Research fellow Elisa Felsche (U-M Psychology) will give a talk on "The origins of inference: A hierarchical Bayesian modelling approach to imitation and abstraction in children and primates."

NOTE: If you would like to attend this meeting, please send an email to cogsi-seminar@umich.edu to receive the passcode.

ABSTRACT

Humans have an immense behavioural and cognitive repertoire that has been shaped by cumulative cultural evolution. In my PhD project I investigated two cognitive abilities that crucially enlarge the efficiency of skill and knowledge acquisition: 1) the capability for abstraction that enables powerful generalization of information to make wide ranging predictions in new situations and 2) the ability to imitate others which allows the quick and low-risk adoption of new behavioural strategies. Despite decades of accumulating data in both domains, it is still debated to what extent other species share these abilities and how they develop in humans. Solving these persisting disagreements requires an alteration of how data are generated and analysed.
In my dissertation project, I introduced the approach of hierarchical Bayesian modelling to the field of comparative psychology to investigate abstract rule formation and action copying in capuchin monkeys, chimpanzees (only abstract rule formation) and children. In the first two studies participants had to use sampled evidence to infer abstract rules about the item distributions in containers and efficiently guide behaviour in novel test situations. In a third study, we investigated children's and capuchin monkeys' ability to integrate causal and social information when copying a goal-directed behaviour. Whereas children’s performance was mostly in line with the predictions of the computational models, showing that they are capable of abstraction and consider causal information when imitating, capuchin monkeys performed in all experiments at chance and chimpanzees showed some understanding of abstract rules.