Speakers, listeners, languages: Patterns of variability and contrast in spoken language dynamics

我们提出的项目的目标是揭示语言产生和感知的基本动态是如何被语言特定的语言结构所塑造的，同时，在同一语言社区的个体之间以系统的方式变化。人类口语的一个决定性特征是个体有意义的语音单位（音素）动态地、相互交织地产生，它们共同构成单词和句子。虽然我们可以将音素正字法表示为字母序列，但语音生产中音素的实现并不像串在弦上的珠子那样随时间展开。相反，我们发出这些单独的语音的方式受到它们之前和之后的声音的影响，并依赖于它们，这种现象被称为协同发音。协同发音确保了快速有效的沟通——它是积极计划的，是习得的、特定语言的说话者知识的一部分。同样在感知方面，听者在确定说话者说了哪些词时，会积极地在说话者的话语中寻找协同发音的动态线索。虽然一再有人认为语音对比——即构成语言清单一部分的声音——限制了协同发音的程度，但这一说法的经验证据仍然含糊不清，因为语言之间的差异并不一定像语音结构所预测的那样。与此同时，同一种语言的个体使用者之间的协同发音模式也有很大的差异。最近的研究发现，语音对比信号的个体差异可能不仅仅是“噪音”，而是一种高维的、特定于语言的生产感知动态的一部分，目前人们对这种动态知之甚少。拟议的项目将使我们能够全面了解个体语言使用者感知和产生的语音之间的关系，其规模将使我们能够评估语言内部和语言之间个体之间的结构变化水平。我们的工作将集中在三种语言（英语，德语，法语）的两种语音对比（鼻音和唇形），并且包括比大多数先前研究更多的参与者。因此，该项目前所未有的范围有可能为以下方面开辟新的视角：1)语言为何以及如何彼此不同；2)语言变化的哪些方面对所有人类语言使用者都是普遍的；3)声音系统如何进化；4)语音产生和感知之间的认知关系。该项目的另一个好处是向科学界提供一个可用于研究和教学目的的综合跨语言数据库。

The goal of our proposed project is to shed new light on how the basic dynamics of speech production and perception are shaped by language-specific linguistic structure and may, at the same time, vary between individuals of the same language community in a systematic fashion. A defining characteristic of human spoken language is the dynamic, interwoven production of individual meaningful speech units (phonemes) which together form words and sentences. Although we may represent phonemes orthographically as sequences of letters, the realization of phonemes in speech production does not unfold in time like beads on a string. Rather, the way that we produce these individual speech sounds is affected by and dependent upon the sounds that precede and follow them in time, a phenomenon known as coarticulation. Coarticulation ensures rapid and effective communication — it is actively planned and is part of learned, language-specific speaker knowledge. Also in perception, listeners actively seek out the dynamic cues of coarticulation in a speaker's production when determining which words were spoken. While it has repeatedly been argued that phonological contrast — i.e., which sounds form part of a language's inventory — constrains the degree of coarticulation, the empirical evidence for this claim has remained ambiguous, since variation between languages does not necessarily pattern as predicted by phonological structure. At the same time, there is substantial variation in coarticulatory patterns among individual speakers of the same language. Recent studies have uncovered that this individual variation in how phonological contrast is signaled may be not just ‘noise’ but instead be part of a high-dimensional, language-specific production-perception dynamic which is currently little understood. The proposed project will enable us obtain a comprehensive view on the relation between perceived and produced speech for individual language users, at a scale that will allow us to assess levels of structured variability among individuals both within and between languages. Our work will focus on two phonological contrasts (nasality and lip rounding), in three languages (English, German, French), and include a larger number of participants than most previous studies. The project's unprecedented scope has therefore the potential to open new perspectives on 1) why and how languages differ from each other, 2) what aspects of language variation are universal to all human language users, 3) how sound systems evolve, and 4) the cognitive relationship between speech production and perception. The further benefit of the project lies in making available to the scientific community a comprehensive cross-linguistic database that can be used for research and teaching purposes.