Abstract

How could one justify the claim that a machine had a mind? This "problem of machine minds" is the subject of my dissertation. ;It begins with an examination of the traditional answers to the closely related problem of other minds. One way to solve this problem is to infer that other people have minds by drawing an analogy between oneself and others. This sort of approach can lead to an unacceptable skepticism about other minds. On the other hand, behaviorism tells us that one can be logically certain that others have minds. However, behaviorist theories of the mind are generally too liberal in their attributions of mentality. I argue that neither physicalism nor functionalism offers a solution to the problem of other minds that avoids both excessive skepticism and excessive liberality. ;The dissertation then examines two distinguished positions on machine mentality. John Searle's "Chinese Room Argument" purports to show that it is impossible to program a machine to understand human languages. It is shown that the Chinese Room Argument exploits the skeptical possibilities of the argument from analogy for other minds. In contrast, Daniel Dennett's theory of the mind is excessively liberal. This liberality makes his attributions of mentality to machines unconvincing. ;I draw on insights from both Searle and Dennett to produce a set of conditions under which we would be justified in attributing mentality to a machine on the basis of an argument from analogy. The importance of these conditions is explained and defended. I describe the general functional and behavioral characteristics of a robot that would satisfy these conditions. I then use an analogical argument to justify the literal attribution of thoughts, beliefs, and desires to this machine. This argument avoids Searle's excessive skepticism about machine minds by discounting the importance of physical differences between ourselves and machines, and it avoids Dennett's excessive liberality by taking the possession of interests to be necessary for rationality. Proper regard for the argument from analogy thus helps us to resolve disputes about the alleged mentality of machines