Wewalaarachchi Thilanga D, Singh Leher
Department of Psychology, National University of Singapore, Singapore Singapore.
Front Psychol. 2016 May 3;7:627. doi: 10.3389/fpsyg.2016.00627. eCollection 2016.
Early language acquisition is potentially complicated by the presence of many sources of variability in the speech signal. A frequent example of variability is phonological alternations, which can lead to context-driven changes in the realization of a word. The aim of the current study was to investigate effects of a highly frequent yet scarcely researched type of suprasegmental phonological alternation - tone Sandhi - on early spoken word recognition. The tone Sandhi rule investigated herein involves a tone change of the first syllable in a disyllabic unit. In accordance with third tone Sandhi, when two dipping tone syllables are juxtaposed in connected speech, the first syllable is dissimilated to a high rising tone. For example, 'flour mill' (unaltered pre-Sandhi form [(214) (214)]) undergoes tonal alternation resulting in the altered post-Sandhi form [(35) (214)]. In the current study, preschoolers' sensitivity to the effects of tone Sandhi when processing familiar words was investigated via a preferential looking paradigm. Words varied in their phonological form: one set of words was labeled with a phonological alternation due to Sandhi (Post Sandhi), one set of words was labeled with an unaltered Sandhi form (Pre Sandhi), one set consisted of non Sandhi words (Correct Pronunciation, and one set were labeled with a tonal alternation not associated with Sandhi rules (Mispronunciation). Post-Sandhi forms and correct pronunciations were associated with visual referents with comparable strength, with only a subtle processing cost observed for post-Sandhi forms in the time course of lexical selection. Likewise, pre-Sandhi forms and true mispronunciations were rejected as labels for visual references with comparable strength, with only subtle differences observed in the time course of lexical selection. Findings are discussed in terms of their impact on prevailing theories of lexical representation.
早期语言习得可能会因语音信号中存在多种变异性来源而变得复杂。变异性的一个常见例子是音系交替，它可能导致单词实现过程中由语境驱动的变化。本研究的目的是调查一种高度频繁但研究较少的超音段音系交替——连读变调——对早期口语单词识别的影响。本文研究的连读变调规则涉及双音节单元中第一个音节的声调变化。根据第三声连读变调规则，当两个上声音节在连贯语流中相邻时，第一个音节会异化变为阳平。例如，“面粉厂”（连读变调前未改变的形式[(214) (214)]）会经历声调交替，产生连读变调后的形式[(35) (214)]。在本研究中，通过优先注视范式研究了学龄前儿童在处理熟悉单词时对连读变调影响的敏感度。单词在语音形式上有所不同：一组单词因连读变调而带有音系交替（连读变调后），一组单词带有未改变的连读变调形式（连读变调前），一组由非连读变调单词组成（正确发音），还有一组带有与连读变调规则无关的声调交替（错误发音）。连读变调后的形式和正确发音与视觉参照的关联强度相当，在词汇选择的时间进程中，仅观察到连读变调后的形式有细微的加工成本。同样，连读变调前的形式和真正的错误发音作为视觉参照的标签被以相当的强度拒绝，在词汇选择的时间进程中仅观察到细微差异。研究结果将根据其对现行词汇表征理论的影响进行讨论。
