What are they up to?
Differential contributions of sensory information and prior knowledge to understanding others' intentions
When observing the behaviour of our conspecifics, we are often able to retrieve the intentions that motivate it. How is that possible? The discovery, in the monkey and then in humans, of a system of cerebral regions endowed with 'mirror' properties was hailed as evidence that action observation can provide a direct access to the intentions of an agent. However, this 'mirror' account of action understanding implicitly assumes a one-to-one mapping between intention and action. Yet, as a rule, the relation between mental states and observable behaviour is more complex than a one-to-one correspondence. Besides, information conveyed by kinematic movement is often noisy, ambiguous, or incomplete. As a result, visual information generally under-constraints the space of candidate intentions that are logically consistent with what is observed.
One way to solve this problem is to assume that this space of possible intentions is further constrained by the observerís prior expectations. These expectations are derived from a prior knowledge that may originate from the past experience of the viewer, from her intuitive theories or reputational knowledge as well as from contextual information surrounding the action scene. By exerting top-down influences onto the observerís perceptual representations, this prior knowledge makes possible inductive inferences about the agentís intentions, even in case of noisy signals or incomplete data, thus largely accounting for the robustness of our everyday inferences.
Although most authors agree that prior knowledge and perceptual information both contribute to the process of intention reading, the precise contribution of each type of information as well the exact way they interact remain ill-understood. The problem appears to stem in part from the fact that the term "intention", as it used in the empirical and philosophical literature, is an umbrella term, that may embrace many sub-types of intentions with different properties. One important dimension of variation concerns the scope of the intention; i.e., the more or less complex nature of its goal, from basic intentions directed at simple motor goals to superordinate intentions directed at somewhat complex or general goals, the achievement of which typically involves the completion of a number of subgoals or substeps. A second relevant dimension of variation concerns the target of the intention. Intentions may target inanimate objects but may also target a third party or be achieved in a context of social interaction Intentional contents are thus also modulated by the relational structure in which an action takes place.
I will argue that these two dimensions of variation may be crucial to understanding how the respective contributions of visual kinematics and prior knowledge interact in the intention reading process and how their contributions are modulated. In particular, I will present some recent empirical results providing evidence for an increased contribution of prior knowledge to the reading of intentions with larger scope and social targets.