Abstract

In this article I argue that the intentional explanation of actions is a central task of the social sciences. However, the attribution of intentional states is highly underdetermined by observational data. Such attributions are forms of inference to the best explanation, and they have to cohere with our background knowledge of the conditions under which the agent acts. In addition to this ‚outer coherence’ we have to ask for an ‚inner coherence’ within the intentional profile of the agent to gain a meaningful understanding. Both forms of coherence can help to reduce underdetermination.