Speech Perception and Memory in an Episodic Lexicon

描述（由申请人提供）：我们的研究研究了长期记忆中的词汇表征，以及这些表征在语音感知、产生和记忆中的编码和检索过程。大多数理论认为，语音波形被转换成音素串，然后与心理词汇进行比较。这种观点强调单词的抽象属性：包含特殊信息（例如说话人的声音）的语音信号被归一化为规范形式，然后与抽象的词汇条目相匹配。然而，情景记忆理论认为，感知经验的细节不会被遗忘。相反，它们储存起来，支持以后的感知。拟议的研究包括五个具体目标。首先，我们将准备一系列计算模型，从强情景模型到混合抽象-情景模型。其余四个目标对应于四个实证项目。这些实验使用言语感知、生产和记忆测量来测试情景观点。在项目1中，知觉和记忆实验将测试影响情景内容的因素，考察选择性注意和工作记忆的作用。实验将评估记忆痕迹形成中感知-概念权衡的性质和程度。在项目2中，多维尺度实验与记忆测试相结合，旨在发现情景词启动的相似-距离函数。项目3提供了情景视图的关键测试，使用眼动追踪程序来评估知觉中情景效应的时间进程。项目四将使用完整的句子作为刺激，检查更大的情景单元的记忆。除了标准测试，句子记忆将使用一种新的语音生成方法来评估，检查口语反应的声学-语音内容。根据我们之前的研究，我们期望人们自发地模仿刺激句的声音模式。情景理论预测，模仿程度将受到句子的“抽象”特征的影响，例如语义复杂性，这表明语音声学反映了语言刺激及其情景表征的复杂相互作用。

DESCRIPTION (provided by applicant): Our research investigates lexical representations in long-term memory and the processes that both encode and retrieve those representations in speech perception, production and memory. Most theories assume that speech waveforms are converted to phoneme strings, which are then compared to the mental lexicon. This view emphasizes the abstract properties of words: Speech signals containing idiosyncratic information (e.g., speaker's voice) are normalized into canonical forms, then matched to abstract lexical entries. However, episodic memory theories posit that details of perceptual experiences are not forgotten. Instead, they stored, and they support later perception. The proposed research consists of five specific aims. First, we will prepare a collection of computational models, ranging from strongly episodic to hybrid abstract-episodic models. The remaining four aims correspond to four empirical projects. These experiments test the episodic view using speech perception, production, and memory measures. In Project 1, perception and memory experiments will test factors that affect episodic content, examining roles of selective attention and working memory. The experiments will assess the nature and extent of perceptual-conceptual tradeoffs in memory trace formation. In Project 2, multidimensional scaling experiments are coupled with memory tests, aiming to discover the similarity-distance function in episodic word priming. Project 3 provides critical tests of the episodic view, using eye-tracking procedures to assess the timecourse of episodic effects in perception. Project 4 will examine memory for larger episodic units, using full sentences as stimuli. In addition to standard tests, sentence memory will be assessed using a novel speech-production method, examining the acoustic-phonetic content of spoken responses. Following our previous research, we expect people to spontaneously imitate acoustic patterns of the stimulus sentences. The episodic theory predicts that degrees of imitation will be affected by "abstract" characteristics of the sentences, such as semantic complexity, suggesting that speech acoustics reflect a complex interplay of linguistic stimuli and their episodic representations.