From agillies@U.Arizona.EDU Mon Feb 1 11:04:11 1999 X-Sender: agillies@pop.u.arizona.edu Date: Mon, 1 Feb 1999 11:49:16 -0700 To: Recipient.List.Suppressed:; From: "Anthony S. Gillies" Subject: discussion: worry about epistemic arguments Status: RO Hi all, I have a sort of ill-defined, amorphous worry about epistemic arguments against materialism. But, be warned, the worry seems to come and go as it pleases, and I may not have it by the time I get to the end of this message. In general epistemic arguments against materialism have the form: 1. Imagine someone (typically someone named 'Mary') who knows all there is to know about X (color vision, etc.) 2. But, imagine Mary has cognitive deficiency Y (color blind, raised in a black and white room, etc.) 3. One day she's released (or her color vision miraculously is restored, or ...) and sees *green* grass for the very first time. 4. She's learned something new; ergo materialism is false. Here's my worry (I think). (Incidentally, it probably rests on another amorphous, ill-defined worry about conceivability.) How can we be sure we're getting our inferences about Mary's inferences right? The argument needs it to be that Mary learns nothing new, but as far as I can tell this requires us to track Mary's inferencing. But this is hard for us since *we don't know* all there is to know about X. And how much someone knows about X has an obvious impact on the conclusions one can draw, no? Here's an example of what I have in mind. The Churchland's have a pretty cool parody of a thought experiment called (I think) the "Dark Room Argument". I'll change their example to make it an epistemic argument. 1. Imagine Mary knows all there is to know about light. 2. But Mary is in a completely dark room. 3. Suppose Mary is swinging a bar magnetic abover her head in the dark room. Nothing happens. Then someone turns on a light. 4. She's learned something new: light isn't electromagnetic wave propogation after all. Now, since we know more about light and how it worls, the argument is uncompelling. But we can imagine that it may have been compelling to (say) folks in the middle ages. Like I said, the worry comes and goes. But at the moment I am a bit worried about this sort of thing. Thony +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ "Curious green ideas sleep furiously." +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ From josh@ame2.math.arizona.edu Mon Feb 1 12:23:06 1999 From: Josh Cowley Subject: epist. and analysis arguments To: bradt@U.Arizona.EDU (Brad Thompson) Date: Mon, 1 Feb 1999 13:22:21 -0700 (MST) Cc: chalmers@ling.ucsc.edu, aburnett@U.Arizona.EDU, agillies@U.Arizona.EDU, akolers@U.Arizona.EDU, atlane@U.Arizona.EDU, bayne@U.Arizona.EDU, cowley@U.Arizona.EDU, erikh@U.Arizona.EDU, erikl@U.Arizona.EDU, jtismael@U.Arizona.EDU, kcreath@primenet.com, laj@U.Arizona.EDU, patrickr@U.Arizona.EDU, rachaelp@U.Arizona.EDU, sch@U.Arizona.EDU, shaughan@ns.arizona.edu, tolliver@U.Arizona.EDU, bradt@U.Arizona.EDU Status: RO I have some concerns about the epistemic and analysis arguments. In particular it seems that the two epistemic arguments and part of the analysis argument ultimately come back to conceivability arguments. If one is still hesitent about conceivability arguments, then it isn't so clear that the other arguments will take its place. I'll take each in turn. 1) The Argument from Epistemic Asymmetry The argument is that my knowledge that conscious experience exists derives primarily from my own case. External evidence doesn't play a role in it. (pg 102). This is unlike how we know about other things. From this we can conclude that consciousness can't logically supervene because, "If it were logically supervenient, there would be no such epistemic asymmetry; a logically supervenient property can be detected straightforwardly on the basis of external evidence." (pg 102) My probelm is with the word "straightforwardly." On one interpretation electrons cannot be detected straightforwardly from external evidence. In order to detect electons you first have to postulate their existence and build a partial theory describing some of their observable properties. Then you have to build some elaborate tools to see one. Why shouldn't I think that we just don't have a partial theory are the right tools to detect consciousness from external evidence. Most of the rest of the mind is an unsolved problem, why not this one? It seems that my only reason for thinking this can't be done is the conceivability arguments. 2) The Knowledge Argument: Mary knows everything about the physics/functionality of the mind, but there is something she still doesn't know until she actually sees red. I have no idea what it would be like to have a complete theory of the physics and functionality of the mind. So my intuitions about Mary are not comming from an understanding that *the* theory of mind doesn't explain consciousness. Rather, it is that I cannot immagine a completed physical/functional theory of mind explaining consciousness. Again it comes to a conceivability argument. 3) From the absence of analysis. In order for a reductive explanation to work there needs to be an analysis of consciousness whose satisfaction physical facts could imply. But there is no such analysis. Functionalism gives the best hope, but then to explain consciousness is just to explain our ability to manifest some capacity. "But on the face of it, it is entirely conceivable that one could explain all these things without explaining the experience that accompanies the report or the discrimination." (pg 105) The first part of the absence of analysis argument pretty clearly rests on a conceivability argument. The remainder of the argument is quite interesting and I don't think it is subject to the line I'm running here. Even if you don't agree with what I've said, there is one point I've used in this argument that I think is important. The mind (even without consciousnes) is possibly the most complex thing human's have ever studied. Vision is probably the most well developed area in the theory of mind and there are still many ongoing debates and just plain holes in that theory. When we imagine a completed functional or physical theory of the mind, I would suggest that we are not even comming close to immagining what such a theory will actually look like. That isn't to say that such imaginings are pointless, but we need to keep the limitations of our imagination in mind. From chalmers Tue Feb 2 03:36:27 1999 Date: Tue, 2 Feb 1999 03:36:24 -0800 (PST) From: David Chalmers To: chalmers Subject: Epistemic arguments Status: RO Thanks to Thony and Josh for their messages. I'm looking forward to more comments and to others' reactions, but in the meantime here are some thoughts (I hope they'll be helpful in the discussion Tuesday). I'll begin with some background and stage-setting, putting the discussion here within the context of the more general metaphysical issues in the seminar. Modal arguments against materialism, and against other ontological doctrines, take off from the central point that ontological views have modal consequences. For example, if materialism is true, so that in some sense the fundamental constituents of the world are physical, then the physical facts must necessitate all the facts. So one can argue against materialism by arguing that the physical facts don't necessitate all the facts. One can similarly argue for and against other ontological views, by considering the question of whether facts in the (putative) reductive domain (call it the A domain) necessitate facts in the to-be-reduced domain (the B domain). How does one argue for such a modal claim? This comes down to some deep issues about the epistemology of modality, issues which we'll be concentrating on in this seminar. But the most common way to argue for such a claim is to make an *epistemic* argument. Such an argument tries to establish that there is an epistemic gap between the A domain and the B domain. The thrust of such an argument is usually that there is no a priori entailment from A truths to B truths. From here, it is inferred that there is no necessary entailment from A truths to B truths, so the ontological reduction of B to A can't succeed. Perhaps the central questions we'll be concentrating on in this seminar is whether the move from an epistemic gap to an ontological gap (from failure of a priori entailment to failure of necessary entailment, for example) is valid, and if so, when and why. But in this section, we're looking at what's involved in the antecedent of such a move -- i.e. the argument for an epistemic gap, which will always be the first step in such an argument. There are a number of ways such arguments can be run. One can use epistemological arguments, arguing directly that knowledge of A truths isn't sufficient to yield knowledge of B truths (even when combined with arbitrary a priori reflection). One can use conceivability arguments, arguing that there is no incoherence in the suppositions that the A truths hold but certain B truths do not. And one can use arguments from conceptual analysis, arguing that B concepts are not of the right type to support a priori entailment from A to B. The case of consciousness provides some paradigm examples. The question at issue is whether there is an epistemic gap between physical and phenomenal, and in particular whether there is an a priori entailment from physical to phenomenal. Here arguments from conceivability, epistemology, and conceptual analysis can all be invoked. All of these arguments involve an appeal to some sort of epistemic or conceptual intuition at some point (as do most modal arguments, and indeed most arguments anywhere in philosophy). I'm not going to try too hard to convince you that there is indeed an epistemic gap in these cases, as we'll mostly be interested in the independent question of whether the epistemic-to-ontic move is valid (a question that remains of general interest for all sorts of issues even if one rejects the epistemic gap in this particular case). But it's good to get a sense of what's involved in supporting such a claim. Thony and Josh both raise the question: how can we know that there is an epistemic gap between A and B when we don't yet know all the A facts. An anti-materialist will surely concede that we don't yet know all the physical facts, so presumably he or she will have to be making a general claim about what can follow from physical facts, a claim independent of the specifics. I think Jackson intends to make it plausible, for example, that no matter what physical facts Mary knows (even future physical facts), that can't be enough to teach her what red experiences are like. I take it there is at least a strong intuition here (i.e., to the effect that Mary's epistemic gap isn't merely due to her not having the right specific physical knowledge). And I think an opponent would be fighting an uphill battle if they were to argue that with the right specific physical knowledge, Mary really could know what red experiences (and the rest) feel like from within her room. So, we seem to have an intuition of a general epistemic gap here, not just a specific epistemic gap. But it is interesting to think about what might be grounding such an intuition. One possible general way of grounding the Mary intuition would be: (1) Physical knowledge is all descriptive knowledge; (2) No amount of descriptive knowledge could on its own enable Mary to know what it's like to see red; so (3) No amount of physical knowledge could enable Mary to know what it's like to see red. Of course there are questions about just what "descriptive knowledge" comes to here. One might try to cash it out as "knowledge acquirable in a black-and-white room", or "knowledge acquirable without relevant experiences", or "knowledge about the objective structure of causal systems". For any such reading one could then of course at least argue about whether premise (2) is compelling. My own view is that the general Mary claim has a force which doesn't need to be grounded in such an argument, but such arguments can certainly be used to lend force (and often are, by proponents of epistemic arguments). Something similar goes for conceivability arguments. When it's claimed that zombies are conceivable, for example, an opponent might respond be saying that we don't know all the relevant physical facts, and therefore don't know just what to conceive of when conceiving of a zombie. In response, the conceivability-arguer might make a general claim such as: for any physical P, it's conceivable that P could be true in the absence of experience. And it might be claimed that this has some strong intuitive plausibility. It seems to many, for example, that the conceivability of zombie scenarios is more or less independent of the specific physical facts in question, and that varying those facts in the conception makes no difference at all to the presence or absence of experience in the conception. I have some sympathy for this myself, but it's true that such a point is more compelling if accompanied by some sort of general argument, or some sort of analysis of the underlying relation between physical and phenomenal concepts. For my part, I'd support the claim roughly as follows: (1) Physical concepts are all structural-dispositional concepts; (2) If B truths are to be entailed a priori by structural-dispositional truths, there must be some analysis of B concepts in structural-dispositional terms; (3) There is no analysis of phenomenal concepts in structural-dispositional terms; so (4) Phenomenal truths are not entailed a priori by physical truths. In a way, this corresponds to the point stressed in the book that the argument from the absence of analysis is what lies at the root of the epistemic arguments; and premise 2 in particular corresponds to the point that "structure and function only ever adds up to more structure and function". Of course, an opponent could question premise 3. (They might also question premises 1 or 2, but I think this would probably not be fruitful in the end.) I take it that a dispositional (functional) analysis would be the only candidate of any promise, and that ultimately there are only a few promising dispositions in the vicinity. Here, though, note that the epistemic arguer can mount a case against such analyses without needing to worry about the precise nature of specific physical facts. The claim will just be that for any disposition in the vicinity, there is an epistemic gap between that disposition and experience. For example, for any such disposition, it's conceivable that it be present without consciousness; or one could know all the facts about such dispositions without knowing all the facts about consciousness. Someone might still question this, but at least we have narrowed down the scope of the discussion from worries about future physical facts. All this bears directly on Josh's question about priority between arguments from conceivability, epistemology, and analysis. To my mind, the question of priority is subtle. I think that each of the claims have a certain prima facie force that needn't be grounded in the others, but that underneath there is something of a relation of mutual support going on. It would be surprising to me if the conceivability argument were seen to ultimately carry the burden, as there are a number of people who have the Mary intuition and the analysis intuition while being unsure about or even denying the full conceivability intuition. But no doubt there is something of a circle of mutual support here, and it is interesting to uncover its structure. If I were to try to articulate the substratum underneath the intuitions, I guess I would hold as above that the full conceivability claim is grounded in the deeper point about absence of the relevant sort of analysis, i.e. in a claim about the nature of phenomenal concepts, and it may be that the epistemological arguments are too (though I think that both of these arguments can have prima facie force to one who is uncommitted on the abstract underlying point). It's arguable in turn that the point about analysis is itself grounded to some extent in intuitions about conceivability and epistemology, but note that these are now much more limited claims than the full-blown claims (concerning arbitrary physical facts) which we started with. I also think, though, that the argument from absence of analysis has independent support, e.g. in the sort of reasoning which holds that "for any function, we can always raise the question, why is the performance of this function accompanied by experience, and this will always be a nontrivial further question?". This sort of further-question reasoning (which I rely on most in my paper "Facing Up the the Problem of Consciousness") can arguably be supported by conceivability and epistemological considerations, but I'd argue that it has intuitive support which is more or less independent of and prior to such considerations. So if I were to rest on any point as supporting the central burden, it might be on points like this concerning the conceptual distance (at least in a priori space) between functional and phenomenal concepts. But I think it is most accurate to see things in terms of a complex web of support among a number of related considerations. A few specific points re Thony and Josh. Re the "Dark Room": >1. Imagine Mary knows all there is to know about light. >2. But Mary is in a completely dark room. >3. Suppose Mary is swinging a bar magnetic abover her head in the dark >room. Nothing happens. Then someone turns on a light. >4. She's learned something new: light isn't electromagnetic wave >propogation after all. I presume that premise 1 is meant to say that Mary knows everything *physical* about light (in order to engage appropriately with 4, and to provide an appropriate analogy). Here I'd want to know in what intuitive sense "Mary has learned something new". It seems to me that intuitively, the only new thing Mary has learned is what it is like to see light; i.e. some facts about light *experience*. From here, one might well infer that facts about light experience are not reducible to facts about EM wave propagation. But it's hard to see a more general argument about non-experiential aspects of light getting off the ground; so it's hard to see that we have any analogous plausible but fallacious argument in the vicinity. At least, I'd need to see a lot more spelling out of just what is supposed to be going on in this scenario, and what is supposed to be learned, from the point of view even of the 18th century scientist. Re: the epistemic asymmetry argument, and "a logically supervenient property can be detected straightforwardly on the basis of external evidence." This was probably too quick on my part. Properties such has "had meat for breakfast in 10,000 BC" are presumably logically supervenient but not straightforwardly detectable. On the other hand, *given* relevant physical and microphysical knowledge, such properties are fairly straightforwardly detectable. I guess something similar would apply to electrons. They're detectable only nonstraightforwardly from external evidence, but determinable pretty straightforwardly from complete microphysical knowledge. Whereas even complete microphysical knowledge leaves a gap in the consciousness case. Of course an opponent might deny the epistemic intuition here, and say that with the right physical knowledge, one's knowledge of consciousness would be fairly straightforwardly derivable. I agree that this epistemic asymmetry argument is pretty close to both the conceivability and direct epistemological arguments, which is why I don't rely on it much. Re the knowledge argument: I think I'd want to deny that this is a conceivability argument. The argument isn't that one can't conceive of Mary knowing what it's like to see red, it's that we have good reason to believe that she won't know what it's like to see red. See above. Re the analysis argument: I've tried to spell out the extent to which this does and doesn't rely on conceivability considerations. Re the brain being complex: Of course this is so. But still, a physical theory of the brain's functioning has a certain shape, irrespective of the details, and it may be that there are things we can know about what theories of such a shape can and can't explain, by virtue of certain very general principles that are independent of the details. (Think about someone arguing that no theory of pure statics [without a temporal element] could explain dynamics, for example.) E.g., if it were the case that (1) a physical theory can only explain structure and function, and (2) to explain structure and function is not to explain experience, then one could draw conclusions about the limitations of physical theory in explaining experience. Of course work is being done by general premises such as (1) and (2), which can be questioned, but the point is that we can do a good amount of work in assessing such premises even in advance of knowing the specific details of physical theories of the brain. I certainly don't want to say that empirical knowledge of the brain is completely irrelevant in these matters; and it's certainly vital for understanding all sorts of philosophically important things about the mind. But there is a lot of philosophy we can do in advance of understanding all the details. One had better be sure that the philosophy one does is compatible with any details that emerge, of course. My own view is that attention to the details of neurophysiology and other empirical areas tend to give the epistemic considerations in question fairly strong support. One might reasonably disagree about that, or hold that some future neurophysiology will show us something pretty different, but I suspect that for such a view to have much weight, it's going to have to be supported by a good amount of relatively a priori argument on the conceptual and epistemic matters in the vicinity. These are all deep issues that deserve attention, and I'd be interested to hear any thoughts. --Dave. From bradt@u.arizona.edu Wed Feb 3 01:29:52 1999 Date: Wed, 03 Feb 1999 02:23:55 -0700 From: bradt X-Accept-Language: en To: bradt@u.arizona.edu Subject: minutes from seminar meeting Status: RO Hi all: Here is a summary of our discussion Tuesday--feel free to add anything that I may have left out. First, questions about the unidirectionality of the supervenience relation were addressed, relating to some of our earlier online discussion. We also briefly touched on the definition of materialism and how it is that the logical possibility of zombies would entail the falsity of materialism. This lead us into a discussion of the relationship between conceivability and possibility. It was pointed out that the relevant notion of conceivability has to be stated carefully, so as to rule-out the conceivability of Goldbach's conjecture being false if in fact it is true, for example. Some other remarks were made to the effect that the close relationship between conceivability and possibility is (perhaps pragmatically?) desirable, since any gap between conceivability and possibility might lead to mysteriously inaccessible possibilities or impossibilities. Everyone seemed to be clear about supervenience, possibililty, and necessity. We spent most of our time discussing epistemic arguments against materialism, especially the case of what Mary didn't know. Thony was asked to restate the issues he had raised in online discussion. We then debated some of the points raised there. In particular, we discussed whether or not epistemic arguments were weakened by the fact that the physical facts which are claimed to be insufficient for knowledge of phenomenal facts are left unspecified. Some claimed that this did make epistemic arguments weak. Others were essentially in agreement with Dave that it was sufficient to notice that the relevant physical facts would merely provide more information about structure and dynamics, and that such information would always fall short of providing knowledge of the phenomenal facts. We also ended up discussing subjectivity and objectivity, as well as concepts (and whether or not they had both subjective and objective aspects to them). Somehow (!) this also lead us into discussing Wittgenstein on private language and on the beetle in the box, and Dennett with regard to purported differences that "don't make a difference" (such as Orwellian vs. Stalinesque revisions). Brad From owner-modality@LISTSERV.ARIZONA.EDU Wed Feb 3 06:02:08 1999 Date: Wed, 3 Feb 1999 06:02:04 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Perplexity To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Thanks to Brad for the minutes. I would be interested to see more on these points. I'm a little perplexed that so little is being posted to the mailing list. Only a couple of messages in the last 6 or 7 days, even though I asked that everybody post something about the current round of readings by now. If this goes on, we are going to fall a long way behind. Is there a problem? If so, please let me know (private e-mail is fine). Any suggestions as to methodology are welcome. As things stand, I think I have to be more explicit. There is a weekly writing assignment, which involves posting at least a page or so of questions and/or comments on the current round of readings (and feel free to post more). Students are also expected to engage in a good amount of mutual discussion of each others' contributions, in addition to this (N.B. "in addition", i.e., the initial writing assignment shouldn't be primarily reactive). These writing assignments are a core part of the seminar. I know the online seminar is an unusual arrangement and is not ideal, and I hope we'll be meeting as a normal seminar soon (no news on the visa yet, alas). In the meantime, though, I think some effort at participation on everyone's part can make this a fruitful and productive arrangement. Again, any suggestions are welcome. I'll look forward to seeing everyone's questions and comments on the second set of readings in the next day or so, followed by a good round of discussion. You should also be reading over _Naming and Necessity_ for the third round; your thoughts on that will be due early next week. Although we won't be focusing on the 2-D material until the fourth round, I suggest that those of you who have read N&N already read over the 2-D material first (esp. in TCM and The Components of Content), and then read over N&N with that material in mind, trying to understand it in those terms. I'll give advance notice that at some point fairly soon there will be an assignment asking you to translate all or most of the central examples in N&N into the 2-D framework. That should help you get a working understand of what the 2-D framework is all about and of how to use it, which will be important for the rest of the seminar. --Dave. P.S. We're now ready to switch to the listserv, so all future contributions should go to modality@listserv.arizona.edu. From owner-modality@LISTSERV.ARIZONA.EDU Wed Feb 3 09:33:42 1999 Date: Wed, 3 Feb 1999 10:33:00 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: On Mary and Experience and Concepts To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO All, In reply to Dave's request for more info on Tuesday's discussion, let me attempt to elaborate the line that Thony and I were attempting to develop. (Though don't hold Thony accountable for what follows!) I don't think that the argument would convince anyone not already convinced that Mary can know what it's like to see red, but perhaps it focuses the argument on the issues on which it should be focused: the nature of *concepts* about the phenomenal. (1) Assume that Mary is not only an expert neuroscientist, but an expert developmental psychologist as well. She knows how infants develop all the concepts that they do on the basis of their innate mental structure and the sensory input that they receive. Not a contentious assumption, I hold. (2) Assume that infants develop all concepts on the basis of phenomenal states and their innately given structure. This is controversial. The idea is that sensory input leads to phenomenal states in infants, and they form their concepts on the basis of those phenomenal states (probably together with the feedback they get from action - which we could regard as simply more sensory input). Note: I don't claim that this is the only way in which one could develop concepts. Surely you could (logically) build a machine that develops concepts on the basis of input without going via the sensory states that that input generates, but it does not *seem* as though this is how we do it. (3) Mary has the innate mental structure that infants do. (4) Mary has the concepts that, say, the normal adult human being has. In particular, she has the concept red. After all, she knows everything that there is to know about colors, and to do this she must have all the color concepts. (5) So, Mary will be able to substract from her concept of red the innate structure that went into it, and come up with the phenomenal experience on the basis of which it was constructed. In other words, she can simply reverse-engineer her concept of red, and do in reverse what the infant does in forward. Thus, Mary-the-developmentalist does know what it's like to see red. Here's the predicted reply from Jackson and friends: (4) is false. Mary does not have the concept of red. Or, perhaps more accurately, she does not have the concept of red that is important for this discussion. She may have *a* concept of red in the sense that a blind person can have a concept of red (its a color, people see it, tomatoes are red, as are fire engines, etc.) but she does not have the phenomenal concept of red, of which red qualia is a *constituent.* You cannot have the phenomenal concept of red without qualia because the red quale is a (proper?) part of it. Knowing-that can include phenomenal states, because it includes phenomenal concepts, and phenomenal states are part of (some) phenomenal concepts. And the only way you can generate these concept is by having certain selective types of input from the environment. Mary is in her dark room (or is color blind), and thus cannot get (or operate on) this type of input. (And she cannot provide this input endogenously either - she cannot visualize in red.) Hence she could never concept of red from which to sift out or distill the phenomenal experience of red. As I say, this is a fair reply on the part of Jackson. (And actually, there are other problems with the argument too.) It seems to me that the essential issue is now the nature of phenomenal concepts. If we accept the claim that qualie are constituenta of phenomenal concepts, then we can say that there are two ways one can come to know a phenomenal property: via the experience itself, or via the phenomenal concept. Indeed, we might even be so bold as to suggest that qualia simply *are* phenomenal concepts. Perhaps qualia can have inferential roles, enter into reasons, and do some of the other nifty things that concepts do. Actually, it occurs to me that this is Lycan's line on qualia (I think): they are concepts that are only intra-subjectively potent; they are, so to speak, words in the experiencers langauge of experience, that can't make the leap into public discourse. (But perhaps I have Lycan wrong.) What got me thinking along these lines was the attempt to bridge the gap between the objective (concepts - available for intersubjective discourse, the box a la Wittgenstein) and the subjective (qualia - not available for intersubjective discourse - the beetle). Now: if infants can build up from the subjective to the objective, then why can't Mary go in the other direction. Of course, Fodor's reply is that infants don't build up from the subjective to the objective: all concepts are innate. But as long as you're with Auntie on this one and think that Fodor's wrong, maybe there is something to the thought that Mary can do in reverse what we - as infants - do foward. Tim Timothy J. Bayne RM. 213 Social Science Department of Philosophy University of Arizona Tucson, AZ 85721 USA Hm ph. (520) 298 1930 From owner-modality@LISTSERV.ARIZONA.EDU Wed Feb 3 10:27:49 1999 x-sender: agillies@pop.u.arizona.edu Date: Wed, 3 Feb 1999 11:31:02 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony S Gillies Subject: On Mary and Experience and Concepts To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO >Here's the predicted reply from Jackson and friends: (4) is false. Mary >does not have the concept of red. Or, perhaps more accurately, she does >not have the concept of red that is important for this discussion. She may >have *a* concept of red in the sense that a blind person can have a >concept of red (its a color, people see it, tomatoes are red, as are fire >engines, etc.) but she does not have the phenomenal concept of red, of >which red qualia is a *constituent.* You cannot have the phenomenal >concept of red without qualia because the red quale is a (proper?) part of >it. Knowing-that can include phenomenal states, because it includes >phenomenal concepts, and phenomenal states are part of (some) phenomenal >concepts. And the only way you can generate these concept is by having >certain selective types of input from the environment. Mary is in her dark >room (or is color blind), and thus cannot get (or operate on) this type >of input. (And she cannot provide this input endogenously either - she >cannot visualize in red.) Hence she could never concept of red from which >to sift out or distill the phenomenal experience of red. I think Tim is right in predicting this reply. One worry: if we say (in the premises) that Mary doesn't have the full concept of red but just the non-phenomenal component, then aren't we begging the question against the materialist? After all, the *conclusion* is supposed to be something along those lines. On the other hand, the materialist's use of premise (4) (i.e., that Mary has the full concept of red) seems to beg things the other way; so this might be an indication not that this Tim's argument is an argument *for* materialism, but rather an argument that the Mary-the-neuroscientist-case isn't a good argument *against* materialism. Thony "Curious green ideas sleep furiously." From owner-modality@LISTSERV.ARIZONA.EDU Wed Feb 3 10:43:08 1999 Date: Wed, 3 Feb 1999 11:32:26 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik J Larson Subject: Re: Perplexity To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I have a question regarding Thony's example of the government or "body-politic" (I can't remember the precise phrase now). Did you mean that, even in principle, there are no illuminating accounts of a political process in terms of basic physics. If you meant in principle, I guess I'm confused. In practice, explaining a high-level phenomenon like a political process might be very cumbersome using physics, but it seems like, if that political organization you spoke of is entirely physical, there is nothing precluding such an explanation in principle. "not very illuminating" is perplexing in it's own right. What makes something "illuminating"? That's a strange notion once you've conceded that the process referred to is nothing over and above physics. I guess what I'm asking is, what is the role--ontologically, functionally, explanatorily or whatever--of concepts in explanations and why should they be NECESSARY if a supervenience relation holds between the phenomenon and a set of entirely physical facts? For that matter, WHAT IS A CONCEPT in this sense that we keep using it? Erik L. From owner-modality@LISTSERV.ARIZONA.EDU Wed Feb 3 12:24:02 1999 Date: Wed, 3 Feb 1999 13:27:42 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Conceivability and Knowledge arguments To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Dave suggests that conceivability arguments are distinct from knowledge arguments. The Mary argument is a knowledge argument: Mary knows all the physical facts (A facts) but she doesn't know all the facts about experience (B facts), therefore . . .. Zombie arguments are conceivability arguments: I can conceive of my zombie, i.e. of a being with all my physical properties but none of my phenomenal properties, therefore . . . I'm not so sure that these arguments are really distinct at any deep level of analysis (although there's a caveat below). First, note that one has to conceive of Mary and locate her in the distant future of neurosience. What does this involve? Well, it's hard to say. It might involve visualizing a certain scene (say, Mary wearing a white lab coat), and labelling it *Mary the expert neuroscientist who knows everything about the brain*. What does conceiving of my zombie involve? Well, perhaps something similar. I have a picture of someone just like me, but there's nothing it's like to be him - nobody's home and the lights aren't on - and I stick a label on him: *My zombie twin.* But in fact, I doubt that we do anything *like* this in either case. Conceivability does *not* seem to involve the imagination - at least that's what it seems like to me in *these* cases. It's *not* like trying to work out in your head whether you can fit a ball of diameter two inches inside a triangle of sides 3, 4 and 4 inches which has an outer perimeter of 1/2 an inch thick. So what is going on in these cases? Well, it seems to me that in both cases you have the following intuitions or you don't. (a) Necessarily, all physical properties are functional or structural properties, and phenomenal properties aren't functional/ structural properties; OR, the very similar intuition: (b) Nec., all physical facts (or, perhaps, propositions) are exclusively composed out of functional /structural concepts, and phenomenal concepts are not structural/functional concepts. (a) is at work in the zombie cases, (b) is at work in the knowledge cases. If there is a one-one corresondence between facts and properties as is commonly supposed in setting these cases up, then it is hard to see why one should have (a) and not (b), or vice-versa. (Caveat: Of course, some do reject the claim that if you've got two facts you've also got two properties, but then you can accept that Mary does learn something new without rejecting physicalism.) But imagination does seem to play some role in conceivability. Consider the following problems: (1) is logically it possible to be an agent without, being a perceiver (or vice-versa)? (2) Is it possible for one subject of experience to have two distinct (spatially separated) bodies at the same time, so that, e.g., one of my bodies might be in Tucson, and the other one in Texas at the same time, and they are simultaneously conscious? Well, people seem to have very different intuitions about both of these cases? Why? Are they: (1) actually imagining different scenarios? or (2) imagining the same sorts of scenarios, but putting different labels on them because of the differing structure of their concepts? I dunno, but I'm tempted towards (2). Perhaps conceivability and knowledge arguments are different types of arguments because they are different ways of pumping the same intuitions (i.e. the same concepts), but I'm not even sure that this is true. Tim Timothy J. Bayne RM. 213 Social Science Department of Philosophy University of Arizona Tucson, AZ 85721 USA Hm ph. (520) 298 1930 From owner-modality@LISTSERV.ARIZONA.EDU Thu Feb 4 03:31:18 1999 Date: Thu, 4 Feb 1999 03:31:14 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: On Mary and Experience and Concepts To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Tim's and Thony's line of argument is interesting. Speaking as someone sympathetic with Jackson, I'd be tempted to endorse the predicted reply to Tim's argument, contesting (4). Actually, I'd be tempted by a somewhat fuller two-part reply. (1) There are certain concepts of "red" that Mary has in the black-and-white room, but these are not concepts that require the experience of red for their possession. They might be concepts analyzable as "experiences of the sort caused is normal observers by certain paradigm objects (the objects which normal observers label 'red')" (for her concept "red experience") or "objects of the sort that produce red experiences (i.e. the same sort of experiences as are produced by the paradigm objects) in normal observers under normal conditions" (for her concept "red object"). It's plausible that the "red" concepts that Mary possesses have this sort of nature, but I think it's clear that neither of these concepts requires Mary to have ever had a red experience, and there's no reason to believe that possession of either concept will allow her to know what a red experience is like. (2) There are other concepts of "red" that Mary doesn't have in the room, but acquires later on having her first red experience. Such a concept might include the concept R corresponding to her new belief, tomatoes look R. (Of course she already knew they look red, in the above senses.) Mary doesn't have this concept in the room, so she won't be able to use it to figure out what red experiences are like. I think that (1) alone is a sufficient reply to the argument, so the anti-materialist doesn't need to put special weight on (2). But I think that (2) is plausible all the same. I think it's more or less common sense to say that a blind person can have only an "indirect" concept of red experiences, as opposed to the "direct" concept that is possessed by someone who has actually had the experiences. Is this question-begging? I hope not. (1) seems more or less like common sense to me, and (2) isn't really going out on a limb. It seems to me that both of these are claims that even many materialists might endorse. One needn't even rely on claims about what Mary could and couldn't figure out in principle; one can just think about the ordinary concepts possessed by a non-omniscient colorblind Mary, and the concepts possessed by an ordinary person who can see colors. (Incidentally, "begging the question" is a very interesting phenomenon in its own right. Almost any philosophical argument can be accused of begging the question. Say it is a valid argument from premise P to conclusion not-C, arguing against a C-ist. Then it's obviously equally the case that if C is true, then P is false. So clearly on the C-ist position, not-P. So a C-ist can accuse the original arguer of begging the question against them by asserting P. Obviously this can't be quite right. I think the most important distinction between question-begging and non-question-begging arguments is whether the premises have plausible independent support antecedently to considering the conclusion.) In any case Tim is right that these issues about phenomenal concepts are very important. I try to develop an account of phenomenal concepts of the sort discussed in (2), on which the concept R is partly constituted by the red experiential quality itself, in my "The Content and Epistemology of Phenomenal Belief" (online as the third Princeton talk). --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Thu Feb 4 23:29:53 1999 X-Accept-Language: en Date: Fri, 5 Feb 1999 00:12:29 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Brad Thompson Subject: Our ignorance of "all the physical facts" To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO It seems to me that the most defensible materialist response to the anti materialist arguments we've been discussing and reading about is the one that the Churchlands and Dennett have given (I think all 3 philosophers have given this basic response to Jackson's argument, Nagel's argument, and the conceivability of zombies). The response is that we don't know "all the physical facts", so that what we *succeed* in imagining when we try to imagine Mary's situation (or similarly being a bat, or a physical duplicate of me without consciousness) is not what the thought experiment *asks* us to imagine. The materialist assumption that goes with this reply is that *if* we did know all the relevant physical facts, zombies would be inconceivable (and we'd know what it is like to be a bat, and we'd be able to imagine being in Mary's situation and wouldn't conclude that she didn't know what seeing red was like). Dave's response to this objection appears to be that even though we do not know the specific physical facts needed, we know what *types* of facts that additional knowledge would be--more structure and dynamics. And adding more structure and dynamics would still not give us a reductive explanation. Look at functionalist theories of qualia. They radically fail to give an account of why a red experience has the particular quale that it does, or any quale at all. The inverted spectrum argument shows this, but the shortcomings of functionalist accounts doesn't really require argument--they simply fail to provide any explanation for why there are qualia or for why mental states have the particular qualia that they do. But in the case of functionalism, I think it is pretty easy to see that adding extra functional information could not possibly remedy the situation. But what about some yet unknown neurophysiological information--could this close the epistemic gap (is this what Mary knows, that we don't)? Well, it is difficult to imagine how such a view could resist positing what Dennett would call "wonder tissue". What non-functional property does the brain possess which makes it conscious? Though I think that the above considerations help against the materialist response I've been discussing, I think we have to concede that the anti materialist arguments are weaker than they would be if we did in fact have total physical knowledge. The materialist might insist that this concession entails that the anti materialist arguments simply fail. But it seems to me that they should not be too comfortable in resting their own view on a check which, prima facie, they probably cannot cash. One can view the anti materialist thought experiments as providing the prima facie evidence that that check won't be cashed. Perhaps we are left then with the question of where the burden of proof lies--on the materialist or on those who deny materialism. PS-- On another note, there is a wonderful example of what Dave was talking about with regard to begging the question. In _Philosophical Naturalism_, David Papineau gives an argument for materialism which has the falsity of epiphenomenalism as one of its premises. Jackson, on the other hand, argued against materialism and then uses that to argue for epiphenomenalism. From owner-modality@LISTSERV.ARIZONA.EDU Fri Feb 5 10:12:42 1999 Date: Fri, 5 Feb 1999 10:59:06 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Josh Cowley Subject: Re: Our ignorance of "all the physical facts" To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO In response to Brad. > Dave's response to this objection appears to be that even though we do > not know the specific physical facts needed, we know what *types* of > facts that additional knowledge would be--more structure and dynamics. > And adding more structure and dynamics would still not give us a > reductive explanation. What is it that `structural' means here? Does it mean physical structure or is it more general? If it is more general, them I'm not sure why consciousness isn't structured. One note on the materialist insufficient-info rebuttal. We have been discussing the rebuttal as if it has the following form. 1. We don't know all the physical facts. 2. You can't have a good conception of Mary (zombies etc.) unless you know all the physical facts. C. Therefore the arguments fail. I think an important step is being left out of the real argument though. 3. Everything other than consciousness seems to be explainable in terms of physics 4. It would be weird if there were only a single phenomina that couldn't be explained by physics. C. Therefore I have better reason for thinking I'm missing something than for thinking physics can't explain qualia. You can, of course, argue whether you actually have better reason for this. But its important to see that the materialist is comparing possibilities and judging one to be more plausable. Josh From owner-modality@LISTSERV.ARIZONA.EDU Fri Feb 5 19:42:54 1999 Date: Fri, 5 Feb 1999 19:42:50 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: Our ignorance of "all the physical facts" To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I'm heading out of town for a couple of days, but a quick reply to Brad and Josh. Brad says: >It seems to me that the most defensible materialist response to the anti >materialist arguments we've been discussing and reading about is the one >that the Churchlands and Dennett have given (I think all 3 philosophers >have given this basic response to Jackson's argument, Nagel's argument, >and the conceivability of zombies). The response is that we don't know >"all the physical facts", so that what we *succeed* in imagining when >we try to imagine Mary's situation (or similarly being a bat, or a >physical duplicate of me without consciousness) is not what the thought >experiment *asks* us to imagine. The materialist assumption that goes >with this reply is that *if* we did know all the relevant physical >facts, zombies would be inconceivable (and we'd know what it is like to >be a bat, and we'd be able to imagine being in Mary's situation and >wouldn't conclude that she didn't know what seeing red was like). Actually, I don't think this is really Dennett's main line of response (though I know he says something like this somewherte), and for good reason. Dennett holds that there is nothing more to qualia than certain reactive dispositions -- i.e. he holds that qualia are functionally definable, and/or that any sense in which qualia aren't functional concepts is a sense in which qualia don't exist. In folows from this that on Dennett's view, once we know the appropriate facts about which functions are performed -- the facts about discrimination, integration, higher-order monitoring, control of behavior, and verbal report -- we know everything. So Mary knows everything as soon as she knows that much (the apparent residue is just an illusion), and it's not really coherent to suppose that something could have that functional profile without having qualia (or: any sense of "qualia" in which this is coherent is a sense in which we don't have qualia). So relying on further unknown physical facts is more or less irrelevant to Dennett's position. I respect this position, as I think it is the most coherent line for a materialist to take. If qualia are to be conceptually entailed by physical facts, it will be by virtue of functional analysis of phenomenal concepts, so what matters are the functions, not the underlying physical facts. And the relevant work in seeing how the entailment can go through is in a sense conceptual (an a priori entailment will be grounded in an a priori (if deep) conceptual analysis, so one doesn't need to wave one's hands to future physical discoveries). Someone like Lewis is also admirably clear about this. Something like this comes out in what Brad notes: >But what about some yet unknown neurophysiological information--could >this close the epistemic gap (is this what Mary knows, that we don't)? >Well, it is difficult to imagine how such a view could resist positing >what Dennett would call "wonder tissue". What non-functional property >does the brain possess which makes it conscious? The obvious trouble is that the explanatory relevance of neurophysiology is always by virtue of the functions that the neurophysiology performs. So it's not the neurophysiology itself which is crucial to crossing the gap, it's the function, and we're back to the previous sort of position. To suppose that there's some other way that the neurophysiology might be relevant seems to lead to "wonder tissue" -- perhaps conceptual entailment by virtue of intrinsic properties rather than by virtue of functional role? -- and doesn't seem to hold up to a close analysis of what neurophysiology is all about. (Actually, I'm not completely hostile to the idea that there could be some sort of intrinsic entailment, but taken seriously the idea leads straight to "panprotopsychism"). Of course the "we don't know all the physical facts yet" position is very tempting at first blush, but it's not clear to me that this temptation should survive a careful analysis, and whether one can make coherent sense of the hopeful gesture. Entailment by virtue of a functional analysis I can understand, but then the real burden will be carried by analysis, not by the empirical details. And it's not at all clear how any other sort of entailment from physical to phenomenal could work, given the nature of physical concepts (and in any case, the central work will still be done by an analysis, which is conceptual). Of course a materialist might say that future physical discovery will lead to something utterly unlike the kind of physical knowledge we currently have, and which won't be subject to the same constraints. I think of this as more of a "new physics" than a "new neurophysiology" approach, and certainly not the sort of thing a Churchland or Dennett would be tempted by (it would require the sort of "real humdinger" than Churchland derides). Even here, I think that close analysis reveals that the options are pretty limited. All this being said, if someone wants to try and develop the "future physical knowledge" line into some sort of coherent position that evades the obvious objections, I think that would be a very useful service indeed, and I'd be very interested to see it. Josh writes: >What is it that `structural' means here? Does it mean physical >structure or is it more general? If it is more general, them I'm not >sure why consciousness isn't structured. Well, consciousness certainly has structure but (a) it's structural properties don't seem to exhaust the explananda (which most centrally involve intrinsic non-structural features), and (b) the structure it has doesn't seem to be the right sort to hook into entailment by physical facts. As you suggest here, the structure isn't physical structure. Physical structure seems to involve (a) spatiotemporal relations and (b) relations in certain underlying physical spaces, such as Hilbert spaces and the like. But this sort of structure is clearly conceptually compatible with the absence of consciousness. Even if one could get some sort of structural match-up (e.g. one might suppose that *given* that we have experience, one will have some sort of structural coherence), it's even less clear how one could exhaustively analyze the concept of experience in terms of static physical-structural concepts than in terms of functional concepts. (But still some interesting conceptual space to play with, so one might like to try!) >One note on the materialist insufficient-info rebuttal. We have been >discussing the rebuttal as if it has the following form. > >1. We don't know all the physical facts. >2. You can't have a good conception of Mary (zombies etc.) unless you >know all the physical facts. >C. Therefore the arguments fail. > >I think an important step is being left out of the real argument >though. > >3. Everything other than consciousness seems to be explainable in >terms of physics >4. It would be weird if there were only a single phenomina that >couldn't be explained by physics. >C. Therefore I have better reason for thinking I'm missing something >than for thinking physics can't explain qualia. > >You can, of course, argue whether you actually have better reason for >this. But its important to see that the materialist is comparing >possibilities and judging one to be more plausable. Well, of course there are lots of good reasons to be a materialist. That's more or less taken for granted in these discussions, by virtue of the fact that materialism is always regarded as the plausible default position which has to be dislodged, rather than vice versa. But of course the materialist still has to answer counterarguments. And as a counterargument, "maybe we're missing something" is not the strongest. Of course things would be different if one could provide some sort of faint gesture to just what we might be missing (not asking for an explanation, just for the conceptual room for an explanation). The anti-materialist can also try to "explain away" the force of (3) by noting that all those "everything else"s that have been explained seem to come down to the explanation of physical structures and of functions, and quite obviously so, whereas experience is prima facie not a matter of explaining either. So one has grounds right there for not giving the inductive argument too much force. I'm not really trying to convert anyone to anti-materialism here, and there are all sorts of reasons why materialism is an attractive position. But I do think the materialist has to face up to bullets that need to be bitten. As I said, I think the most coherent sort of materialism is the sort that just denies the epistemic intuitions, and is either functional or eliminative about the concept of qualia. Of course what the position gains in coherence it may lose in plausibility. There are other intuitively "more plausible" positions out there at least at first blush, but the question is whether they are philosophically coherent. Hopefully the rest of the course will provide us with some of the tools to examine such positions in some depth. --Dave. P.S. Kripke questions and comments are due by Tuesday. In the meantime this discussion is interesting, so please go ahead with any further thoughts in this vicinity or elsewhere re the second set of readings. A number of people still owe comments re this week -- soon, please! From owner-modality@LISTSERV.ARIZONA.EDU Fri Feb 5 01:35:36 1999 Date: Fri, 5 Feb 1999 02:24:54 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik A Herman Subject: Re: On Mary and Experience and Concepts To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I agree with Jackson but I don't find his examples very compelling. Learning from feeling and learing from being told/reading it are two totally different ways to experience something. They are different in kind and so can't really be compared. It's like telling someone blue feels cold-- you aren't gonna get it unless there's a one to one correspondence between blue and cold or sight categories and olfactory categories in general. If there WERE this correspondence then I would say that it's fair to translate visual information into verbal information in which case I think we WOULD be able to know red2. The situation with Mary: Again, I think she certainly will learn something when she sees the color TV. But what won't be immediately obvious to her is how the colors correspond to her prior interpretation of all the black and white things, other than there just being differences. As for the Modality Argument, I am concerned that there are inherent lawlike connections among physical things, and perhaps consciousness is one of these-- that by virtue of having a brain physically identical to mine, a duplicate must be conscious and in the same consious state. Oh, not to jump around but back to Fred, does it ruin everything if qualia is unique to each individual? Could it be that what you feel is your relation to that thing (so it involves both partys to the interaction)? Also, I can't help but think that Mary would dream in color-- does this count as a qualia? Erik H. From owner-modality@LISTSERV.ARIZONA.EDU Sat Feb 6 16:01:18 1999 Date: Sat, 6 Feb 1999 16:50:35 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony T Lane Subject: Re: Perplexity To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Here are a few general questions about last weeks material. A bit of a regression in the discussion. I'm afraid, but here goes... In section IV of *Epiphenomenal Qualia* Jackson discusses three arguments that might be presented to show that qualia might have some role in determining brain states. I suppose that his commitment to the impossiblity of qualia having any causal efficacy stems from the rest of his argument. If Mary were to know everything about physics, neurophysiology and so forth, but had not had the experience of seeing red, and seeing red had some causal role in determining brain states, then, presumably, she would not have all of the physical knowledge there is to have. There are two points I am not clear about here: First, I am a bit confused about the phenomenal/ psychological division. It seems to me that the phenomenal quality of out experiences does play a causal role in determining behavior. Isn't it the feeling of pleasure that one has when one sees a particularly beautiful sunset that leads one to try to have such experiences? I think this is probably a misunderstanding on my part, but it does seem that when I want to eat a pintof Ben and Jerry's, it is the phenomenal quality of the experience that I find particularly appealing. How does this work if we speak in purely psychological terms? This also gave me a worry in regard to the zombie argument. We are asked to admit the logical possibility of a world in which there are beings physically and functionally identical to us, but without consciousness. But why would these beings look at sunsets or eat ice cream? I don't know how behavior such as we exhibit is to be explained in terms of only functional states. Secondly, is it a fact that prior to having a particular kind of sensory experience we are incapable of having experiences of that sort? Could Mary, having read about the experience of seeing red and having examined various MRIs and what not of peole as they experience red, devise some sort of device to stimulate her brain in such a way as to have the phenomenal experience of red? Sorry to be rambling somewhat. A somewhat unrelated question... In the first part of chapter you talk about the form of dualism you are advancing. You suggest that phenomenal properties are basic and that there must be fundamental psychophysical laws relating these phenomenal properties to physical properties. Will these laws relate particular physical (brain) states to certain conscious experiences? Is the point of this that it is a contingent matter what psychophysical laws actually exist? Anthony From owner-modality@LISTSERV.ARIZONA.EDU Sun Feb 7 23:58:47 1999 Date: Sun, 7 Feb 1999 23:58:43 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Qualia To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Nobody needs to worry about "retarding discussion". It's important to be clear about things, and we can easily have numerous discussion topics open at the same time, so it's OK for a discussion of one's weeks topics to drag over into the next week. As long as the next week's discussion goes ahead just the same, things should be fine (N.B. everyone should make their initial contribution near the beginning of the week, though, to maximize discussion potential). Here are some random notes on comments, mostly about qualia, by Erik H. and Anthony. Erik writes: >I agree with Jackson but I don't find his examples very compelling. >Learning from feeling and learing from being told/reading it are two >totally different ways to experience something. They are different in >kind and so can't really be compared. It's like telling someone blue >feels cold-- you aren't gonna get it unless there's a one to one >correspondence between blue and cold or sight categories and olfactory >categories in general. If there WERE this correspondence then I would say >that it's fair to translate visual information into verbal information in >which case I think we WOULD be able to know red2. Well, for our purposes, Jackson is just trying to argue that the phenomenal facts are not entailed a priori by the physical facts. As far I can tell, you're agreeing with this here (e.g. you agree that Mary gains new knowledge that complete knowledge of physics and physiology can't give her). Of course it's then a further question whether this epistemic gap leads to an ontological gap, and that's one of the things we'll be looking at a lot more deeply. >The situation with Mary: Again, I think she certainly will learn >something when she sees the color TV. But what won't be immediately >obvious to her is how the colors correspond to her prior interpretation of >all the black and white things, other than there just being differences. Right, that may not be obvious from introspection. On the other hand, Dennett has a nice passage where he suggests that Mary might be able to figure out which color is which, etc, because she might know that different colors will cause slightly different reactions (e.g. emotions, twinges, flickers, etc) in her which she can identify. So when the scientist's try to trick her by presenting her with a blue banana, she isn't fooled into thinking it's yellow. I think that's entirely possible (it's an empirically open question), though I don't think it impacts the materialism debate. >As for the Modality Argument, I am concerned that there are inherent >lawlike connections among physical things, and perhaps consciousness is >one of these-- that by virtue of having a brain physically identical to >mine, a duplicate must be conscious and in the same consious state. Well, even a non-materialist like me thinks there are inherent lawlike connections between brain processes and consciousness. But for materialism to be true there has to be more than a lawlike connection -- witness the distinction between natural supervenience and logical supervenience. >Oh, not to jump around but back to Fred, does it ruin everything if qualia >is unique to each individual? Could it be that what you feel is your >relation to that thing (so it involves both partys to the interaction)? > >Also, I can't help but think that Mary would dream in color-- does this >count as a qualia? I think for the purposes of the issues re materialism, it doesn't matter whether individual's qualia are unique or not. It may well be that our qualia are at least "colored" by individual factors. Re Mary's dreams, you're right that this could well happen in practice (similarly she could presumably experience color patches when rubbing her eyes, and she'd see reddish qualities in her skin, etc). We just have to abstract away from those things for the purposes of the thought-experiment (or make Mary cortically colorblind, which would presumably do the job). ---------------------------------------------- Anthony writes: >In section IV of *Epiphenomenal Qualia* Jackson discusses three arguments >that might be presented to show that qualia might have some role in >determining brain states. I suppose that his commitment to the >impossiblity of qualia having any causal efficacy stems from the rest of >his argument. If Mary were to know everything about physics, >neurophysiology and so forth, but had not had the experience of seeing >red, and seeing red had some causal role in determining brain states, >then, presumably, she would not have all of the physical knowledge there >is to have. There are two points I am not clear about here: Well, I think Jackson is assuming that physics is causally closed. It's an interesting question what to say if interactionist dualism is true, i.e. if qualia are nonphysical and play a causal role with respect to a non-causally-closed physics. Maybe here one could still make a case that Mary could know (at least in principle) the physical facts about a system; she just wouldn't know about the nonphysical causes of some of the physical events. Anyway, if interactionist dualism is true, materialism is already false, so for the purposes of the argument one can take it as a concession to the materialist to rule that view out. Of course it then becomes a further question whether, upon rejecting materialism, one should accept epiphenomenalism or interactionism. (Actually, I think of the choice as a threeway disjunction between epiphenomenalism, interactionaism, and "panprotopsychism".) >First, I am a bit confused about the phenomenal/ psychological division. >It seems to me that the phenomenal quality of out experiences does play a >causal role in determining behavior. Isn't it the feeling of pleasure that >one has when one sees a particularly beautiful sunset that leads one to >try to have such experiences? I think this is probably a misunderstanding >on my part, but it does seem that when I want to eat a pintof Ben and >Jerry's, it is the phenomenal quality of the experience that I find >particularly appealing. How does this work if we speak in purely >psychological terms? To say that phenomenal concepts aren't functional concepts (or even to say that phenomenal states aren't functional states) doesn't automatically imply that phenomenal states don't play a causal role. It merely says, in effect, that they aren't *defined* by their causal roles. A concept such as "poison" is arguably a functional concept: to be a poison is roughly to be the sort of thing that has certain sickening effects on certain biological systems. So there's nothing more to the idea of a poison than the idea of something which plays a certain functional role. (Once Mary knows that something is the sort of thing that plays the role, she knows it's poison; one can't conceive of something being the sort of thing that plays the role without it being poison; etc, etc.) That's how it is arguably for very many concepts. But (according to the epistemic arguments) not so for qualia. Maybe these play a causal role, but that causal role doesn't define them. The point can be put somewhat controversially by saying that qualia are intrinsic properties rather than functional properties. Functional properties are defined by their causal role. Intrinsic properties are defined by their intrinsic nature. Of course intrinsic properties can still play a causal role. Indeed, presumably any causal role will ultimately be played by some intrinsic state or property. E.g. the "poison" role in any given case will be played by such-and-such a chemical structure. It's plausible that the chemical properties in question are intrinsic properties. (N.B. There is a case that these may be functional properties at a lower-level of analysis, but I leave that aside for now.) >This also gave me a worry in regard to the zombie argument. We are asked >to admit the logical possibility of a world in which there are beings >physically and functionally identical to us, but without consciousness. >But why would these beings look at sunsets or eat ice cream? I don't know >how behavior such as we exhibit is to be explained in terms of only >functional states. Well, the first thing to note is that it's not obvious that the same strictures of "explanation" hold in logically possible worlds as hold in the actual world. E.g. there are presumably possible worlds where monkeys coincidentally type "Hamlet", where the world explodes for no reason at all, etc. But anyway: presuming that actual physics is causally closed, I suppose the zombie's behavior might be explained simply in terms of its physical states. When it eats ice creams, certain receptors are triggered which leads to a state which it "likes", i.e. which leads it to attempt to return to the state again. It's not obvious that there's a problem in principle with explaining the behavioral part in physical terms. Of course one may find it counterintuitive that qualia don't play a causal role in our actual behavior, in which case there may be at least some reason for the nonmaterialist to consider alternatives on which they play such a role (such as interactionism or panprotopsychism). >Secondly, is it a fact that prior to having a particular kind of sensory >experience we are incapable of having experiences of that sort? Could >Mary, having read about the experience of seeing red and having examined >various MRIs and what not of peole as they experience red, devise some >sort of device to stimulate her brain in such a way as to have the >phenomenal experience of red? It's very likely that she could. But I don't think this affects the central dialectic, which concerns what Mary can know merely by virtue of knowing the physical facts and reasoning about them. Direct brain stimulation in the production of experience is obviously a further "empirical" sort of knowledge. Crucially, the fact that Mary can do this in no way suggests that the phenomenal facts are entailed a priori by the physical facts. At best this relies on some sort of contingent a posteriori connection between them, the sort that even a non-materialist might embrace. (E.g., if there are contingent psychophysical laws connecting brain states to experiences, this method ought to work.) >Sorry to be rambling somewhat. A somewhat unrelated question... In the >first part of chapter you talk about the form of dualism you are >advancing. You suggest that phenomenal properties are basic and that there >must be fundamental psychophysical laws relating these phenomenal >properties to physical properties. Will these laws relate particular >physical (brain) states to certain conscious experiences? Is the point of >this that it is a contingent matter what psychophysical laws actually >exist? Well, I think there will be psychophysical laws relating brain states to experiences; e.g., it will be a law that when you have such-and-such a brain state, you have an experience of red. But those laws probably won't be fundamental. It's unlikely that universal laws of nature will make reference to brains, etc. So one would expect there to be some much simpler underlying principles (e.g. principles linking information and experience?), from which the specific high-level laws are consequences. I take it that their nature is very much an open question, though. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Sun Feb 7 21:15:06 1999 Date: Sun, 7 Feb 1999 22:04:33 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Rachael J Parkinson Subject: more on reductive explanation To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Hello All- Like Anthony, I hope I am not holding discussion back here... In 3.3 of TCM, Chalmers criticizes attempts at reductive explanations of consciousness. In his critique of Baars cognitive model he argues, "But there is no reductive explanation of *experience* to be found here. The question of why these processes should give rise to experience is simply not addressed."(112) I take it that this is because the explanation of lower-level facts does not (and cannot) explain the higher-level phenomena of consciousness. I am curious as to how this is different from the problem of reductive explanation in the case of the football game or the cocktail party. It does not seem that in giving a complete description of the lower level facts of a football game, I am really explaining how the lower level facts give rise to a football game. What is necessary for a complete explanation is not only a description of the lower level facts but also a description of the higher level phenomena (in terms of plays, players, touchdowns, etc.) >From this, we may propose that what is necessary to supplement a cognitive or neurobiological model is a complete description of the higher level phenomena in question, specifically, consciousness. I assume that a description of this sort would rely heavily on psychology. It seems that combining a reliable cognitive or neurobiological model with a comprehensive psychology should be able to play the same sort of explanatory role as describing a football game in terms of both the lower and higher level facts. It is tempting to conclude that this sort of explanation of consciousness, in terms of lower level and higher level facts is satisfactory. However, objections to explanations of this type are given by Chalmers in this chapter. To take just one, we can conceive of something physically identical to ourselves not having consciousness (a zombie). But can we conceive of something physically identical to a football game not being a football game? Although it is conceivable that such a thing not be *called* a football game, it is safe to say that the higher level phenomena of a football game will be in place once all of the lower level facts are fixed. My question is one which arises from Chalmers objection to cognitive models. Chalmers says, "The question of why these processes should give rise to experience is simply not addressed."(112) I am curious as to why and how these processes *should* give rise to experience, but as I have not read all of TCM, I take it the answer will surface at some point in my reading. -Rachael From owner-modality@LISTSERV.ARIZONA.EDU Mon Feb 8 04:45:34 1999 Date: Mon, 8 Feb 1999 04:45:30 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: more on reductive explanation To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Some thoughts in responses to Rachael's and Erik L's notes on reductive explanation: Rachael writes: >I am curious as to how this is different from the problem of reductive >explanation in the case of the football game or the cocktail party. It >does not seem that in giving a complete description of the lower level >facts of a football game, I am really explaining how the lower level facts >give rise to a football game. What is necessary for a complete >explanation is not only a description of the lower level facts but also a >description of the higher level phenomena (in terms of plays, players, >touchdowns, etc.) Well, giving the low-level facts alone will never be enough for a reductive explanation, since you have to connect them to the high-level story. That's the part which is played by explication of the high-level concepts (e.g. as functional concepts), and by showing how the low-level properties can entail the high-level properties. For functional concepts such as "poison", "life", and the like, this is fairly straightforward. Things are mnore complex for football games and cocktail parties, mostly because a complete explanation of these things centrally involves various *intentional* properties, characterizing people's attitudes, beliefs, desires, etc. And it's not at al obvious how these intentional properties are to be reductively explained. But leaving aside these worries stemming from the "mental" aspects of the games and parties, it seems pretty clear that the residual high-level phenomena in question (the movement of people and balls and drinks, the production of utterances, etc) can be accounted for in physical terms. As before, it may not be the most illuminating explanation of all time, but we'll remove any ontological mystery. The intentional aspects pose special worries, but here of course we're back to the mind-body problem itself. Personally I am hopeful that the intentional can be explained in terms of the functional and the phenomenal, but I don't say this is completly obvious. (N.B. The book is officially neutral about whether intentionality can be functionally explained, but I was probably more sympathetic to the idea then than I am now. I now tend to think that phenomenology may be central to a full understanding of intentionality.) >From this, we may propose that what is necessary to supplement a cognitive >or neurobiological model is a complete description of the higher level >phenomena in question, specifically, consciousness. I assume that a >description of this sort would rely heavily on psychology. It seems that >combining a reliable cognitive or neurobiological model with a >comprehensive psychology should be able to play the same sort of >explanatory role as describing a football game in terms of both the lower >and higher level facts. Well, it's not just a matter of giving the low-level facts and the high-level facts at the same time. One needs to have the right sort of relationship of entailment between them. And as you point out below, it's this epistemic relationship that seems to be missing in the case of consciousness. >My question is one which arises from Chalmers objection to cognitive >models. Chalmers says, "The question of why these processes should give >rise to experience is simply not addressed."(112) I am curious as to why >and how these processes *should* give rise to experience, but as I have >not read all of TCM, I take it the answer will surface at some point in my >reading. I'm not certain what the question is here. If it's about why and how the processes give rise to experience on my view, it's because there are fundamental psychophysical laws linking the physical and the phenomenal. At some point it's like "why do apples fall? Because of the law of gravity". The apple and the earth stand in the right sort of relation for the law to apply; and the brain is presumably in the right sort of state for psychophysical laws to apply. As to why the law exists, at some point one has to take that as a brute fact, just as in physics. Of course one wants the residual "brute" part to be as simple as possible. Actually, we won't be talking much about these issues in the rest of the seminar, as we'll be focusing more on modality than on mind, but I'm sure some of these things will come up. ---------------- Erik writes: >"not very >illuminating" is perplexing in it's own right. What makes something >"illuminating"? That's a strange notion once you've conceded that the >process referred to is nothing over and above physics. I guess what I'm >asking is, what is the role--ontologically, functionally, explanatorily or >whatever--of concepts in explanations and why should they be NECESSARY if >a supervenience relation holds between the phenomenon and a set of >entirely physical facts? For that matter, WHAT IS A CONCEPT in this sense >that we keep using it? Re "illuminating", presumably this is an epistemic notion, and (unlike "a priori") not an idealized one, so relative to human psychology. So A can a priori entail B, but the entailment may not be terribly illuminating to us, because we're not in a position to easily grasp the deep structure in what is entailed, or something like that, at least not without a lot of further work (and that further work is what one might call "high-level explanation"). Whereas perhaps a more cognitively skilled race of beings might see these things much more easily. Re the role of concepts: If the physical facts entail the B facts a priori, then one who merely possesses the B concepts ought to be able to figure out the B facts from the A facts. If this person can do this merely in virtue of concept possession (plus reasoning)m it seems clear that the concept has to be of the right sort to enable the entailment. In particular, there ought to be some sort of analysis of the concept's conditions of application so we can see why they are satisfied automatically whenever the A facts are satisfied. The main point is that a priori entailment is an epistemic relation, and epistemic relations seem to be grounded in relations between concepts. The question of how these concepts (in our mind) correspond to properties (in the world) and what this means about the bridge between epistemoloy and ontology is one we'll be thinking about. What's a concept? That's a hard one. We've been more or less helping ourselves to the notion at an intuitive level, requiring only that there is a concept corresponding to (most) words, that concepts have conditions of application, and that these conditions of application when composed yield the conditions of application of corresponding sentences. Or something like that. The underlying metaphysics of concepts is far from clear (though maybe the 2-D framework helps a little in clarifying their structure). But I hope even the intuitive level, thinking of concepts in terms of their pattern of application to possibilities, is helpful for our purposes. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Mon Feb 8 05:30:48 1999 Date: Mon, 8 Feb 1999 05:30:45 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: Conceivability and Knowledge arguments To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Sorry to the slow reply to Tim's very interesting message about conceivability. First, Tim suggests that the knowledge argument may be a conceivability argument, since one has to conceive of Mary in her lab coat, etc. Personally, I think this act of conceiving is pretty inessential (certainly I make not the slightest effort to really conceive of Mary herself when thinking about this). What's more central is the claim that conceivability or no, it's extremely plausible that someone with full physical knowledge of the brain is not thereby in a position to know what it's like to see red, and will learn something new when they see red for the first time. I don't think one needs to engage in detailed acts of conceiving to find this plausible. The zombie case, on the other hand, at least involves a token attempt at conceiving the details of the case. I try to conceive of someone functioning just like me, behaving like me, with mechanisms of perceptual discrimination, attention and integration, higher-order monitoring, control of behavior, verbal report, etc, but no consciousness. Of course one's attention to detail is highly imperfect here, but as before there may be reaons to believe that the details can't make the crucial difference. >But in fact, I doubt that we do anything *like* this in either case. >Conceivability does *not* seem to involve the imagination - at least >that's what it seems like to me in *these* cases. It's *not* like trying >to work out in your head whether you can fit a ball of diameter two inches >inside a triangle of sides 3, 4 and 4 inches which has an outer perimeter >of 1/2 an inch thick. I'm not surely exactly why you say this. Certainly there are some senses in which conceivability needn't involve imagination -- e.g. if one acquaints imagination with visual or sensory imagery. But it seems plausible to me that conceivability involves *something like* imagination -- e.g. a sort of "conceptual imagination", involving considering a lot of details about a case, and seeing if they are coherent and compossible. >So what is going on in these cases? Well, it seems to me that in both >cases you have the following intuitions or you don't. > >(a) Necessarily, all physical properties are functional or structural >properties, and phenomenal properties aren't functional/ structural >properties; OR, the very similar intuition: (b) Nec., all physical facts >(or, perhaps, propositions) are exclusively composed out of functional >/structural concepts, and phenomenal concepts are not >structural/functional concepts. > > >(a) is at work in the zombie cases, (b) is at work in the knowledge cases. >If there is a one-one corresondence between facts and properties as is >commonly supposed in setting these cases up, then it is hard to see why >one should have (a) and not (b), or vice-versa. (Caveat: Of course, some >do reject the claim that if you've got two facts you've also got two >properties, but then you can accept that Mary does learn something new >without rejecting physicalism.) Hmm, this is interesting. It may well be the case that a tacit acceptance of something like your principles (a) and (b) underlies our acceptance of the conceivability and knowledge intuitions. But I think it would at least be highly tacit. Plenty of people would be unsure whether to accept the abstract principle, but quite confident about the conceivability or knowledge intuition. And it's certainly the case that a good number of people accept the Mary intuition while being uncertain about the zombie intuition, at least in its full-blown form. So I think there's reason to believe that something is going on in our evaluation of the conceivability and knowledge intuition that is at least somewhat different from each other and somewhat different from our evaluation of the abstract principles, though it may be the case that the three are a priori connectable and that someone who accepts one should accept all three. Incidentally I don't quite accept your diagnosis of (a) going with zombie intuitions and (b) with knowledge intuitions. It seems to me that (b) is weaker than (a), as its a claim about concepts, not properties. The type-B materialist who holds that there is an epistemic gap but not an ontological gap may well accept (b) but not (a), for example. But this type-B materialist may well still have the zombie intuition, i.e. will find zombies conceivable; they'll just deny that conceivability implies possibility. So that suggests that (b), not (a), may underlie the zombie intuition, if anything does. (Of course, one may call the "zombie intuition" the claim that zombies are metaphysically possible; then accepting the intuition may well go along with accepting (a). But then the zombie intuition isn't the pure epistemic intuition. I prefer to reserve the term "intuition" for the epistemic part, as the residual part is a matter of straight philosophy, not really intuition.) Another slight difference between zombie and knowledge intuitions. Take someone who holds the epistemic theory of vagueness, according to which there is a sharp line in the application of vague predicates, although we can't know where that line is. So e.g. someone 5'10" is either tall or not tall (abstracting away from issues about context), but we can't know which through any amount of reasoning. Such a theorist will hold that complete physical knowledge won't enable someone (e.g. Mary) to know whether the person is tall. But it's not clear that such a person will hold that a 5'10" tall person and a 5'10" short person are both conceivable in the way that me and my zombie twin are both conceivable. Maybe they are "negatively conceivable" in that we can't rule them out a priori, but they don't seem to be "positively conceivable" in the sense that we can clearly and distinctly conceive of both scenarios. Rather there seems to be just one scenario that we conceive of, such that we can interpret it in two different ways. So it seems that zombie intuitions involve a slightly stronger claim than knowledge intuitions. The former requires "positive conceivability", the latter mere "negative conceivability" or absence of contradiction. So maybe that is more evidence that the two are slightly distinct. The distinction between these two sorts of conceivability is something we'll be discussing quite a bit in a few weeks. >But imagination does seem to play some role in conceivability. Consider >the following problems: (1) is logically it possible to be an agent >without, being a perceiver (or vice-versa)? (2) Is it possible for one >subject of experience to have two distinct (spatially separated) bodies at >the same time, so that, e.g., one of my bodies might be in Tucson, and the >other one in Texas at the same time, and they are simultaneously >conscious? Well, people seem to have very different intuitions about both >of these cases? Why? Are they: (1) actually imagining different scenarios? >or (2) imagining the same sorts of scenarios, but putting different labels >on them because of the differing structure of their concepts? I dunno, but >I'm tempted towards (2). I'm tempted to agree with you here. It seems that at least in (1), the scenario one conceives of is pretty clear; the question is then how to describe and interpret it. One positively conceives a being who can't see/hear/etc, but who seems to move around, etc, and one considers whether that being is an agent. So we have a core conceived scenario and an issue of how to describe it. A little like the epistemic vagueness case, actually, in that there's just one positively conceivable scenario in the vicinity. Contrast the zombie case, where there seem to be two (at least to some people -- at least here the issue is about whether there is one or two, rather than about how to interpret the obvious one). Maybe something similar goes in (2). >Perhaps conceivability and knowledge arguments are different types of >arguments because they are different ways of pumping the same intuitions >(i.e. the same concepts), but I'm not even sure that this is true. I tend to think they are ways of pumping similar intuitions, though perhaps not exactly the same intuitions. They certainly play separable roles as argument and a rhetorical devices, though of course that's often the case with "intuition pumps". As often in philosophy, there can be two closely related ways to reach the same conclusion, I agree, though, that the key underlying point in both cases concerns the gap between our concepts. And the central questions are then (a) is there an a priori gap between the concepts and (b) if som does it lead to a gap in properties. The type-A materialist denies the gap in concepts. The type-B materialist accepts the gap in concepts but denies that it leads to a gap in properties. This type-B materialist will typically lean on a posteriori necessities ("water" and "H2O" are arguably different concepts picking out the same properties) and on Kripke's discussion thereof. Which leads nicely into this week's topic. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Tue Feb 9 06:59:06 1999 x-sender: agillies@pop.u.arizona.edu Date: Tue, 9 Feb 1999 08:10:51 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony S Gillies Subject: twins and such To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO All, Here's a puzzle I have. My trouble is in imagining my zombie twin. (1) I think it's pretty clear that for purposes of the materialism debate it won't do for me to imagine (say) your zombie twin. You might *be* a zombie for all I know. Since I don't have any immediate access to your subjective states, a case where you don't consciousness isn't any different (from my vantage point) than a case where you do. So I have to imagine my own zombie twin. (2) OK, now how do I do that? I imagine zombie twin earth. The problem is that there are lots of (zombie) folks there. How do I pick out my zombie twin? Since it's crucial that I pick out *my* zombie twin, I figure that I tag him *Zombie Thony*, and use the name to identify him. Imagining my zombie twin comes down to evaluating the truth of (*) Necessarily, Thony is conscious Materialism is false if (*) is false according to the 1-intensions. (3) The trouble with this is (as I think you say in TCM) consciousness is at the core of our epistemic universe--we can't help but think we're conscious. But then the 1-intension of 'Thony' (when thought by me) will include 'is conscious'. And so (*) is true after all. Apparently something has gone wrong: my zombie twin *seems* possible, but in fact isn't when I try to imagine him. Conceivability and possibility seem to come apart. But I don't think things are really that simple. My guess is that one might be tempted to say that what I need to imagine is not really *my twin*, just any old thing that is physically identical to me. I'm probably missing something in this would-be reply. I thought that what's at stake is the truth of modal statements like (*), and imagining something physically identical to me doesn't decide that. Thony "Curious green ideas sleep furiously." From owner-modality@LISTSERV.ARIZONA.EDU Tue Feb 9 09:40:17 1999 Date: Tue, 9 Feb 1999 10:43:54 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Re: twins and such To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Reply to Thony on twins and such. (Thony knows my line on this, but I thought it might be useful to make our discussion public.) I think the false move (if there is one here) is the first move. Why won't it do to imagine somebody else's zombie twin? As far as I can see, it will do. I take it that the argument from zombiehood is meant to show that you can (logically) have all the functional/physical properties that we have, without having conscious experience. The reason why the zombie arguments ask you to imagine your own zombie, is that they hope that in this way they *guarentee* that you've got the sort of physical/functional properties on which consciousness naturally supervenes (since one supposes that one is oneself consciousness). I read Thony's worry re imaging his zombie twin in the following way: if consciousness is one of my essential properties, then I don't have a zombie twin - there is no world in which there I exist and am not conscious (or, perhaps, capable of having conscious experience). I think Thony reads Dave is endorsing this premise or something like it. Whether or not Dave does, it's not an unreasonable assumption. Suppose that we grant this point. I take is that the zombie argument can make do with *qualitative* identity rather than *numerical* identity. Suppose there is an organism just like you, twin Thony. Twin Thony is physically and functionally identical to you. That is, for every physical property that you instantiates, he also instantiates, and vice-versa. But we can conceive that twin Thony is not conscious, and if conceivability= logical possibility, consciousness does not (logically) supervene on the physical. As I say, I think that you can probably run the argument without getting into worries about trans-world identity and individuation (which is good!): what matters is that isolate the sort of physical facts that the reductive materialist insists that consciousness is logically supervenient (and then you go to work with the zombie intuitions. . Talking about *My* zombie twin is just used (perhaps wrongly) as a convenient way to isolate those facts. Tim 'm not sure that the first step of On Tue, 9 Feb 1999, Anthony S Gillies wrote: > All, > > Here's a puzzle I have. My trouble is in imagining my zombie twin. > > (1) I think it's pretty clear that for purposes of the materialism > debate it won't do for me to imagine (say) your zombie twin. You might > *be* a zombie for all I know. Since I don't have any immediate access to > your subjective states, a case where you don't consciousness isn't any > different (from my vantage point) than a case where you do. So I have to > imagine my own zombie twin. > > (2) OK, now how do I do that? I imagine zombie twin earth. The problem > is that there are lots of (zombie) folks there. How do I pick out my > zombie twin? Since it's crucial that I pick out *my* zombie twin, I > figure that I tag him *Zombie Thony*, and use the name to identify him. > Imagining my zombie twin comes down to evaluating the truth of > (*) Necessarily, Thony is conscious > Materialism is false if (*) is false according to the 1-intensions. > > (3) The trouble with this is (as I think you say in TCM) consciousness > is at the core of our epistemic universe--we can't help but think we're > conscious. But then the 1-intension of 'Thony' (when thought by me) will > include 'is conscious'. And so (*) is true after all. > > Apparently something has gone wrong: my zombie twin *seems* possible, but > in fact isn't when I try to imagine him. Conceivability and possibility > seem to come apart. But I don't think things are really that simple. My > guess is that one might be tempted to say that what I need to imagine is > not really *my twin*, just any old thing that is physically identical to > me. I'm probably missing something in this would-be reply. I thought > that what's at stake is the truth of modal statements like (*), and > imagining something physically identical to me doesn't decide that. > > Thony > > > > "Curious green ideas sleep furiously." > Timothy J. Bayne RM. 213 Social Science Department of Philosophy University of Arizona Tucson, AZ 85721 USA Hm ph. (520) 298 1930 From owner-modality@LISTSERV.ARIZONA.EDU Tue Feb 9 11:25:50 1999 Date: Tue, 9 Feb 1999 12:25:07 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik J Larson Subject: Re: twins and such To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Some further thoughts on Thony and Tim's comments re "twins and such" I admit it is puzzling to imagine imagining a physical replica of yourself without conscious experience. The gut feeling is something like, What would it mean to imagine MYSELF without my consciousness? I suppose in so far as one is trying to imagine one's twin "from within", there is no way to get the physical-replica-without-consciousness conception off the ground. So I suppose the conception of a zombie-twin or a physical replica of oneself sans consciousness when you want to construct it from the first-person point of view--that is the very perspective that is supposed to be excluded. So I suppose the job is to determine whether this perspective--the "from within" perspective"--can be detached from the rest of the picture in imagining a replica of oneself. I can imagine myself without an arm--all else being equal. And so on for all sorts of physical alterations. So the problem is not one of simply subtracting physical constituents--we can do that. Personality is trickier. Can one imagine ONESELF without certain memories, or dispositions, etc? Now I suppose that's a question of personal identity, and I think it is more puzzling than the case at hand. That is because, imagining yourself without conscious experience is tantamount to imagining yourself qua physical THING, or, imagining yourself as someone else would see you--from the third-person so to speak. And what is so tricky about that? We do it all the time, or can do it all the time, as even a moments reflection should convince one of. I personally do not have any trouble imagining myself qua physical thing--looking exactly the same, moving the same, behaving and reacting the same--only without any conscious experience whatever. Now I can't imagine ME in any strong sense this way, since I think that part of my personal identity is my conscousness, or that my consciousness is a precondition of having a personal identity, perhaps. So that is not a possiblity. But, I think it is also not at issue here. What is at issue is whether we might conceive of ourselves qua physical thing, identical in all physical (and behavioral) aspects but lacking conscious experience. Would this be you in any strong sense? No. But would it be you in precisely the sense required--having the physical properties en mass but without conscious experience. So, replying to Thony's point directly, I think he is conflating questions of personal identity with questions of purely physical conceivability, where the latter is less troublesome. Erik L. On Tue, 9 Feb 1999, Timothy J Bayne wrote: > Reply to Thony on twins and such. > > (Thony knows my line on this, but I thought it might be useful to make our > discussion public.) > > I think the false move (if there is one here) is the first move. Why > won't it do to imagine somebody else's zombie twin? As far as I can see, > it will do. I take it that the argument from zombiehood is meant to show > that you can (logically) have all the functional/physical properties that > we have, without having conscious experience. The reason why the zombie > arguments ask you to imagine your own zombie, is that they hope that in > this way they *guarentee* that you've got the sort of physical/functional > properties on which consciousness naturally supervenes (since one supposes > that one is oneself consciousness). > > I read Thony's worry re imaging his zombie twin in the following way: if > consciousness is one of my essential properties, then I don't have a > zombie twin - there is no world in which there I exist and am not > conscious (or, perhaps, capable of having conscious experience). I think > Thony reads Dave is endorsing this premise or something like it. Whether > or not Dave does, it's not an unreasonable assumption. > > Suppose that we grant this point. I take is that the zombie argument can > make do with *qualitative* identity rather than *numerical* identity. > Suppose there is an organism just like you, twin Thony. Twin Thony is > physically and functionally identical to you. That is, for every physical > property that you instantiates, he also instantiates, and vice-versa. But > we can conceive that twin Thony is not conscious, and if conceivability= > logical possibility, consciousness does not (logically) supervene on the > physical. > > As I say, I think that you can probably run the argument without > getting into worries about trans-world identity and individuation (which > is good!): what matters is that isolate the sort of physical facts > that the reductive materialist insists that consciousness is logically > supervenient (and then you go to work with the zombie intuitions. . > Talking about *My* zombie twin is just used (perhaps wrongly) as a > convenient way to isolate those facts. > > > Tim > > > 'm not sure that the first step of > > On Tue, 9 Feb 1999, Anthony S Gillies wrote: > > > All, > > > > Here's a puzzle I have. My trouble is in imagining my zombie twin. > > > > (1) I think it's pretty clear that for purposes of the materialism > > debate it won't do for me to imagine (say) your zombie twin. You might > > *be* a zombie for all I know. Since I don't have any immediate access to > > your subjective states, a case where you don't consciousness isn't any > > different (from my vantage point) than a case where you do. So I have to > > imagine my own zombie twin. > > > > (2) OK, now how do I do that? I imagine zombie twin earth. The problem > > is that there are lots of (zombie) folks there. How do I pick out my > > zombie twin? Since it's crucial that I pick out *my* zombie twin, I > > figure that I tag him *Zombie Thony*, and use the name to identify him. > > Imagining my zombie twin comes down to evaluating the truth of > > (*) Necessarily, Thony is conscious > > Materialism is false if (*) is false according to the 1-intensions. > > > > (3) The trouble with this is (as I think you say in TCM) consciousness > > is at the core of our epistemic universe--we can't help but think we're > > conscious. But then the 1-intension of 'Thony' (when thought by me) will > > include 'is conscious'. And so (*) is true after all. > > > > Apparently something has gone wrong: my zombie twin *seems* possible, but > > in fact isn't when I try to imagine him. Conceivability and possibility > > seem to come apart. But I don't think things are really that simple. My > > guess is that one might be tempted to say that what I need to imagine is > > not really *my twin*, just any old thing that is physically identical to > > me. I'm probably missing something in this would-be reply. I thought > > that what's at stake is the truth of modal statements like (*), and > > imagining something physically identical to me doesn't decide that. > > > > Thony > > > > > > > > "Curious green ideas sleep furiously." > > > > Timothy J. Bayne > RM. 213 Social Science > Department of Philosophy > University of Arizona > Tucson, AZ 85721 > USA > > Hm ph. (520) 298 1930 > From owner-modality@LISTSERV.ARIZONA.EDU Tue Feb 9 23:53:44 1999 Date: Tue, 9 Feb 1999 23:53:40 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: twins and such To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I'm sympathetic with what Tim and Erik say in reply to Thony's question, and I don't have a whole lot to add to their reply on the central point. But there are a number of very interesting issues in the vicinity, some of which tie in with current and forthcoming issues on Kripke and 2-D semantics, so here are a few thoughts. In the conceivability argument, I don't think too much turns on whether the conceived being is one's own zombie twin. It can just as well be the zombie twin of any conscious being. Of course in one's own case there is the advantage that one is certain the actual being in question is conscious. But on the plausible assumption that other beings are conscious, the argument will work just as well for them. Thony suggests a problem: imagining someone else to be a zombie won't be any different from imagining them to be conscious, because both will seem the same from my point of view. I think I disagree with this. There seem to be two pretty distinct epistemic possibilities that I can entertain re that being: that they are conscious or that they are not. It's true that both would "look" the same to me, but this just means that we not should individuate epistemic possibilities according to the way they look, or the way they seem, or according to the evidence one would have. I take it that to suppose that conceivability or epistemic possibility is determined by appearances or evidence involves a sort of verificationism; and I take it that verificationism is usually taken to be implausible precisely because there seem to be distinct epistemic possibilities that would yield exactly the same evidence. E.g., empirically equivalent theories in science, or possibiities involving entities that leave no trace, or indeed possibilities involving consciousness. So I think doing the thought-experiment on others is OK. Still, I suppose one doesn't want to be *forced* to do the thought-experiment only for others, as if one did, then presumably someone could evade the argument by accepting solipsism (i.e. denying that any other beings are conscious), and the conclusion of the argument would be "materialism is false or solipsism is true". Solipsism isn't exactly attractive, but this would be messy, so it would be best if the argument works for one's own twin. Thony raises the worry: if it has to be *me* that is a zombie, maybe this can't happen, because consciousness may be part of the primary intension of our concept of ourselves, or some such. Here I would endorse what Tim and Erik say in reply. The epistemic argument against materialism merely involves the claim that "P -> Q" is not a priori, or that P-and-not-Q is conceivable, where P is the complete physical truth about the world and Q is a phenomenal truth. Nothing here requires that the zombie world contains me, or that the zombie twin be me; it merely has to be a physical duplicate of our world, and he merely has to be a physical duplicate of me. And here, Thony's worry doesn't arise. Still, it's an interesting question: could my zombie twin be me? Speaking for myself, I don't usually think of my zombie twin as being me; I think of him perhaps as some sort of close relative. But could I be a zombie? Matters depend here on whether one is thinking of standard "subjunctive possibility" (secondary-intension possibility) or "epistemic possibility" (primary-intension possibility). In the subjunctive case, we're considering the standard sort of possibility that Kripke and others discuss. Here, I can imagine someone arguing that it is essential to sentient human beings (like me) that they be at least potentially sentient, so that the hypothesis that I could be a zombie is like the hypothesis (discussed by Kripke) that I could have descended from a different sperm and egg, etc. If so, the 2-intension of "me" and "DC", etc, can only pick out potentially conscious beings, and it's (secondarily) impossible that I be a zombie; counterfactual possibilities concerning a zombie should not be described as possibilities concerning me. (I say "potentially conscious" as it's quite possible that I be temporarily unconscious, and perhaps that I died before becoming conscious.) I'm not sure whether this essentialist claim is right, but it has a certain possibility. The epistemic possibility (primary intension) case is closer to what matters in the epistemic arguments, though, since we're dealing with a priori connections, and in such epistemic matters, primary intensions are central. Is it epistemically possible that I'm a zombie? By this, we're in effect asking (modulo a few small frills): is there a conceivable scenario satisfying the primary intension of "I am a zombie"? (Note that this isn't quite the standard sense of "epistemic possibility", where P is epistemically possible if it's true for all I know. Obviously, I know I'm not a zombie, so it isn't epistemically possible in that sense. Rather, I'm invoking the Kripke-esque sense in which it's epistemically possible that water is XYZ, even though we know that it isn't, and so on. We could call this "broad epistemic possibility" -- it is broadly epistemically possible that P when P isn't ruled out a priori. If you don't like calling this "epistemic possibility", just think of it as "1-possibility" instead, though I think it is epistemic in some deep sense. We'll come back to this.) On my view, it is 1-possible that I'm a zombie. There is a centered world centered on a zombie, and the primary intension of my term "I" picks out the zombie in that world, and the primary intension of my term "consciousness" doesn't apply to that being. Thony may not like this, on the grounds that the primary intensions of our conceptions of ourselves include consciousness. But on my view, that isn't quite right: the primary intension of my "I" concept just picks out the person at the center of a centered world. Of course I very strongly bnelieve (even know) that I am conscious, but this isn't part of the semantics of "I". Similarly (but more so) for concepts such as "DC", "Thony", etc. So in a centered world centered on an unconscious being, the primary intension of "I" picks out that being. One can come at this slightly differently, by forgetting the 2-D details for a moment, and asking directly whether it is broadly epistemically possible that I am a zombie, i.e. whether it is a priori that I am not a zombie, i.e. whether it is a priori that I am conscious. On my view, this isn't a priori. This is something we know, and know directly and in some sense "non-empirically", but we don't know it through reason alone; we know it in virtue of having certain experiences. I can't know that I am conscious independently of experience, so it isn't a priori. So I can still consider the broad epistemic possibility that I am not conscious (even though it is a possibility easily ruled out by things that I know), and I can use my concepts to describe that epistemic possibility, etc. The broad epistemic possibility in question (in effect, the hypothesis that my world is centered on a zombie) is describable as an instance of the epistemic possibility that I am a zombie, an instance of the possibility that I am not conscious, etc. So this is a 1-possibility, and is a broad epistemic possibility, even though I know it doesn't obtain, so it isn't an epistemic possibility in the standard sense. So, on my view, it probably isn't 2-possible that I'm a zombie, but it is 1-possible that I'm a zombie. None of this matters much for the argument against materialism (for which the semantics of "I" is mostly tangential), but it's independently interesting, especially for the illustration it gives of the way 2-D semantics and modality works. Obviously this hooks up nicely with the Kripke discussion, and some of it turns out to be important later. What I say above certainly isn't trivial or obvious or uncontestable, so feel free to fire away with questions and reactions. --Dave. P.S. A few of these issues come up very briefly around p. 133 of the book (middle paragraph), where I suggest that (in effect to avoid the technicalities above which arise from considering epistemic possibilities in which I'm not conscious), it's easiest to consider conceivable scenarios in which I'm at the center, conscious, and in which some other people are zombies. So I suppose my official version runs with third-person zombies (and so officially has the solipsism loophole). In principle, I think things work even with zombies at the center, though I don't talk about that much in the book. The above gets at the beginning of how this works, though there are some more subtleties I haven't mentioned yet about whether concepts are required at the center of centered worlds. That turns out to be important for various purposes, too. P.P.S. More on the Kripke messages shortly. It would be good to get something from everybody soon (I have Thony's and Tim's contributions posted a couple of weeks ago already). I also look forward to getting a detailed summary of Tuesday's discussion (the more detailed the better).