New Perspectives on Mental State Attribution

Mental state attribution is one of the most interesting human capacities: we instinctively see other agents as driven by inner states such as beliefs, goals and desires. We are not always perfectly accurate in our assessments of what others are feeling and thinking, but our natural social intelligence is strong enough to enable us to compete and cooperate with others across many contexts. Researchers across disciplines have used a great variety of methods to analyze this natural social intelligence: in particular, mental state attribution has been the focus of intense and interestingly different forms of scrutiny within philosophy, linguistics, and four branches of psychology (developmental, comparative, social and cognitive).

The aim of this workshop is to bring together leading and emerging researchers from across these disciplines to share the most innovative new research on mental state attribution, with a special focus on gaining understanding of the philosophical significance of recent scientific findings. Presenters at this workshop are using a terrific variety of methods, from computer simulation of artificial agents to fieldwork with social animals, from research on the expression of mental states in aboriginal languages to abstract reflection on the nature of mental states themselves. Both established and emerging scholars will participate in the exchange of new ideas and new methods of inquiry into the core of human social intelligence.

We have a range of innovative presenters, starting with Canadian philosopher Kristin Andrews, of York University, who works on animal minds. Informed by her fieldwork with orangutans in the wild, the research she is presenting focuses on the extent to which moral dimensions of human social intelligence are shared with other animals. The other philosophy presenter at the workshop will be Johannes Roessler of the University of Warwick, UK, examining the relationship between the early development of self-awareness and the human capacity to attribute states of knowledge. From Google's DeepMind Project we have Neil Rabinowitz, lead author on a new paper explaining how mental state attribution can be executed in a Deep Learning paradigm: his research group has trained a simulated neural network to attribute mental states to simulated agents. The workshop features two developmental psychologists who take interestingly opposed positions on the early development of mental state attribution: Ted Ruffman of the University of Otago, New Zealand, and Hannes Rakoczy of Göttingen University, Germany. From MIT in the United States, we have social cognitive neuroscientist Rebecca Saxe, who was one of the first researchers to localize mental state attribution functions in the brain. Her latest work has focused on the question of how functional specialization develops in the brain to support the variety of tasks needed for mature human social intelligence. From linguistics, we have Elizabeth Bogal-Allbritten of the University of Gothenburg, Sweden, whose work examines the linguistic expression of mental state attributions, in particular attributions of uncertainty, belief and desire. Her work on Navajo, in particular, raises interesting questions about the diversity of linguistic structures that can be used in talking about the mental states of others. We will also have Rachel Dudley from the Ecole Normale Supérieure, France, whose research examines how children acquire semantic and pragmatic competence, in particular, she has studied how children discover that certain linguistic expressions such as "know" trigger presuppositions.