Modeling Polite Speech: Tradeoff between Communicative Goals
2018 Dissertation Fellowship
Erica's dissertation work focuses on people’s production and understanding of polite speech: Why and how do people speak politely, and how do they infer intentions behind polite speech? To explore these issues, Erica uses computational and developmental approaches. First, she and her collaborators built a utility-theoretic formal model of polite speech that explains polite speech emerges from speakers’ desires to be and to appear informative and kind. Using the model, they were able to successfully capture key patterns of simple white lie and indirect speech production in adults.
Second, Erica is looking at how people attribute intentions behind polite speech, and how that ability develops from childhood to adulthood. To do that, she first looks at when the ability to attribute intention to be polite emerges in very young children (2-4-year-olds): when do they know that it is more polite to use words like “please”? Then she looks at the development of understanding white lies, where there is a conflict between the want to convey the truth (honesty) versus the want to avoid hurting the listener’s feelings (kindness