Martin Ip
Linguistic Data Consortium
University of Pennsylvania

Searching for the “Edges” of Speech: How Prosody and Speaker Identity Support Language Processing

Human speech contains a continual cascade of information, from the physical properties of the speech sounds to the sequencing of words and the wider discourse context. To anticipate the likely continuations of what we hear, we must constantly build up knowledge about the incoming signal and integrate information from all parts of the language structure. A key challenge in the psychology of spoken language is to reveal how the mind can turn such a multifaceted stream of sounds into meaningful words, sentences, and dialogues. Of course, success in vocal communication and listening in everyday conversations is not just about attending to the speech sounds (i.e., the vowels and consonants), and it is also not just about decoding the literal meanings of words. In this talk, I will discuss two types of cues in the speech stream that had once been considered to be the “edges” of speech and language, and I will present experimental evidence showing how they may serve an important role in language processing and learning. First, I will discuss cross-linguistic experiments demonstrating how attention to prosody, the suprasegmental structure of speech, may provide a likely universal strategy for listeners to encode information structure in real time and to learn the meanings of words. I will also discuss how prosody can potentially be used in clinical settings to capture language differences between neurotypical and neuroatypical individuals (e.g., autistic children), and how social contexts may interact with prosody to affect these speech differences. In the second part of the talk, I will present evidence showing how listeners form social judgments from the speech input by integrating a richer network of speaker abilities and preferences that forms part of the speaker’s perceived social identity.