X-Message-Number: 7956 From: (Thomas Donaldson) Subject: Re: CryoNet #7941 - #7944 Date: Thu, 27 Mar 1997 00:17:13 -0800 (PST) To Mr. Dratzman: I don't know how much of the discussion you have followed, but there are a couple of points I would make. Going backwards, the Turing Test has faults. Its main fault is that it is far too based on language rather than behavior. (I'm hardly surprised that some thought that humans were computers, using it, either --- for exactly that reason). Clearly the Turing Test also assumes that the computer (or person) shares a language with the person on the other side. That is, they have something in common. If we are allowed to admit that somehow the computer and me (the interrogator?) have something in common, here is a test which I would take much more seriously. I would require that the computer, or something guided by it by some means, be able not only to converse with me but to take several walks through city streets and perhaps also through large parks, and we would both discuss what we had seen. At the end of these walks I would take it home with me and we would both prepare a small snack. If the computer can't eat what I eat, then perhaps there is some other thing which it or its "moving version" might do --- say, oil itself, or plug itself into a power outlet, or whatever. Naturally we would continue our conversation. Then I would bid it goodbye and it would find its own way back. (It's OK if it has to ask me for a map first). If it can do all of these things, and everything they imply, then I would seriously consider it to be aware. The point of this test, unlike the Turing Test, is that it must not only interact with me, but also interact with the world, in a serious way, not as a game. The city we walked through would not be virtual, nor would the park. I will summarize what I am saying. Because everything we do is real rather than symbolic (when stripped down to its basics) we must interact with a world which cannot be defined fully by any amount of symbols or theory. Clearly some systems of interacting atoms can do this, and some are aware. But then trees are systems of interacting atoms, and they are not aware. I am saying first of all that requiring that such a system be aware imposes some conditions on its structure. Those conditions are much stronger than simply the ability to manipulate symbols --- which is exactly what the objects we normally call computers can do, and do very well. It must be able to interact with the real world; and if we are talking about computer intelligence rather than simple awareness, it must do so at least as well as a human being. The problem with symbols is that they have no INTRINSIC meaning to anyone. Some other person must be present to interpret the symbols produced by a computer program. That interpretation, one way or another, connects them somehow to something in the world... but it is not the computer which does that, it is its user. Not only that, but symbols by their nature can only have a floppy, fuzzy, changeable interpretation. God never made a dictionary, we did. And ultimately words cannot be defined by other words. (One problem with the Turing Test is that it plays on our human tendency to react to symbols --- that is, the language produced by the person or computer on the other side --- quite automatically, as if they automatically meant something to the object/person producing them). And I'm actually making a strong statement here: not about computers --- I think that this part of what I am saying should be trivial, once we liberate ourselves from the notion that symbols are real independent of us -- but about awareness. WE are structured in a particular way which allows us to interact with the world. So are mammals. And I am saying that a brain which can do that as well as we do, or even similarly, will also produce awareness. In animals, perhaps not intelligence --- though compared to most animals, a rat is really quite bright, but awareness. Perhaps as we explore the universe further we will find creatures which are not aware, but can still function in the world. That will be interesting. But we have NOT made them when we made computers. As for artificial intelligence or artificial awareness, I have no problem with either. My problem is with the notion that we could program computers to have either. Robots, though, they are a separate question, and the more independent we can make those robots, the more awareness they may have. Long long life, Thomas Donaldson Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=7956