- Slides: 34
First, Scale Up to the Robotic Turing Test, Then Worry About Feeling
Consciousness is feeling,
and the problem of consciousness is the problem of explaining how and why some of the functions underlying some of our performance capacities are felt rather than just “functed. ”
But unless we are prepared to assign feeling a telekinetic power (which all evidence contradicts),
feeling cannot be assigned any causal power at all.
We cannot explain how or why we feel.
Hence the empirical target of cognitive science can only be to scale up to the robotic Turing Test.
Consciousness is Feeling Performance Biochemical robots The other-minds problem The Causal Role of Feeling The mind/matter problem Forward &Reverse Engineering Vitalism Correlation and Causation Feeling Versus “Functing”: How and Why Do We Feel? Telekinesis
• Are models of consciousness useful for AI? • Are AI systems useful for understanding consciousness? • What are theoretical foundations of machine consciousness? • Is machine phenomenology possible? • Will conscious systems perform better than unconscious systems? • What are the implementation issues of current AI systems inspired by
Can a machine be conscious? (How? ) Depends what we mean by “machine? man-made devices? toasters? ovens? cars? computers? today’s robots? "Almost certainly not. "
Why “almost"? (1) “empirical risk. ” (2) “other-minds” problem” Descartes on certainty and uncertainty: Certain: (i) mathematics, (ii) my-own-mind (the "cogito"), and Uncertain: everything else other-minds problem haunts robotics
"Is any machine we have built to date conscious? " "Can a man-made artifact be conscious? " Genetic or other biological engineering? Toasters: man-made vs. tree-grown
Machine = any causal physical system Includes biological organisms Hence we are conscious machines. So the right question is : “What kinds of machines can and cannot be conscious, and How? ” (cognitive science) (And How can we tell? The "other-minds problem. ”)
"forward-engineering" vs. "reverse-engineering" causal explanation How?
Explaining a cardiac system (a heart): F-eng: build a mechanism that can do what the heart can do R-eng: F-eng + explain (and/or build) the structure and the function of the biological heart itself: what the heart is (made out of) how it in particular happens to do what hearts can do. Explaining a conscious system (the brain): F-eng: build a mechanism that can do what the brain can do R-eng: F-eng + explain (and/or build) the structure and the function of the biological brain itself: what the brain is (made out of) how it in particular happens to do what brains can do. Either way, the causal explanation is a structural/functional one.
The ghost of the other-minds problem: Note: Cardiac research program is completely unproblematic. The “cardiac vitalist” asks: "Can a machine be cardiac? " (“What kind of machine can and cannot be cardiac, and how? ”)
At no point would the cardiac vitalist have any basis for saying: "But how do we know that this machine is really cardiac? " There is no way left (other than ordinary empirical risk) for any difference even to be defined. Same would be true if our question had been "Can a machine be alive? " When we ask -- of the man-made, R-eng clone: "But how do we know that this machine is really alive? "
Two structurally and functionally indistinguishable systems, one natural and the other man-made, their full causal mechanism known and understood: What does it mean to ask: "But what if one of them is really alive, whereas the other is not? " What property is at issue that one has and the other lacks, when all empirical properties have already been captured? Vitalism sounds like the other-minds problem -- and may be the other-minds problem: the animism at the true heart (soul) of vitalism
The (nonanimist) vitalist who accepts that plants are not conscious would be in exactly the same untenable position if sceptical about the R-eng plant as the sceptic about the R-eng heart: What vital property is at issue (if it is not consciousness itself)? But the same is most definitely not true in the case of consciousness itself.
Forward-engineering the brain: Build an F-eng robot that passes the Turing Test: It can do everything a real human can do, for a lifetime, indistinguishably from a real human (except appearance: we will return to that). Is it really conscious? It is indistinguishable from us in everything it can do. But conscious is something I am, not something I do. In particular, it is something I feel; indeed, it is the fact that I feel.
The sceptic wants to say that F-eng is the wrong kind of machine, that it lacks something essential that we humans have. The robot does not feel, it merely behaves -- behaves exactly, indeed Turing-indistinguishably -- as if it feels, but without actually feeling a thing.
Whether or not a Turing robot feels is an ontic question rather than merely an epistemic question. Having conceded this point regarding certainty, however, only a fool argues with the Turing-Indistinguishable:
Indistinguishable? If you prick us, do we not bleed? Perhaps the sceptic about the F-eng machine should hold out for R-eng machine, made out of the right stuff, Turing-indistinguishable both inside and out at both the macro and micro levels. But would we be right to kick robots that don’t bleed, because we infer that they therefore don’t feel?
The Other-Minds Barrier TELEPATHY: mind-reading vs. turing-testing
"theory of mind" or "mind-reading" in animals and children. (not philosophy or parapsychology but "other-mind perception”) “Mind-reading” is Turing-testing: inferring mental states from behavior. (Language (also a behavior) is the most powerful and direct means of mind-reading!)
The only mind we can read other than by Turing-testing is our own! Don’t make ontic/epistemic conflation here: Turing-testing does not mean that all there is to mind is behavior (as the blinkered behaviorists thought)! It means the only way to read others' minds is through their behavior.
Indistinguishable? What about R-eng. and rest of the neuromolecular facts about the brain? Which facts?
"What kinds of machines can and cannot be conscious? " We know brains can be, but how? What are their relevant properties (if their weight, for example, is not)? If we pare down the properties of the brain to test which ones are and are not needed to be conscious: What will our test be? We are right back to Turing-testing again, the only non-telepathic methodology available to us, because of the other-minds problem.
Will "correlations" do instead? Brain imaging to find the areas and activities that covary with conscious states, as the necessary and sufficient conditions of consciousness? But how did we identify those correlates? Because they were correlates of behavior: and of our own feelings i. e. , by Turing-testing. What is our basis for favoring R-eng over F-eng then, if they are Turing-Indistinguishable (behaviorally) and the Turing Test is our only face-valid criterion.
Conclusion: "What kinds of machines can be conscious (and how)? " The kinds that can pass the Turing Test, and by whatever means are necessary and sufficient to pass the Turing Test.
Two consolations for residual worries about Zombies passing the Turing Test: (1) Darwin is no more capable of telepathy than we are. Evolution can’t distinguish sentients from zombies either. (2) the problem of explaining how (and why) we are not zombies (the “mind/body problem”) is “hard” (probably insoluble)
Why the Mind/Body Problem is “hard”? Causality The Scylla of TELEKINESIS: Feeling as a Primal Force The Charybdis of EPIPHENOMENALISM: Feeling as an Inert Fact
It would be easy if telekinetic powers existed: Then feelings would be physical forces like everything else. But there is no evidence that feelings are causal forces. Hence both F-eng and R-eng can only explain how it is that we can do things, not how it is that we can feel things. And that is why the ghost in the machine is destined to continue to haunt us even after all cognitive science’s empirical work is done. This paper (and linked references) is available at: http: //www. ecs. soton. ac. uk/~harnad/Temp/machine. htm
First, Scale Up to the Robotic Turing Test, Then Worry About Feeling Consciousness is feeling, and the problem of consciousness is the problem of explaining how and why some of the functions underlying some of our performance capacities are felt rather than just “functed. ” But unless we are prepared to assign feeling a telekinetic power (which all evidence contradicts), feeling cannot be assigned any causal power at all. We cannot explain how or why we feel. Hence the empirical target of cognitive science can only be to scale up to the robotic Turing Test.