In his paper on the “human interaction engine”, Levinson famously asserted that, in social interaction, people’s responses “are to actions and intentions, not to behaviors” (2006: 45). Indeed human beings attribute intentions/goals to the production of signals and parsing other’s signals means simulating others’ mental worlds, at least to some degree. But how do speakers calibrate their interactional moves in first position so that they are more likely to elicit their preferred response? Which variables do they take into account?
In this lecture I present observational and experimental data on how human and non-human animals (chimpanzees, bonobos, orangutans, dogs) calibrate requests for actions and for objects. I will discuss the role of prospection, entitlement, epistemic asymmetries and accountability in the calibration of adult requests. I will show how children’s request format changes over time, and finally when and how we pursue responses when they are missing. In doing so, I will provide an overview of factors to consider in the design of agents/machines that are supposed to interact with humans.
Federico Rossano is an associate professor in the Cognitive Science department at UC San Diego and the director of the Comparative Cognition Laboratory. His work is highly interdisciplinary. He has conducted studies on cooperation and communication in all great apes, baboons and macaques, dogs, cats, goats horses and rats. He has studied the development of joint attention and social norms (in particular property/ownership concerns) in young children from all continents and has published several papers on multimodal communication in adult humans (gaze, gestures, posture) and on how knowledge is negotiated in conversation (epistemics). He has collected the largest video longitudinal dataset of baby apes (4000 hours of videos from 31 animals ) and hundreds of hours of video footage of preschoolers interacting with peers. He conducts both observational and experimental work in several countries. He is the PI of How.TheyCanTalk, the largest citizen science study ever attempted on Animal Communication, collecting longitudinal data on 5000 dogs and 1000 cats (from 47 countries) learning to communicate with humans through a soundboard. He has published in Science Advances, Science Robotics, PNAS, Psychological Science, Cognition, Proceedings of the Royal Society B, Animal Behavior, Animal Cognition, Child Development, Developmental Psychology, Journal of Linguistics, etc.