Are tones in the expressive lexicon iconic? Evidence from three Chinese languages

Recent advances in the literature have focused on sketching phonosemantic mappings of imitative or iconic utterances by relying on vowels and consonants, leaving the suprasegmental information unexplored. To begin bridging this gap, this study looks at the interaction of lexical tone and iconicity by comparing sound symbolic (i.e., mimetic, expressive, ideophonic) strata and general (i.e., arbitrary, prosaic, non-iconic) strata from three Chinese languages (Mandarin, Taiwanese Southern Min, Hong Kong Cantonese) using corpus-based means. For all three languages the distribution of tones in the sound symbolic strata are skewed so that the majority of syllables are largely confined to two tonal categories per language, one of which is high level, while the general strata exhibit no such tonal bias. These results indicate that phonological systematicity at the prosodic level might play an important role in demarcating an iconic class of words. This cross-linguistic tendency towards high tone mappings may be derived from phonotactic strategies to facilitate prosodic foregrounding of iconic utterances as well as an embodiment of expressive voice and marked pitch use like that of Infant Directed Speech.

现有研究的最新进展多聚焦于依托元音与辅音，梳理模仿性或象似性话语的语音语义映射（phonosemantic mappings）关系，却尚未对超音段信息（suprasegmental information）展开探索。为填补这一研究空白，本研究依托基于语料库的研究方法（corpus-based means），对比三种汉语（普通话、台湾闽南语、香港粤语）的语音象征层级（sound symbolic strata，即模仿性、表情性、象声性层级）与普通层级（general strata，即任意性、平实性、非象似性层级），以此探讨词汇调（lexical tone）与象似性（iconicity）之间的交互关系。三种汉语的语音象征层级中调类（tonal categories）分布均呈现偏斜特征：多数音节大多被限定在每种语言的两个调类范围内，其中一个为高平调（high level）；而普通层级则未表现出此类调类偏向（tonal bias）。上述结果表明，韵律层面（prosodic level）的语音系统性（phonological systematicity）或许在划定象似性词汇类别中发挥着重要作用。这种偏好高音映射的跨语言倾向（cross-linguistic tendency），或许源于两类动因：一是依托音位策略（phonotactic strategies）以实现象似性话语的韵律凸显（prosodic foregrounding）；二是呼应了婴儿导向言语（Infant Directed Speech）中富有表现力的发声与标记性语调使用的特征。