mcdermott-drew@cs.yale.edu (Drew McDermott) (10/19/90)
In article <VINSCI.90Oct18224945@nic.soft.fi> vinsci@soft.fi (Leonard Norrgard) writes: >Your wrote: >>[...] >>Seriously, though (and maybe I should say that I haven't looked at >>comp.ai in a couple of years), what could ever be sufficient evidence >>for machine consciousness? > >How about first finding evidence that other humans are conscious? When >you have that, apply it to machines. (Hint: #1 is a little bit troublesome.) Okay, it's time for: Why There is No Other-Minds Problem Copyright (1990) Drew McDermott \section{The Traditional Other-Minds Problem} The other-minds problem is supposed to be something like this: How can I tell whether there are any minds in the universe besides mine? What I wish to argue is that this problem cannot be coherently stated, and hence is no problem at all. Let me hedge a bit. There is an other-minds problem to the extent that there is an ``outside-reality'' problem. Descartes (or maybe Aquinas) was the first to raise the issue whether all of reality might be an illusion. All I can know about the world comes through my senses, so perhaps some Evil Deceiver is feeding systematically false information through my senses, and there is in fact no real world at all. This kind of Total Skepticism is supposed to be independent of the other-minds problem. That is, someone could pooh-pooh Cartesian doubt about reality as a whole, and still claim to entertain doubt that there are any minds but his. It is this claim that I dispute. I begin by observing that, to the untrained observer, there is a lot of mind around. Every other person has a mind (excepting the usual people in comas, and such), and many animals have minds, too. Indeed, the naive observer is likely to get carried away and see mind at work in the actions of storms and planets. As we get sophisticated, we are able to draw distinctions more carefully, and we get less tempted to ascribe mind to the weather. However, we never get tempted to discard the concept of mind completely. (Contrast the concept of God, which many people are sincerely willing to discard.) One may as an idle speculation suppose that mind is much scarcer than it appears to be, but when it comes to dealing with other creatures we never cease for an instant to assume that they have minds. So why in the world is there some issue about whether other minds exist? I think the problem is arrived at like this: I know I have a mind. (I think, therefore I am.) But my belief that others have minds is due to a chain of inference that might be faulty. I see someone heading for the freezer, and I infer that she wants ice cream. I see someone writhing on the floor, and I infer that he is in pain. But the inferences could be wrong. These people could be just going through the motions, and not have minds at all. There are two flaws with this way of stating the problem. First, the perception of mind in these cases sometimes requires a deep chain of thought, but usually it does not. Sometimes we have to infer that someone is in pain by observing how stiffly he moves, but usually the pain is just as observable as the stiffness of motion. We see enough of the phenomenon to recognize it; from the part we infer the whole. It's like seeing the front of an animal and inferring the presence of the entire animal. We are occasionally wrong, but we couldn't be wrong all the time without recourse to the sort of boring skepticism I put aside early on. In other words, it's just a faulty analysis of the situation to assume that the subjective experience is the primary definition of mind, and all other instances are recognized by inference to the primary definition. Mental phenomena tend to have three aspects: subjective experience, behavioral consequences, and implementation mechanisms. There is no particular reason to assign primacy to any one of these aspects (except for a hangover from Cartesian philosophy; see below). Mental phenomena are ``natural kinds.'' As Hilary Putnam has pointed out, such entities are defined by ostension to some extent, and not defined by a set of necessary and sufficient properties. At one point in history, water was known as a certain abundant transparent liquid. Later we found out it was H_2O. The result was not an inference that in the presence of water one often finds H_2O, but instead the discovery that water {\it is} H_2O. We have similar discoveries to make about mental phenomena. Consider an analogy to photosynthesis. Suppose humankind had noted that plants were capable of surviving on sunshine, water, air, and dirt, and had given the name ``photosynthesis'' to this ability. (This is not intended to be historically accurate.) Then people would have expected, with much justification, that when plants were opened up and examined carefully, mechanisms capable of photosynthesis would be found. Similarly, we have no reason not to expect mechanisms capable of mental phenomena to be found when we open up brains. Of course, in each case, the mechanisms responsible do not leap out at us, although by now we've explained photosynthesis pretty well. For some reason, when people talk about opening up brains, they start talking about looking for ``correlates'' of mind instead of ``mechanisms'' of mind. The only possible reason is some preexisting philosophical bias. Anyway, the first flaw may be summarized thus: The perception of mental phenomena does not require an inference {\it from} behavior {\it to} subjectivity. Evolution has shaped us to recognize certain real things, such as the three-dimensionality of the world. Mental phenomena are one such thing. It takes special training and perversity to pretend that all we really are sure of is the subjective aspect of mind, and everything else is mere evidence for this aspect. It's just like pretending that two-dimensional images are all we're really certain of, and the three-dimensional world is an ``inference'' from it. In the nineteenth century, psychologists really did try to convince themselves of such statements, based on certain radical-empiricist preconceptions. We've gotten over those preconceptions. I hope we've gotten over the corresponding ones for the perception of minds. The second flaw in the statement of the problem is that the idea of ``just going through the motions'' is not sufficiently clear. It does not mean ``faking it.'' Suppose someone were faking pain or intentionality very convincingly. Then that would not mean he didn't have a mind; quite the contrary. I think the phrase is supposed to mean ``behaving automatically, like a robot.'' But this phrase is still ambiguous. If it means, ``behaving like a Walt Disney audioanimatron,'' then it's quite obvious that people are not ``just going through the motions,'' because Disney technology is not sophisticated enough to duplicate people. No one above a certain age attributes mind to the Abe Lincoln in the Disneyworld Hall of Presidents. It just cannot be the case that a person is ``just going through the motions'' in this sense. If the phrase means ``behaving like the most sophisticated robot it is possible to build,'' then it is not clear that there is actually a contrast between ``really having a mind'' and ``just going through the motions.'' That's because we don't know what the most sophisticated possible robot is capable of. If you simply want to assume that there is a contrast between having a real mind and being a robot, then you have assumed away a very interesting question. You haven't said anything about what the contrast {\it is}. I will say more about this in Section 2. I hope at this point you are feeling a certain frustation. ``I know the distinction I want to make, but I just can't put it into words.'' Try this: Suppose other people are just hollow shells; there's nothing ``inside.'' But this isn't getting us anywhere. We know there's a lot inside. Okay, about this phrasing: As granted above, mental phenomena have three aspects, subjective, behavioral, and implementational. Why couldn't it simply be the case that the first one is often or always absent? The only agent I'm sure has subjective experience is me, so what's my evidence that anyone else has? The problem with this objection is that it misconstrues {\it aspects} as {\it parts}. This construal would make sense for problems like the ``other-planets problem'': We observe one star with planets and ask what our evidence is that some or any other stars have any. But subjective experience is not an appendage of mentation in this way. A mental event can be experienced from several different angles, including (although the categories are fairly arbitrary), the subjective, the behavioral, and the implementational. But the experiences are all {\it of the same event}. For example, consider a parent dealing with a child's earache. Both persons are experiencing the same earache. We could if we wanted sort the experiences into bins. One classification would be into behavioral, subjective, and implementational. (Behavioral: an observation of a squeal; subjective: an observation of a twinge of pain; implementational: $\ldots$ unlikely, under the circumstances). Another would be into who owns what (my squeal; her squeal; our twinge). On any occasion any angle of observation could be missing, but to suppose that a particular angle is missing {\it normally} is like supposing that objects normally do not have insides. Once we understand geometry, we do not need to cut open objects and keep a tally in order to verify that every object has an inside; having an inside is part of what it means to be an object in the sort of space we inhabit. Of course, some objects could actually be weird Klein-bottlish things which, when cut open, reveal their outsides again. It's not logically necessary that all objects have an inside, nor is it logically necessary that all agents with minds have subjective experience. It's just extremely likely that agents capable of experiencing the world would also experience some of their own workings. My problem, you now might claim, is that I'm ``objectifying'' subjectivity. The issue is not whether we can observe our own mental phenomena, as someone might observe aircraft, but whether we can {\it feel} them. If I were {\it counting} twinges of pain, then I might grant that your pain was as in principle observable as mine, but when {\it experiencing} them the issue of whose they is takes on an altogether different meaning. My ``raw feels'' are intrinsically {\it private}. I can observe my own feelings --- or, more precisely, I can {\it have} my own feelings --- in a way that no one else can. And other people's feelings are forever hidden from me. So what makes me think those feelings exist? Our intuitions, through centuries of training, have become warped regarding this point. Suppose a brain surgeon has my skull open, and touches an electrode to a particular point, causing a pleasant tingling sensation. Suppose the surgeon is equipped with a good theory of how the brain works (much better than what we now have), so that she can see the reverbations die away, and get reported to the memory log as pleasant, exactly as I report subjectively. We're both observing my sensations directly, but with different ``instruments'' and from different distances. The pleasant tingling sensation seems quite different to me than to her (e.g., it's not particularly pleasant for her), but then again my voice sounds different to me than to her. It's true that she's not ``having'' my feelings, but only because she's not me. If she wished, she could probably wire up an apparatus to cause her brain to experience my brain's states in a way very close to the way my brain experiences them, but she still couldn't have my feelings, any more than she could have my location. The fact is that other people's sensations are more ``directly'' observable than a lot of things we grant reality to with much less ``direct'' evidence, like quarks or electromagnetic fields. So why are we so curiously persnickety about the matter? I think the answer ultimately derives from the epistemology of philosophers like Descartes and Kant. Those guys took it for granted that each of us is in the strange epistemological position of having to reconstruct the world from our sense data. This theory gets us into some ridiculous quandaries (which is why --- in case you haven't heard --- most philosophers have abandoned it). The most ridiculous quandary is that you have to be reconstructed from my sense data, and I have to be reconstructed from yours. On this theory, it's easy to say who has a mind: Those doing these reconstructions are the actual minds, and the other things in the world are just passive objects of recognition. The other-minds problem then just becomes the question of how many reconstructors there are. In other words, when I see someone with an apparent mind, I have to ask ``Is he actually building the world from sense data, or is he just an aggregation of --- or inference from --- my sense data?'' Unfortunately, while this may have seemed an acceptable question to Descartes, I assume (er, I hope) that any modern person would be embarrassed by it. Taking this question seriously presupposes the ``movie theater'' view of the world, in which the human race consists of a bunch of consciousnesses all watching the same movie (the world). What's absurd about this view is that we have to reconcile it with the fact that we're all {\it in} the movie. The original reconciliation was to assume that it's only our bodies that are in the movie --- the minds are out in the theater, connected up to the bodies somehow. If you want to buy this picture, then, yes, you have an other-minds problem (namely, how many seats are occupied?). But in the twentieth century it's gradually becoming clear that the minds and the bodies are all mixed up together. So we're left with frustation: How do you state the other-minds problem? \section{The Individual Other-Mind Problem} You may be willing to grant that the traditional other-minds problem is ill-posed, yet claim that when faced with a particular entity we still have the problem of determining whether it has a mind. This is the ``individual other-mind'' problem. The usual way this problem is posed is to present a Gedanken experiment involving an intelligent robot, about which there is doubt whether it has a mind. What's odd about this setting is that when the individual other-mind problem comes up in everyday life, what we are confronted with a is a {\it borderline case}, such as a nonhuman mammal or a person in a coma. The evidence we cite one way or the other has to do with completely mundane phenomena like whether the creature appears to fear pain, or shows brain activity of a certain kind. There is nothing special about the individual other-mind problem in this setting. It's like the problem of distinguishing between animals and plants. We believe there's a difference; we believe there are borderline cases; we believe that by studying animals and plants more we'll learn more about how to --- and whether to --- classify the borderline cases. End of story. By contrast, in the Gedanken experiments I referred to, we are supposed to imagine the existence of a creature whose mental status would not be in doubt at all, if such creatures could actually be created. Suppose in the remote future you had an acquaintance who was a computerized mathematician, a pretty good mathematician, whose talent was somewhat blunted by its preoccupation with whether other mathematicians had plagiarized its results. This computer would not be able to walk around, but it would be able to attend conferences (via remote TV hookups), and carry on a completely colloquial spoken conversation about mathematics, and about the daily affairs and conspiracies of mathematicians. I've painted this picture for a variety of reasons, besides the fact that it's fun to engage in this sort of science fiction. What I'm trying to convey is the idea of an entity that probably couldn't pass Turing's Test, but that no one would deny was intelligent. (The computer might rebuff all attempts to Turing Test it by saying things like, ``I admit I'm not a human; now can we discuss unreachable cardinals?'') One might deny that we can ever build such a thing. Of course, there is little evidence that we can. But people who are fond of the other-minds problem are usually eager to place this topic on the table: the notion of an entity that would, if it existed, seem overwhelmingly intelligent, and yet {\it still} would not {\it really} ``have a mind.'' The reason they are so eager is because they have a vivid image of the possibility of such an entity existing without ``anybody being at home.'' That is, the robot mathematician could be a ``hollow shell,'' with no subjective experience. But the problems I raised in Section 1 now arise again with this sort of vivid image. There is also a new problem, which is that the whole experiment assumes we'll know a lot more in the future about how minds work, and then asks us to predict what it is we'll know. If we could actually build such a robot mathematician, we would presumably have a theory of mind that would make twentieth-century cognitive science seem puny. So what good is it to speculate about --- let alone set limits to --- the answers that future cognitive science would give us regarding the state of the robot's mind? Here's an analogy: Suppose someone in the year 1850 speculated that some day we might understand the chemical basis of life. Now suppose a vitalist proposed a Gedanken experiment in which some goo is created that behaves remarkably similar to an ameba. ``How would you know,'' he challenges, ``whether it was really alive?'' The proper response from the anti-vitalist is to say, ``Don't ask {\it me}; wait a hundred years and ask Watson and Crick.'' (And of course, Watson and Crick would say, ``Now that we understand what's going on, the question whether something is really alive has lost any interest or meaning.'') Unfortunately, a hundred years before the success of molecular biology, this answer doesn't sound very convincing. So we have to say a little more. If we return to the issues raised in Section 1, they appear here in a slightly different form, but are still relevant. First, consider the fact that mental phenomena form natural kinds. In the Gedanken experiment, that means that we have to consider whether the future cognitive scientists have succeeded in creating intelligence and paranoia in the robot mathematician, or creating other things that only {\it appear} to be intelligence and paranoia. It's odd that people often take it for granted that there could {\it be} pseudo-intelligence or pseudo-paranoia. Let's crank up the photosynthesis analogy again. Suppose someone built an artificial plant. How would we tell whether it was doing real photosynthesis or pseudo-photosynthesis? Well, chances are it would be doing something close to real photosynthesis, but it would also be interesting if there were entirely different mechanisms that looked macroscopically just like photosynthesis. On the other hand, it might turn out that there simply are no such mechanisms. When it comes to mind, we don't know enough yet to say whether it's possible for there to be pseudo-intelligence or pseudo-paranoia. That is, it might turn out that there's essentially just one way to create these things, or that there are a million quite different ways. Hence, if we ever build intelligent robots, it may be that they are doing things quite similar to what we do, or it may turn out that there are lots of design choices; {\it but these questions are entirely empirical}. When you go to the robot store, you may be able to select (e.g.), a companion that experiences jealousy, a companion that experiences something rather different from jealousy in an interesting way, or a companion that's quite good at faking jealousy for fantasy-game purposes. Now let's turn to the ``just going through the motions'' question. Here we come, I suppose, to the crux of the matter. Obviously, if someone believes that the robot might be ``just going through the motions,'' he's not alluding to the possibility that the robot is really remotely controlled by a person. (Cf. the old mechanical-chess-player hoax.) In that case, the debate would be just about the location of the mind, not its reality. No, the skeptic believes that the robot is actually really doing what it appears to be doing, but that ``doing'' is not sufficient. The robot's mechanisms might be executing exactly the computational steps that the future cognitive science says are those carried out by real brains, and yet the skeptic insists that it might still not really have a mind (or be conscious, or however you want to phrase it). ``In the end,'' says the skeptic, ``only the robot can be completely sure whether it is conscious.'' Okay, we'll ask it: ``Are you conscious?'' or, more specifically, ``Are you really afraid that Prof. Potter sneaked a peek at your proof, or are you just faking it?'' The skeptic will turn up his nose at {\it this}, obviously. All we're going to get out of this experiment is more observable data, both verbal behavior (e.g., the robot stutters as it denies it's paranoid) and observations of data structures and neuronal transmissions. The ultimate skeptic insists that all of these things are ultimately irrelevant, that subjective experience is fundamentally unobservable by all but the experiencer. Here I must part company with him. For one thing, as I said in Section 1, if it weren't for warped philosophical intuitions we would all grant that observations of mental implementational mechanisms {\it are} observations of subjective experience. I just don't know what we're talking about if we're not talking about an observable phenomenon. The skeptic claims simultaneously that his subjective experience can never be observed by me, and that when he casually alludes to it I know exactly what he is referring to. Surely the second part of this claim relies on an implicit belief that his experience and mine are ultimately mediated by the same mechanisms, and we could in principle open up our brains and verify that. If the skeptic denies this point, then I am afraid he is committed to a dualistic position, in which mental substances are connected in fairly arbitrary ways to physical objects. Suppose an entity does have a mind in this sense --- a subjectivity arbitrarily associated with it. Then there is nothing linking this kind of mind with any information-processing capacity of the system. Suppose the computers at the National Weather Service do have this kind of subjective minds, in the same sense that trees or rocks might. These minds might be dreaming about God; the chances that they are thinking about the weather are negligible. Somehow our theory of the mind associated with an entity has got to incorporate the idea that if sensors transduce information from the world into the entity, then what the mind of the entity knows about is the information so transduced. The idea that you could know all about the operation of the entity, and {\it still know nothing about its subjective experience}, carries with it the consequence that its subjective experience need have nothing to do with its operation. We would do well to steer clear of such a nightmarish form of the mind-body problem. \section{Conclusion} The other-minds problem is supposed to be the problem of deciding whether other people, or a particular entity, is possessed of a mind. However, the problem turns out to be a vestigial organ of an extinct philosophical organism, Cartesian epistemology. The question is invariably posed in a setting where it is completely obvious, prima facie, that the creature in question has a mind, and it presupposes that the alternative --- that it does not ``really'' have a mind --- makes sense. But the alternative cannot be spelled out without presupposing dualism or solipsism.
dave@cogsci.indiana.edu (David Chalmers) (10/20/90)
In article <26852@cs.yale.edu> mcdermott-drew@cs.yale.edu (Drew McDermott) writes: >Why There is No Other-Minds Problem >[...] >Okay, about this phrasing: As granted above, >mental phenomena have three aspects, subjective, behavioral, and >implementational. Why couldn't it simply be the case that the first >one is often or always absent? The only agent I'm sure has subjective >experience is me, so what's my evidence that anyone else has? Yes, this is the right way to state the problem. "Mind" as the term is traditionally used has behavioural, functional, and phenomenological aspects. The "other minds" problem is concerned only with the phenomenological (subjective) aspects. >The problem with this objection is that it misconstrues {\it aspects} >as {\it parts}. This construal would make sense for problems like the >``other-planets problem'': We observe one star with planets and ask >what our evidence is that some or any other stars have any. But >subjective experience is not an appendage of mentation in this way. A >mental event can be experienced from several different angles, >including (although the categories are fairly arbitrary), the >subjective, the behavioral, and the implementational. But the >experiences are all {\it of the same event}. The flaw in your argument is right here -- the rest is just wrapping. It may be that in many cases -- in particular, the cases where subjective phenomena exist -- the subjective, behavioural, and functional aspects are all aspects of the same event. But this does not imply that in all cases, they must be tied together. Certainly, in *my* case the three aspects go together. Does that allow be to deduce that in your case, or in a robot's case, they must also? Of course not. The coincidence of functional organization, behavioural consequences and subjectivity in some cases is quite compatible, a priori, with the coincidence of functional organization, behavioural consequences and non-subjectivity in others. Analogy time: In certain cases, temperature coincides with the motion of gas molecules -- they are literally two aspects of the same thing. In other cases, temperature exists without the motion of gas molecules. Of course, you may be wanting to make a *claim* that in all cases where at least one of them occur, the three aspects of "mind" will be tied together as different aspects of the same thing. This is a non-trivial claim, whose truth-value may or may not be "true". The non-triviality of this claim is the reason why there is an Other-Minds Problem. (I in fact believe that the claim is true -- more specifically, I believe that the phenomenological is supervenient on the functional. But I can't prove it, at least not easily, and so the Other Minds problem can still be raised.) >It's just extremely likely that agents capable >of experiencing the world would also experience some of their own >workings. Your phrasing right here concedes that the OMP is really a problem. "Extremely likely" is not good enough. It was always extremely likely that you and other people all have subjective experience. The OMP is "How can you know *for sure*?" As far as I'm concerned, the best answer for now is "You can't, but you can be pretty confident on various inductive grounds". >That is, the robot mathematician could be >a ``hollow shell,'' with no subjective experience. > >So what good is it to speculate about --- let alone set >limits to --- the answers that future cognitive science would give us >regarding the state of the robot's mind? Of course we can't predict these answer in advance. But few advocates of the OMP (Searle excepted) would argue in advance that we know that certain intelligent-seeming beings could *not* have minds. They just argue that the question is open. >When it comes to mind, we don't know enough yet to say whether it's >possible for there to be pseudo-intelligence or pseudo-paranoia. I hate to get stuck in a rut, but you've once again conceded that there *is* an OMP. >Okay, we'll ask it: ``Are you conscious?'' or, more specifically, >``Are you really afraid that Prof. Potter sneaked a peek at your >proof, or are you just faking it?'' I have a paper that addresses the relationship between claims about consciousness (e.g. "Sure, I have these really weird subjective feels"), and consciousness itself. It's a highly non-trivial issue, but the conclusion I come to is that if we want our claims about consciousness to reflect the properties of consciousness, then consciousness must be supervenient on functional organization (i.e., whenever you have the right functional -- that is, abstract causal -- organization, it must be accompanied by consciousness). If the source of consciousness is something other than functional organization -- e.g. if consciousness was biochemistry-specific, as Searle seems occasionally to believe -- then our consciousness-claims would be deeply irrelevant to consciousness itself. In this case, there'd be little point even talking about the Mind-Body Problem, as we wouldn't know what we were talking about. Shoemaker's paper "Functionalism and Qualia" (reprinted in Block, _Readings in the Philosophy of Psychology_, Vol. 1) touches on issues related to this. >If the skeptic denies this point, then I am afraid he is committed to >a dualistic position, in which mental substances are connected in >fairly arbitrary ways to physical objects. Suppose an entity does >have a mind in this sense --- a subjectivity arbitrarily associated >with it. Then there is nothing linking this kind of mind with any >information-processing capacity of the system. Suppose the computers >at the National Weather Service do have this kind of subjective minds, >in the same sense that trees or rocks might. These minds might be dreaming >about God; the chances that they are thinking about the weather are >negligible. One can be some form of dualist *without* believing that minds are arbitrarily associated with information-processing. My favourite theory of consciousness is sort-of-dualist, but still holds that the causal roots (really, the supervenience base) of consciousness lie in information-processing. There can be a dualistic mind-brain association without it being an arbitrary one. If we could *prove* such a theory, or any theory delineating the specific roots of consciousness in the physical, then there would no longer be an OMP. We would know precisely which entities have, or don't have, minds. Unfortunately no-one yet has done more than offer plausibility arguments for various theories. This is OK -- for all we know, that's the best we can do. But while that's the best we can do, the OMP will remain. >However, the problem turns out to be a vestigial organ of an extinct >philosophical organism, Cartesian epistemology. The question is >invariably posed in a setting where it is completely obvious, prima >facie, that the creature in question has a mind, and it presupposes >that the alternative --- that it does not ``really'' have a mind --- >makes sense. But the alternative cannot be spelled out without >presupposing dualism or solipsism. This is quite false. It's perfectly consistent to hold a quite materalist theory of mind, where subjective phenomena are identical to certain material events -- but only *certain* material events, such as ones that occur in a given biochemistry. I believe that for various reasons this is highly implausible, but it is a coherent position. -- Dave Chalmers (dave@cogsci.indiana.edu) Center for Research on Concepts and Cognition, Indiana University. "It is not the least charm of a theory that it is refutable."
mcdermott-drew@cs.yale.edu (Drew McDermott) (10/25/90)
As usual, my previous posting on Why There is no Other-Minds Problem needs further explanation. David Chalmers (dave@cogsci.indiana.edu) objects thus: In article <26852@cs.yale.edu> mcdermott-drew@cs.yale.edu (Drew McDermott) writes: >Okay, [how] about this phrasing: As granted above, >mental phenomena have three aspects, subjective, behavioral, and >implementational. Why couldn't it simply be the case that the first >one is often or always absent? The only agent I'm sure has subjective >experience is me, so what's my evidence that anyone else has? Yes, this is the right way to state the problem. "Mind" as the term is traditionally used has behavioural, functional, and phenomenological aspects. The "other minds" problem is concerned only with the phenomenological (subjective) aspects. >The problem with this objection is that it misconstrues {\it aspects} >as {\it parts}. This construal would make sense for problems like the >``other-planets problem'': We observe one star with planets and ask >what our evidence is that some or any other stars have any. But >subjective experience is not an appendage of mentation in this way. A >mental event can be experienced from several different angles, >including (although the categories are fairly arbitrary), the >subjective, the behavioral, and the implementational. But the >experiences are all {\it of the same event}. The flaw in your argument is right here -- the rest is just wrapping. Well, I'm not going to acknowledge that my argument is a flaw plus wrapping! I will acknowledge that we have a clash of several different intuitions, and that people are hard to budge from those they're used to. What I want to do is reassure those who think that there's something silly about the other-minds problem that their intuitions are basically healthy. It may be that in many cases -- in particular, the cases where subjective phenomena exist -- the subjective, behavioural, and functional aspects are all aspects of the same event. But this does not imply that in all cases, they must be tied together. Certainly, in *my* case the three aspects go together. Does that allow be to deduce that in your case, or in a robot's case, they must also? Of course not. The coincidence of functional organization, behavioural consequences and subjectivity in some cases is quite compatible, a priori, with the coincidence of functional organization, behavioural consequences and non-subjectivity in others. Analogy time: In certain cases, temperature coincides with the motion of gas molecules -- they are literally two aspects of the same thing. In other cases, temperature exists without the motion of gas molecules. Of course, you may be wanting to make a *claim* that in all cases where at least one of them occur, the three aspects of "mind" will be tied together as different aspects of the same thing. This is a non-trivial claim, whose truth-value may or may not be "true". The non-triviality of this claim is the reason why there is an Other-Minds Problem. (I in fact believe that the claim is true -- more specifically, I believe that the phenomenological is supervenient on the functional. But I can't prove it, at least not easily, and so the Other Minds problem can still be raised.) This restates the problem again pretty well, but fails to convince me it's real. It seems to propose that we could account for all the *observable properties* of subjectivity (or consciousness) and still not be sure we had "really" accounted for consciousness. This is what seems to me to be a preposterously high standard. Suppose we build a robot, and it claims to be subjectively aware. We ask it how it tells the difference between red things and green things, and it says they look different. When we press it, it starts to tell us about qualia. It treats its own decisions as free. Inspection of its blueprints shows that it has the same functional organization as the brain. (We can't do this today, of course.) It's at this point that we hit the intuition that we still couldn't know whether it *really* experienced anything. My claim is that this intuition is empty. When we've accounted for everything, there's nothing left to account for. If necessary, we could even put into the robot the intuition that mere observables are not enough to be sure an entity is subjectively aware. (I don't think it's necessary; I think the intuition is due to faulty education, not wiring. But I could be wrong.) Your phrasing right here concedes that the OMP is really a problem. "Extremely likely" is not good enough. It was always extremely likely that you and other people all have subjective experience. The OMP is "How can you know *for sure*?" As far as I'm concerned, the best answer for now is "You can't, but you can be pretty confident on various inductive grounds". What is the standard? Do we know the theory of evolution "for sure"? Most scientists are considerably annoyed by creationists' refusal to grant that the theory of evolution is as certain as anything ever gets in science. They should also be annoyed by philosophers' attempts to uphold a similar refusal here with respect to a hypothetical future theory of mind. Whence this refusal? I think Chalmers's use of the word "phenomenological" above is quite revealing. This word derives from Kant's notion of "phenomenon," or thing-as-it-appears. Kant and his buddies were obsessed by the distinction between appearance and reality. To our minds this obsession seems quaint. We picture the world as populated by a variety of information-processing systems, some conscious, others simpler. All of them introduce errors and approximations into data, and most manage to cope with these distortions most of the time. We can often come up with a quantitative theory about how close a system is to the truth about a situation. But *this* theory of appearance vs. reality is obviously not what Kant and Descartes were worried about. They were thinking of the situation of a mind that could be absolutely certain only of "appearance," and had to reason back from there to reality. If you take this picture seriously, then a mind isn't a mind unless things "appear" to it in this way. And nothing can "appear" to a robot in this sense, because (a) there's no absolute certainty; and (b) it's not even possible or necessary to single out one part of the robot as the "subject" of these "phenomena." *This,* I think, is what make people nervous about the idea that robots can have minds. All we have to do is junk the whole epistemological framework, and the problem goes away. Consciousness becomes just another part of the world, like photosynthesis. It's perfectly consistent to hold a quite materalist theory of mind, where subjective phenomena are identical to certain material events -- but only *certain* material events, such as ones that occur in a given biochemistry. I believe that for various reasons this is highly implausible, but it is a coherent position. Dave Chalmers (dave@cogsci.indiana.edu) Center for Research on Concepts and Cognition, Indiana University. By the way, I agree with this point. I am emphatically not arguing that a computationalist or functionalist position is correct a priori. It may well turn out (although I doubt it) that consciousness is a biochemical property, or a property of enormous and unintelligible neural nets. It just seems to me that *any* materialist theory is going to be open to the "objection" that there's no way to be *sure* that the objects it predicts are conscious *really* have minds, blah, blah. Drew McDermott mcdermott@cs.yale.edu