X-Message-Number: 12981 From: Date: Wed, 22 Dec 1999 21:13:06 EST Subject: being there Let me try again to throw a bit of light on one of the mistakes of some of the "strong AI" people. Daniel Crevier has quoted or paraphrased Moravec (MIND CHILDREN, 1988), writing : ----------- "you are operated upon by a robot surgeon which analyses a small part of your brain, and constructs an electronic circuit (either analog or digital, it doesn't matter) with the same input-output properties. He then installs the circuit in your brain in such a way that you can switch between it and the original tissue. In this way, you can verify the accuracy of the simulation, which is tuned until the circuit and the original tissue feel exactly the same to you. When this is achieved, the robot removes the original tissue and wires in the circuit permanently. This procedure can be repeated on all parts of your brain, little bit by little bit, in such a way that at every step you are in a position to verify that you are still really yourself." -------------- This is similar to one of my thought experiments in THE PROSPECT OF IMMORTALITY (1962 et seq; the 1964 version is available in full on the CI/IS web site). But it doesn't show what Moravec and Crevier think it shows--that awareness is basically just data processing, and that all significant human data processing can in principle be done electronically. First, the signals sent between one part of the brain and another, and between the brain and other parts of the body, are not only electrical, but also chemical and perhaps mechanical. An "electronic circuit" therefore cannot entail the same "input-output" properties. It could simulate (describe) them, of course, given enough space and time; but a description of a thing is not the same as the thing itself. Whether a description (or a partial isomorphism) is "just as good" as the thing itself is an unanswered question, along with other "philosophical" questions such as the identity of duplicates and other dilemmas of continuity. We simply don't have the answers yet, and it is unscientific to pretend we do just to avoid the discomfort of saying "I don't know." Second (yes, this is partly redundant), the thought experiment dodges the question of time and space relationships. I'll omit repeating the reasons here, but it is reasonably clear that awareness (based on feeling) must bind time and space. A computer--especially a serial computer--does not do this. According the the hard-core strong-AI people, it doesn't matter. Even a basic, low-tech Turing computer--a strip of paper moving back and forth with marks being written and erased on its squares--would feel and think, they claim. One must admire the audacity of the thought, but not the stubbornness that refuses to admit its weaknesses. There are even deeper questions. David Deutsch and others tend to believe that one should not speak cavalierly about what is possible "in principle"--that there is no disjuncture between the physically possible and the logically possible. Again, remains to be seen, and not soon either. Finally, look again at the words, "In this way, you can verify the accuracy of the simulation, which is tuned until the circuit and the original tissue feel exactly the same to you." If qualia reside in the "self circuit," and if that is a special construct such as a particular kind of (electro-chemical?) standing wave, then when that is removed feeling will cease, regardless of what electronic signals continue to be sent between parts of the brain. I realize I have said all this before, many times, but I am searching for phraseology that will be more effective. Hope springs eternal. Robert Ettinger Cryonics Institute Immortalist Society http://www.cryonics.org Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=12981