pausing becomes more native-like as higher language proficiency is achieved. Also, the random pauses produced by EFL-L largely enhance the impression of syllable-timing patterns, whereas EFL-H and ESL were passing through the syllable-timing stage and moving to the other direction of stress-timing. Based on these findings, it can increase the understanding of interlanguage phonetics and phonology and second language acquisition.