The Intelligibility of Action

Abstract

When I play chess with my son, my son plays chess with me. When I play chess with my computer, my computer does not play chess with me. And this in spite of the fact that my computer is a more skilful chess-player than either my son or I. Some have supposed that the only ground which would warrant such an assertion would be a compelling argument to the effect that there are specific types of task which human beings can perform, but machines cannot. But this seems to me to be mistaken and in any case it is clear that we possess no such compelling arguments. Others have claimed that it is because machines must lack anything adequately resembling human consciousness that we are not entitled to understand their performances as actions. For reasons that will emerge much later in this paper this too seems to me to be mistaken. What I shall claim is that characteristically human actions have the property of intelligibility, although some human actions may lack that property without thereby ceasing to be human actions; whereas the performances of machines are neither intelligible nor unintelligible. What then do I mean by intelligibility?

I am indebted for penetrating and constructive criticisms of an earlier version of this paper to Amélie Oksenberg Rorty, Wesley Felix and especially Joseph Margolis.