X-Message-Number: 21670 Date: Sat, 26 Apr 2003 00:56:18 -0700 From: Mike Perry <> Subject: Qualia and Benefit of Doubt Robert Ettinger writes >Absence of evidence may not be conclusive evidence of absence, but it tilts >the scales. When we know the anatomy and physiology of qualia, a system >lacking such should be presumed without feeling. To "know the anatomy and physiology of qualia" appears to mean knowing how qualia or feelings occur in human beings and other applicable, biological life forms--but no more. A system that appears to express feeling may be really doing so ("success") or just imitating ("failure"). So, to show that a system possessed certain anatomical and physiological traits would amount to a proof of success for that system. But it would not decide the question for some other system that lacked these features, a robot for instance, which we assume also seemed to express feelings. It would not then constitute proof of failure for the second system, yet here it is suggested we should presume failure in the absence of proof of success. I imagine I would react differently if acquainted with an actual case of an emotive-seeming robot. My natural reaction would be to presume success--that it *did* have the feeling it seemed to have--unless confronted with proof of failure. If confronted with the possibility of replacing my brain or an important part with an artificial equivalent (which offered significant advantages such as greater durability, say) I think there would be cause to be more conservative, for more would be at stake. But I would not go so far as to demand proof of success. Instead I would accept proof of the impossibility of proof of failure, or something close to this, if such could be gotten. So as long as it seemed unlikely that one could ever conclusively decide against success, I would probably go for the switch. >An important point is that isomorphism is never one-to-one. except in the trivial case of identity, which need not detain us. >The simulation >and the simulated always differ in some ways--otherwise it wouldn't be a >simulation but a duplicate (and even then there would be at least small >differences). It is always just a subset of one mapping on to a subset of the >other. So--how can you confidently assert that only the congruences are >important and not the differences? This is a good question, and the short, incomplete answer is "because you are confident you have the *right* isomorphism." Of course this begs the questions of what the "right" isomorphism is and how do you know that. These I think must be answered on a case by case basis. In the case at hand, we are asking, as part of a thought experiment, whether a certain system exhibits feeling because it is isomorphic, in a certain way, to another system that we agree has feeling. We want to know whether certain suggested isomorphisms would preserve success. As an example, we might have a robot brain which would be an inorganic construct with an organic analogue that could, in principle, also exist. (In practice normally it would not.) The organic analogue, if it did exist, would exhibit consciousness and feeling (again, success). So would the robot also have these traits? Would the isomorphism preserve success? If it mapped events at the quantum level, that is to say, for each possible particle interaction or other quantum event in the organic analogue it assigned a corresponding event in the construct, with appropriate relations and properties holding, I would be inclined to say yes. I think here that failure could never be demonstrated, though quite possibly success could never be demonstrated either. If so I would give the benefit of the doubt. I would accept the isomorphism as one of the "right" kind, in this one case at least, and accept the system as having the feeling it seemed to have. Other cases might be more difficult to decide, but once again I would invoke benefit of doubt and presume success as long as proof of failure seemed unlikely. So one may ask, why do I have this orientation in the first place? But that is a long story and it's getting late--another time. Mike Perry Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=21670