Dr Graham White
The focus of my work is the problem of reasoning about human action. This is a theme which has repeated been tackled, both from the side of philosophy and from the side of AI: it has proved remarkably resistant to solution. Any successful formal account of action would have to do at least the following: one would want to know what the contexts of action were (that is, what the subject knows when acting), and one would want a logical system, parametrised by contexts of action, which would represent the sort of reasoning, about our own actions and those of others, which we are capable of. There are a number of criteria against which one would measure such an account. The mere construction of such a system is quite difficult, and imposes quite a lot of constraints of its own: so do standard metatheoretical desiderata. For example, that the system should be independent of the linguistic primitives that one uses to construct it, and that, too, imposes significant constraints on the formalism. There is also a good deal of empirical evidence, mostly from linguistics. Talmy's work in cognitive semantics describes important linguistic universals governing human language about action, and this ought to play a role. Similarly, the theory of speech acts, and particularly Grice's work on implicatures, ought to be formalisable in a successful theory. (Both of these requirements are extraordinarily ambitious.) What I have already done: I have a description of contexts of action which, although not a complete account of real-world concepts, has a good proof theory. It also allows an interpretation in terms of the sort of equational reasoning that Davidson describes in his paper The Logical Form of Action Sentences. I also have an interesting logic of explanation which, when applied to explanations of action, recreates the standard solution of the frame problem. What I am planning to do next: I hope to extend the work on equational concepts of action to include a sort of Heideggerean account of habitual action. I also want to write a philosophical paper on my logic of explanation and show how it deals with Fodor's problems with nonmonotonic logic. Longer term plans: to incorporate this work with work on causal reasoning about action, and particularly with work on actions causing other actions. This will involve more work on contexts of action: I have some conjectures on what suitable contexts might be. What use is this, anyway? A lot of my recent work on action could probably be translated fairly straightforwardly into description logic; this work, then, could be quite semantic web-friendly. Work on speech acts is also important for software agents: agent communication needs some version of speech acts, but as yet there have been no systems with good theory.
Member of London Mathematical Society
Member of American Philosophical Association
Invited Chapter 2004-01-01 Invited chapter on "The Philosophy of Programming Languages" in The Blackwell Guide to the Philosophy of Programming and Information (L. Floridi ed.)
Invited encyclopaedia article 2002-01-01 Invited chapter on Medieval Theories of Causality in The Stanford Encyclopaedia of Philosophy. Available at http://plato.stanford.edu/entries/causation-medieval