Watson, Come Here, I Want You: featuring Alan Turing, Pontius Pilate, Henry Higgins and the midcentury mainframe of your choice
Does Watson know that he won? Ohhh, I’m sorry…in linking thinking to biology, John Searle offers an answer that is not in the form of a question.
Watson did quite well on Jeopardy, as it was designed to do. Machines tend to play games well, especially when the games involve logical or mechanical procedures, i.e., rules. Or building SUVs. But did doing well mean it knew the answers? Does it know things? Does it know it won? Can we infer self-awareness, consciousness, from a pattern of perceived causes and effects?
Can Watson say, with Rene, “I think, therefore I am?” Or can it only display it?
For that matter, can you?
The problem of other minds extends to the issue of artificial intelligence, rather than the other way around. I do not have privileged access to your mind the way I do to my own, so how do I know that you think? I don’t know it, I can only infer it, from a pattern of behavior and belief-reports on your part. Aside from that, I have to take it on faith. Since this is the same procedure whether the other in question is you or a reasonable facsimile, then how do I know a machine does not think, if it produces a similarly believable pattern of behaviors and claims about itself?
This is the perspective of Alan Turing, as he designed his eponymous Test: behind each of two doors is a man and a machine. The questioner (let’s say, you) submits typed questions to each through a tiny slot (no peeking!), and typed answers are returned the same way. If you can’t guess from the answers which is what, and what is who, then thinking must be attributed to the machine the same way it is attributed to the man. Of course thinking and consciousness might not be the same thing. But regardless, so far no machine has passed the test.
But so far does not have a lock on never, and whether or not the Turing Test will be our ultimate Test, the question is still an open one. So Searle’s recent cri de coeur in the Wall Street Journal turns out to beg the question it claims to consider. Despite observing consciousness’ relationship to biology (so far), Searle does not succeed in establishing a necessary link between the two. “I am” may be an epiphenomenon, or even a congeries of epiphenomena, upon multiple foundations; some carbon-based, some silicon-, and some, perhaps, even more distributed and less tangible.
Cyber-resurrections of 16th century esoteric philosophers, take heart.