โ Scribed by David Cole
No coin nor oath required. For personal study only.
Thinking about thinking and consciousness and the relation of the mental to the physical seems, not surprisingly, to invite thought experiments. Most of these experiments are presented in the course of an argument, and a conclusion is drawn. Here I wish to consider a family of thought experiments, propounded by philosophers from Leibniz to John Searle, along with some new ones intended to shed light (and dark) on the others and on thought, functionalism, and understanding.
My intent is critical. Thought experiments have been used over the past 250 years purportedly to show that no machine could embody thought, understanding or sentience. The machines which, it has been argued, are incapable of these mental manifestations have included mechnical, biological (the human brain), and most recently digital electronic systems. I will argue that these arguments, based on thought experiments, do not succeed.
In particular, John Searle's recent critique a of research in artificial intelligence (A.I.) turns on a thought experiment reproduced below. In a nutshell, Searle's 'Chinese Room' argument is this: since a human simulation of a machine (digital computer) simulation of human behavior which in humans normally evidences understanding would not itself involve that understanding, neither does the machine simulation.
Searle's argument is the most recent and sophisticated of those I consider and will command most of my attention. But I will argue that it is a development of earlier arguments and shares their central weakness. I will criticize Searle's argument on three grounds. First, it is not clear, despite Searle's denials, that his imagined simulation of a machine would not produce understanding. Second, there is an important disanalogy between the machine simulation of a human performance and the human simulation of the machine, and this, in part, accounts for the mistaken conclusion which he draws about the case. And third, if this sort of analogical argument were valid, similar arguments could be constructed (thought experiments 6 and 7, below) against
๐ SIMILAR VOLUMES
The idea that claims about the physical world might be arrived through a priori reasoning has a long history in physics. But it is clear that empiricist notions of the natu~Te of science, and in particular the empirical nature of physics, have held sway in this century. Yet, in the idea of thought e