Sound Patterns and Linguistic Structures at the Transition Space in Conversation

互动是社会行为的基础，话轮转换是互动的重要组成部分。人类的互动是独特的，因为它使用语言作为媒介，特别是在日常对话中。会话中的话轮转换可能很容易被认为是理所当然的，但它有一个复杂的系统性组织，参与会话的人能够在其他说话者结束时迅速开始说话，如此之快以至于人类的认知能力无法达到这样的速度，除非人们能够预测其他人的会话回合何时以及如何结束。拟议中的研究调查在什么时候在会话中这些预测变得相关，也就是说，当一个可能的“话轮转换空间”出现在一个给定的扬声器的回合，以及什么样的语言信息的听众使用，以便作出这样的预测。特别是，我们专注于韵律变化：声音特征，如如何响亮或快速，或与什么样的旋律模式，单词或句子被说出。该项目的第一部分调查的韵律变化的音频记录的对话，调查不同的位置，过渡空间预测可能开始，并在德国和瑞典的比较结果。第二部分在感知实验中测试这些结果，研究人们在听对话时如何对不同类型的声音模式做出反应，特别是来自不同语言的听众如何解释相似的声音模式。该项目的结果将突出不同语言实现类似交际目标的方式，并将对人类认知过程的交流和互动提供有价值的见解。

Interaction is fundamental to social behavior, and turn-taking is an essential component of interaction. Human interaction is unique in that it makes use of language as a medium, most notably in everyday conversation. Turn-taking in conversation may be easily taken for granted, but it has a complex systematic organization, and people who are participating in conversations are able to start speaking quickly when other speakers finish—so quickly that human cognitive abilities could not achieve such speed unless people are able make predictions about when and how other people’s conversational turns will end. The proposed research investigates at what times during conversation these predictions become relevant—that is, when a possible “turn-transition space” arises during a given speaker’s turn—and what kinds of linguistic information listeners use in order to make such predictions. In particular, we focus on prosodic variation: sound characteristics such as how loud or fast, or with what kinds of melodic patterns, words or sentences are spoken. The first portion of the project investigates prosodic variation in audio recordings of conversations, investigating different locations where transition space prediction could begin, and comparing findings in German and Swedish. The second portion tests these results in perception experiments, studying how people listening to conversations respond to different types of sound patterns, and in particular, how listeners from different languages interpret similar sound patterns. The results of the project will highlight ways in which different languages achieve similar communicative goals, and will give valuable insight into processes of human cognition for communication and interaction.