TOWARDS INTERPRETATION l PHILIP PETTIT Dimitri Ginev (1994) argues that the notion of 'normalizing' explanation that I introduced in some earlier papers (Pettit 1986a, 19861)) does not capture the hermeneutic idea of intentional interpretation. I agree; it doesn't. If I suggested otherwise, as Ginev thinks, then I was wrong. So how do normalizing explanation and intentional interpretation relate to one another? What places do they occupy, respectively, in the large network of concepts of explanation? I shall try to deal with these questions here, though in a relatively unargued mode. Answering the questions requires a big picture and a big picture needs some bold strokes (see too Pettit 1993), My paper is in three sections. In the first section I present intentional interpretation as a kind of 'programming' explanation; in the second I cast it, more specifically, as a sort of normalizing explanation; and in the third I characterise it, more specifically still, as the sort of programming and normalizing explanation which directs us to the content of an agent's thoughts and deliberations. 1. Interpretation as programming explanation Intentional interpretation involves the attempt to explain an agent's speech or behaviour by reference to distinctive psychological states: roughly, by reference to states that reflect the information to which the agent gives countenance and the inclination that moves him; by reference, as the stock phrase has it, to beliefs and desires. The first thing to be said in characterisation of such explanation is that it invokes higher-level causal factors, not factors that operate at the most basic level there is. 157