From owner-modality@LISTSERV.ARIZONA.EDU Sun Mar 14 16:57:22 1999 Date: Sun, 14 Mar 1999 16:57:11 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Readings, etc To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Someone asked that I post the readings for next meeting (after spring break). This meeting will be "week 8" from the syllabus, on "Does Conceivability Entail Possibility?". The main readings are: Chalmers, Mind and Modality, part 2. Yablo, Is conceivability a guide to possibility? van Cleve, Conceivability and the Cartesian argument for dualism We won't be going over the Yablo and van Cleve in huge detail, but they provide very interesting and useful background material for the general issues here, so you should go over them carefully. In seminar I will be developing the various distinctions between different kinds of conceivability in section 2.3 of the Princeton lectures, and will probably also go over the varieties over modal rationalism in 2.4. So you should concentrate especially on those sections and try to understand them and come with any questions. We'll be going over the material in 2.5 and later in coming weeks. You should also read the Hill paper on conceivability from Philosophical Studies, if you haven't already, and Loar's paper "Phenomenal States", as these are both quite relevant. We've talked a lot on the mailing list about general issues about consciousness and the 2-D framework, but I'd like to see us start to concentrate a bit more on the issues re modality, modal arguments, the epistemic/modal bridges, strong necessities, modal space and modal rationalism, and so on (i.e., the issues in "week 6" onward). I'd particularly like to see people getting into the details. It's good to have a broad understanding, but at the end of the day the details are where the cash value is, and there are all sorts of interesting things going on in and around the details here. And I'll be expecting term papers that really go into things carefully. So in the meantime I'd be pleased to see comments on the ins and outs of the arguments, distinctions, and the general dialectic here. That goes both for the material I've presented and for the material in the readings. The mailing list might be a particularly good place to go into these things, as it's arguably easier to go into the details carefully outside real time. It might be good to meet twice again the week after spring break, if we can (week 9 is very light on reading, so it would make sense to schedule a double-whammy then). I go out of town on Thursday night, but Thursday at 3 would probably work for me. Let me know if it doesn't work for any of you. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Fri Mar 19 13:10:27 1999 Date: Fri, 19 Mar 1999 14:09:09 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Comments on Yablo To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Some comments on Yablo, (1) I have some worries about the reliability of conceivability that Yablo either doesn't mention, or doesn't take as seriously as he should. The first is the fact that one's modal intuitions can be influenced by how the secenario that one is being asked to imagine is described. The locus classicus of this is Bernard William's wonderful paper *The Self and the Future*. In the first scenario, information from my brain is extracted and placed in your brain, and vice versa. Most people have a very strong intuition that my identity goes with my information, and thus the continuity of one's body is not sufficient for one's continued existence. In the second scenario, I am told that I will be tortured tomorrow, but not to worry, for I will be given a drug before hand such that all of my memories are erased. Most people have an equally strong intuition that I should worry, and thus they think that continuity of one's body is sufficient for one's continued existence. (I'm glossing over a lot of the details here, hopefully most will be familiar with the paper.) The moral that at least some have drawn from this is that, with a few exceptions, thought-experiments (or thought-experiments that go beyond what we know to be nomologically possible) are unreliable. (Kathleen Wilkes says *something* like this, as does Mark Johnston.) There are a number of things that Yablo might say about this case. His first line of response (p. 38) doesn't look promising - nobody is conflating metaphysical possibility or conceivability with epistemic possibility here. He might want to push the second line (p. 39), for he says that 'it is all too easy to believe that much of the current controversy over conditions of personal identity and survival. . .owes more to our meaning slightly different things by "person" and survive" than to any real clash of modal intuitions" (p. 39) But this doesn't look like it will work here, for it is one and the same person who can conceive of a scenario that supports P, and can also conceive of a scenario that also supports not-P. Perhaps what he should say is something like the following: whether one's ability to conceive of P provides good reason for thinking that P is (meaphysically) possible depends on the situation/background against which one is imagining P (see p. 28). The difference between the two scenarios is the context in which the P fact is imagined. To use an analogy, the difference is something like that between imaging a tiger in a zoo and imaging a tiger in a wild, and finding that in the first situation one can imagine the tiger to be polka-doted, but in the second situation one cannot. And this of course is perplexing, because whether or not tigers can be polka-doted should not depend on their immediate environment. I take it that Yablo needs to argue that the imagined context in Williams's scenario's is not as benign as putting one's imagined tiger in the wild or in the zoo - he needs to argue that it actually changes what it is that one is imagining. Here's another example. it would be a bad result for perception if two lines drawn on a piece of paper looked parallel during the week, but appeared to curve on weekends. But it's not so bad when two lines that appear to be parallel when viewed by themselves, seem to curve when embedded in a larger picture of other lines that curve. Then we have grounds for explaing the latter case as one of perceptual illusion. Yablo needs to explain why the context of one (or both) of Williams's scenarios creates a modal illusion. Of course, he may well be able to do that, but no-one seems to have had much succcess at this task. (2) A second worry that I have is one that I've bought up before. It's the problem of inter-subjective disagreement concerning essentialism about one's own identity. Yablo several times refers to the Kripkean intuition that one's origins are essential to one. Some (Aristotelians?) some to think that although one's origins are not essential to one, one's species membership is. Others (an undergrad professor of mine) think that neither one's origins nor one's species membership are essential to one. He thought that he could conceive of himself as being a poached egg (an ordinary poached egg, not an eggy creature that walks and talks). Assuming some form of modal factualism, on which at most one of these positions is correct, how do we decide which one is? Again, it seems to me that neither (1) nor (2) (see p. 39) will work here. Will (3) work? Are there defeaters for one or more of these positions? Perhaps this is where the argument would turn. Perhaps the Kripkean would claim that I couldn't have been a poached egg, for if I could, then you might just as easily have been the same poached egg, but, necessarily, I couldn't have been you. But what is one's reason for thinking that the last claim is true? That one cannot imagine a situation in which I am you? But that seems to be just what is in question. Yablo could also argue that the disagreement between the Kripkeans, Aristotelians and the poached egg people is contaminated by theoretical commitments, but again, this would have to be argued. Now, of course, there are lots of modal problems that (almost) everyone agrees on, e.g., I could have had something different for breakfast this morning. So the argument cannot be that people fail to exhibit agreement on a lot of modal statements. But to say this isn't to say much. The problem seems to be that there are lots of important modal statements that people don't agree on, and it's not clear that they are making some non-modal factual error or logical error that would explain their modal disagrement (p. 39). The situation here seems to be akin to that of religious discourse, and to some extent moral discourse. Most people agree on most moral judgments, but that doesn't make the job of trying to understand why there is a residue of basic and fundamental moral disagreement any easier. (3). Why does Yablo think that 'it is inconceivable that addition facts should vary between possible worlds'? (p. 32) Presumably it's not because of some inductive generalization. It's not as if he's tried to conceive of <2+3=5> in one world, and <2+3 not=5> in some other world and failed, tried to imagine <3+3=6> in one world and <3+3 not=6> in some other world and failed, and so on, and thus induced that all arithmetical statements have their truth-values necessarily. Has he directly tried to imagine all arithmatical truths having their truth-values contingently? How would one do that? I have some sense of what is involved in entertaining the thought , but I have no idea of how to submit this thought to imaginative scrutiny. 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 Tue Mar 23 22:42:43 1999 Date: Tue, 23 Mar 1999 22:41:40 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Admin To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I e-mailed Alvin about using the library tomorrow. He was planning on using it 2-4, and suggested we split the difference and meet at 3:45. So that can be our tentative plan. It occurred to me also that we didn't consider the possibility of meeting earlier, e.g. at 2 or even at 12 (that way maybe Thony could make it). It probably won't be practical to make such a change even if theoretically possible, but just in case, could everyone e-mail me immediately after getting this and tell me whether an earlier time would work? I'll be interested to hear everyone's thought on the issues about conceivability and possibility over e-mail. There are lots of interesting open issues here to discuss -- e.g. concerning (1) the definition of different kinds of conceivability (ideal conceivability, positive conceivability, etc); (2) the gap between prima facie and ideal conceivability, and our epistemic access to the latter; (3) the issue of how good a guide secunda facie conceivability is to possibility, and whether there are some interesting counterexamples (I'm very interested to hear of any!); (4) the difference between negative and positive conceivability and whether there is anything in the "twilight zone" between these (i.e., ideally negatively conceivable but not ideally positively conceivable); and so on. All thoughts on these and other relevant issues are welcome! I should remind people that this is just like any other class where writing is due every week; not handing in work is a black mark. Now that we have real meetings, I'll relax the requirement that everyone should post at least one non-reply contribution per week. At least one reasonably substantial contribution is expected per week, but it can be a reply to or an extension of someone else's contribution. The only reading for tomorrow are sections 2.5 and 2.6 of Mind and Modality (which hopefully you've read already). For next week, the readings are TCM, Section 2.5 Block & Stalnaker, Conceptual analysis and the explanatory gap [web] Byrne, Cosmic hermeneutics. Horgan, Supervenience and cosmic hermeneutics Block and Stalnaker is available on the web via my "online papers on consciousness" page. I have put Byrne and Horgan in the purple folder in the department office; make yourself a copy. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Fri Mar 19 13:10:27 1999 Date: Fri, 19 Mar 1999 14:09:09 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Comments on Yablo To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Some comments on Yablo, (1) I have some worries about the reliability of conceivability that Yablo either doesn't mention, or doesn't take as seriously as he should. The first is the fact that one's modal intuitions can be influenced by how the secenario that one is being asked to imagine is described. The locus classicus of this is Bernard William's wonderful paper *The Self and the Future*. In the first scenario, information from my brain is extracted and placed in your brain, and vice versa. Most people have a very strong intuition that my identity goes with my information, and thus the continuity of one's body is not sufficient for one's continued existence. In the second scenario, I am told that I will be tortured tomorrow, but not to worry, for I will be given a drug before hand such that all of my memories are erased. Most people have an equally strong intuition that I should worry, and thus they think that continuity of one's body is sufficient for one's continued existence. (I'm glossing over a lot of the details here, hopefully most will be familiar with the paper.) The moral that at least some have drawn from this is that, with a few exceptions, thought-experiments (or thought-experiments that go beyond what we know to be nomologically possible) are unreliable. (Kathleen Wilkes says *something* like this, as does Mark Johnston.) There are a number of things that Yablo might say about this case. His first line of response (p. 38) doesn't look promising - nobody is conflating metaphysical possibility or conceivability with epistemic possibility here. He might want to push the second line (p. 39), for he says that 'it is all too easy to believe that much of the current controversy over conditions of personal identity and survival. . .owes more to our meaning slightly different things by "person" and survive" than to any real clash of modal intuitions" (p. 39) But this doesn't look like it will work here, for it is one and the same person who can conceive of a scenario that supports P, and can also conceive of a scenario that also supports not-P. Perhaps what he should say is something like the following: whether one's ability to conceive of P provides good reason for thinking that P is (meaphysically) possible depends on the situation/background against which one is imagining P (see p. 28). The difference between the two scenarios is the context in which the P fact is imagined. To use an analogy, the difference is something like that between imaging a tiger in a zoo and imaging a tiger in a wild, and finding that in the first situation one can imagine the tiger to be polka-doted, but in the second situation one cannot. And this of course is perplexing, because whether or not tigers can be polka-doted should not depend on their immediate environment. I take it that Yablo needs to argue that the imagined context in Williams's scenario's is not as benign as putting one's imagined tiger in the wild or in the zoo - he needs to argue that it actually changes what it is that one is imagining. Here's another example. it would be a bad result for perception if two lines drawn on a piece of paper looked parallel during the week, but appeared to curve on weekends. But it's not so bad when two lines that appear to be parallel when viewed by themselves, seem to curve when embedded in a larger picture of other lines that curve. Then we have grounds for explaing the latter case as one of perceptual illusion. Yablo needs to explain why the context of one (or both) of Williams's scenarios creates a modal illusion. Of course, he may well be able to do that, but no-one seems to have had much succcess at this task. (2) A second worry that I have is one that I've bought up before. It's the problem of inter-subjective disagreement concerning essentialism about one's own identity. Yablo several times refers to the Kripkean intuition that one's origins are essential to one. Some (Aristotelians?) some to think that although one's origins are not essential to one, one's species membership is. Others (an undergrad professor of mine) think that neither one's origins nor one's species membership are essential to one. He thought that he could conceive of himself as being a poached egg (an ordinary poached egg, not an eggy creature that walks and talks). Assuming some form of modal factualism, on which at most one of these positions is correct, how do we decide which one is? Again, it seems to me that neither (1) nor (2) (see p. 39) will work here. Will (3) work? Are there defeaters for one or more of these positions? Perhaps this is where the argument would turn. Perhaps the Kripkean would claim that I couldn't have been a poached egg, for if I could, then you might just as easily have been the same poached egg, but, necessarily, I couldn't have been you. But what is one's reason for thinking that the last claim is true? That one cannot imagine a situation in which I am you? But that seems to be just what is in question. Yablo could also argue that the disagreement between the Kripkeans, Aristotelians and the poached egg people is contaminated by theoretical commitments, but again, this would have to be argued. Now, of course, there are lots of modal problems that (almost) everyone agrees on, e.g., I could have had something different for breakfast this morning. So the argument cannot be that people fail to exhibit agreement on a lot of modal statements. But to say this isn't to say much. The problem seems to be that there are lots of important modal statements that people don't agree on, and it's not clear that they are making some non-modal factual error or logical error that would explain their modal disagrement (p. 39). The situation here seems to be akin to that of religious discourse, and to some extent moral discourse. Most people agree on most moral judgments, but that doesn't make the job of trying to understand why there is a residue of basic and fundamental moral disagreement any easier. (3). Why does Yablo think that 'it is inconceivable that addition facts should vary between possible worlds'? (p. 32) Presumably it's not because of some inductive generalization. It's not as if he's tried to conceive of <2+3=5> in one world, and <2+3 not=5> in some other world and failed, tried to imagine <3+3=6> in one world and <3+3 not=6> in some other world and failed, and so on, and thus induced that all arithmetical statements have their truth-values necessarily. Has he directly tried to imagine all arithmatical truths having their truth-values contingently? How would one do that? I have some sense of what is involved in entertaining the thought , but I have no idea of how to submit this thought to imaginative scrutiny. 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 Tue Mar 23 23:32:48 1999 Date: Tue, 23 Mar 1999 23:32:36 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: Comments on Yablo To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Some brief thoughts re Tim's thoughts on Yablo. All other comments are welcome. >(1) I have some worries about the reliability of conceivability that Yablo >either doesn't mention, or doesn't take as seriously as he should. The >first is the fact that one's modal intuitions can be influenced by how the >secenario that one is being asked to imagine is described. The locus >classicus of this is Bernard William's wonderful paper *The Self and the >Future*. In the first scenario, information from my brain is extracted and >placed in your brain, and vice versa. Most people have a very strong >intuition that my identity goes with my information, and thus the >continuity of one's body is not sufficient for one's continued existence. >In the second scenario, I am told that I will be tortured tomorrow, but >not to worry, for I will be given a drug before hand such that all of my >memories are erased. Most people have an equally strong intuition that I >should worry, and thus they think that continuity of one's body is >sufficient for one's continued existence. (I'm glossing over a lot of the >details here, hopefully most will be familiar with the paper.) This is interesting. Of course not everyone has both intuitions. I'd be surprised if "most people" have the first intuition, for example. I think a lot of people intuitively think that identity goes with the brain, so in the first case I won't survive in the new brain. And the strength of the second intuition is unclear. But I suppose that *if* one has both intuitions, then one seems to be suggesting that continuity of info is sufficient for survival, and that continuity of brain/body is sufficient for survival. As it stands, those intuitions aren't quite contradictory -- maybe one can survive either way? Though it would seem to commit one to the possibility of double survival in certain cases, which seems odd (unless one has a rule for deciding when both criteria are present in separate cases). But anyway, let's ignore the non-contradictoriness and think about what to say if we really did have two contradictory intuitions here: i.e. we find A positively conceivable, B positively conceivable, where the possibility of A contradicts the possibility of B. I think in such ases, one can argue that one is being inconsistent in one's evaluation of the conceivable scenarios, and that further rational reflection will reveal this, so both won't turn out to be ideally positively conceivable. If the A is truly (i.e. ideally) positively conceivable, then analysis of this will reveal that B is not, and vice versa. And indeed I think this mirrors our reaction to the cases. Even if one has a momentary feeling that both intuitions are reasonable, a bit of reflection reveals the conflict and leaves us at best uncertain, or perhaps withdrawing one of the claims completely. Of course that's just the case where both intuitions are had in one person. Where conflicting intuitions re personal identity are had by two different people, it's tricky. Just say A says he can conceive of teletransportation being survival and B says he conceives of it being death. I suppose that's not automatically a contradiction, as maybe they are conceiving different "further facts" in addition to the micro facts: if one is a nonreductionist about PI one might well think that both survival and death are positively conceivable here. So let's take a strong case, where A says teletransportation survival is not conceivable, and B says that it is; and B says that teletransportation death is not conceivable, where A says that it is. These intuitions are clearly in conflict. One possibility is that A and B have different concepts of "survival" and "death": perhaps B has a deflationary concept of "functional survival", whereas A has an inflationary concept of "intrinsic survival", for example. In that case, we might find that on disambiguation, the two agree on the relevant intuitions (teletransportation is always functional survival, never intrinsic survival). Though it may be that B will say that intrinsic survival does not exist; in that case A and B will differ not on what's conceivable but on what's actual. Another possibility is that both are to some degree irrational, or at least are making claims that are infected by theory or wishful thinking as much as direct conceivability. For my part, for example, I think that both teleport survival and teleport death are prima facie negatively conceivable, and I don't have strong intuitions about how best to describe the positively conceivable teleport scenario. So maybe A and B should best be agnostic, though they may have reasonably strong intuitions here. Or maybe it's a place where continued rational reflection can reveal that one is right and the other wrong. Tim also raises the case where P is conceivable against one background but not against the other. In such a case I'm tempted to say that what's here at issue is prima facie conceivability, and that even secunda facie reflection indicates that one ought to treat the two cases equally. Maybe background can bias one in a certain direction prima facie, but this biasing role should be factored out on ideal reflection. Tim's case of conflict over essentialism is tricky. As noted in class, this is a case of secondary conceivability, i.e. intuitions about what is secondarily possible (conceivable/possible in worlds considered as counterfactual). That's often a posteriori, but arguably should be a priori conditional on the empirical non-modal facts being all in or not at issue. So what's the a priori disagreement here? Different concepts of person or object? Maybe, but doesn't seem too promising. Insufficient rational reflection? I think Kripke would say your teacher is just wrong, and not being properly rational, but of course it's not clear how to adjudicate that. Vagueness/indeterminacy/ambiguity? I'm tempted to say that the whole question of what is essential to X is somewhat vague, and I can often go that way. So maybe the fact of the matter is indeterminate, and the two are "precisifying" in different ways. I'd like to say that insofar as the matter is determinate, it's settlable by rational reflection, and vice versa. So any irresolvable rational dispute (given that the empirical facts are in) corresponds to an indeterminacy. That's more or less the scrutability claim we'll be discussing, and it's somewhat controversial, but not wholly implausible, I think. This also touches on the question of disagreement over a priori truths by apparently rational being. E.g., Graham Priest, a very smart philosopher, thinks contradictions are sometimes true. Most others think it is a priori (and obvious) that they never are. Is he being irrational? Well, he at least seem sane and a good reasoner. But I guess one has to say that at some level he is not being rational, even though he is to some extent "procedurally rational". Maybe some conceivability disagreements are analogous to this, where someone says something that is a priori false despite seeming procedurally rational. Maybe even the poached egg case, at a deep level? So there are various resources for somene who endorses a conceivability-possibility link, but clearly none of this is obvious or trivial. All thoughts are welcome. --Dave. P.S. Re mathematics, I think Yablo thinks it is a priori that 3+3=6, so it isn't conceivable otherwise. If P is a priori, not-P is not negatively conceivable or positively conceivable. In any case I take it that there's no positive reason to think it conceivable that 3+3 is not 6; and the fact that it is a priori suggests that any scenario will verify "3+3=6". (Given that arbitrary a priori reasoning is allowed in seeing what verifies what.) From owner-modality@LISTSERV.ARIZONA.EDU Wed Mar 24 07:01:31 1999 Date: Wed, 24 Mar 1999 07:59:10 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Re: Comments on Yablo To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO > > P.S. Re mathematics, I think Yablo thinks it is a priori that 3+3=6, > so it isn't conceivable otherwise. If P is a priori, not-P is not > negatively conceivable or positively conceivable. In any case I take > it that there's no positive reason to think it conceivable that 3+3 is > not 6; and the fact that it is a priori suggests that any scenario > will verify "3+3=6". (Given that arbitrary a priori reasoning is > allowed in seeing what verifies what.) This seems to me to get things around the wrong way. We think that 3+3=6 is a priori *because* its negation is inconceivable, not vice-versa. This is a bit clearer with respect to geometrical truths, it seems to me. Why did Kant think that Euclidian geometry was a priori true of this world? (It's not as if it comes with a label stuck to it: 'a priori'.) He tried to conceive (imagine? visualize?) a physical world in which geometry was non-Euclidean, and failed. It seems to me that thinking P is a priori involves attempting to conceive of not-P, and being unable to. (Perhaps that should be: one is justified in thinking that p is a priori if (only if?) one has tried to conceive of not-P and been unable to. 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 Mar 24 09:40:05 1999 Date: Wed, 24 Mar 1999 10:39:06 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Mary and Negative possibility To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Dave, I didn't understand why the Mary argument is meant to involve negative possibility, while Zombies involve positive possibility. When I think about the Mary argument, (I try to) clearly and distinctly conceive of a scenario in which Mary knows all the physical facts, but can't work out the phenomenal facts. She knows all about red, she doesn't know what red is like. In those moments in which I find the argument compelling, I think that I have C and D conceived of just such a scenario. I don't see the asymmetry between the Zombie case and the Mary case. What am I missing here? 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 Mar 24 11:22:20 1999 Date: Wed, 24 Mar 1999 11:21:38 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: Mary and Negative possibility To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Re Mary: It's true that one positively conceives of Mary, knowing all the physical facts but being unable to figure out the phenomenal facts. So Mary's situation is positively conceivable. But the question I was concerned with is whether "P and not-Q" (P = all the physical truths, Q = a phenomenal truth) is positively or negatively conceivable, as it's the conceivability or otherwise of this statement that hooks into the argument against materialism. Positively conceiving Mary isn't positively conceiving of a situation where P and not Q. Rather, the Mary scenario gives one good reason to believe that Q is not a priori deducible from P. That is, "P -> Q" is not a priori, i.e. "P and not-Q" is negatively conceivable. I.e., what the Mary scenario suggests is that Mary can't a priori rule out the epistemic possibility that P is true and Q is false. That's not yet to say that such a scenario is positively conceivable. It may well be, and personally I think it is (e.g. via conceiving of zombies and inverted spectra), but that takes further argument, somewhat akin to the conceivability argument itself. Re mathematics, Tim suggests interestingly that we say "3+3=6" is a priori precisely because we can't conceive otherwise. One subtlety here: presumably lots of complex mathematical statements M are such that we can't conceive otherwise, i.e. we can't positively conceive that not-M, but only because we can't positively conceive of either M or not-M. So perhaps the claim about "3+3=6" should rather be that in every situation of which we can positively conceive, "3+3=6" is verified. And the suggestion is that it's because "3+3=6" has this property that we say it's a priori. Personally, I'm not certain of this. I agree that if "3+3=6" was *not* verified in some positively conceivable scenario, that would be good reason to say it's not a priori. But I'm not sure that things work the other way. Maybe there can be support for apriority which is not grounded in trying to positively conceive otherwise. E.g., a mathematical proof seems to give support for a priority that doesn't come from trying to positively conceive otherwise. Of course one can still worry about the status of the axioms and inference rules here -- maybe those are so grounded? But this is far from obvious. One worry is that if this were the only support for an a priority claim, such a claim would always be open to the objection that we haven't tried hard enough to conceive of relevant scenarios in which P is false, or that we're suffering from lack of imagination. Maybe that is a reasonable strategy for some apriority claims (e.g. Euclid's), but it's not clear that all apriority claims are vulnerable to this sort of objection. In particular, I think we have grounds for saying that "3+3=6" is a priori without relying solely on our inability to conceive otherwise. Maybe Tim is right that truth of P in all positively conceivable scenarios gives one reason to think P a priori. As such, it might only be defeasible reason, though. E.g. maybe Euclid did have good though defeasible reason to think Euclidean geometry a priori. Of course this eventually got defeated both by people positively conceiving of other geometries and by showing the formal consistency of other principles. It's not obvious to me that support for the a priority of all truths has this form, but it's an interesting claim. Of course the fact that not-P is not *negatively* conceivable would be very good reason to say that P is a priori. But that's pretty trivial: to say that not-P is not negatively conceivable is to say that we can a priori rule out the epistemic possibility that not-P, which is more or less to say that we can a priori rule in P. But it could be the case that P is true in all positively conceivable scenarios, while not-P is still negatively conceivable. E.g. because (1) we can't rule out the possibility of positively conceivable scenarios where P fails, or (2) maybe P is an odd sort of truth that is verified in *all* ideally positively conceivable scenarios, but still is not a priori. (My favorite example of the latter is "There are inconceivable features of the world.) --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Wed Mar 24 12:56:08 1999 Date: Wed, 24 Mar 1999 13:54:16 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Josh Cowley Subject: Ideal conceivability To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I'm still having problems with the notion of ideal conceivability (IC). A couple of the suggested definitions of IC are, "conceivable on ideal rational reflection," "conceivable after all possible reasoning is done," and perhaps "will remain conceivable given any more reasoning." Roughly my worry is that IC winds up being trivially identical with logical possibility. Here is how the argument goes. Being a non-ideal rational agent I don't know how to define ideal rational reflection except as something like, (A) "given some initial premises or concepts, reasoning which results in all and only logically possible conclusions." I don't mean to say that we can't say anything about ideal rationality. We can can give descriptions of what rationality for us is. And the more detailed this description gets the closer we will be to describing ideal rationality. We can even suggest the removal of some of our cognitive constraints such as the lact of time. But having limited rationality places limits on what we can ultimately say about ideal rationality. What we have to go to in the end is that ideal rationality is whatever achieves (A). Is there anything wrong with this? It looks like it is. The current debate is whether conceivability implies or is a guide to logical possibility. I take it our argument is that ideal conceivability is a guide to logical possiblity. Now we need only show that secunda facie conceivability is a good guide to ideal conceivability. But if ideal conceivability is not merely a guide, but is just defined by logical possiblity, the all the arguments that the conceivable doesn't imply the possible can be used to say that the secunda facie conceivable doesn't imply the IC. What is needed is some way of defining IC that doesn't ultimately rest on logical possiblity. From owner-modality@LISTSERV.ARIZONA.EDU Mon Mar 15 23:35:08 1999 Date: Tue, 16 Mar 1999 00:33:42 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik A Herman Subject: Re: Papers To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I have a hypothetical scenario to propose regarding the bounds of ideal a prioricity (consider it actual if you have a difficult time believing it). (joke) Consider an apriori genius, a true "head-case", that spends most of her time in deep meditation and spends little time gathering empirical information. My intuition is that she might eventually come to be able to deduce the physical structure of her brain. After all, all the fundamental physical stuff: space, time, and matter, is right there in her head-- so it might very well be retrievable introspectively (especially in her state of being relatively clear of empirical clutter). In short, my point is that even in meditating with respect only to a priori/ a priori deducible concepts, we might have access to physical truths as we do mathematical ones; my worry is that this "objective viewing of the mind" looks a posteriori minus the senses. My questions: Does the head-case case seem like a logically possible case of ideal a prioricity? If so, how would the a priori theorist deal with my concerns? For instance, what is fundamentally different from her internal monitoring of her brain, than if she stared at her hand? Are BOTH cases a posteriori? Erik H. From owner-modality@LISTSERV.ARIZONA.EDU Thu Mar 25 10:40:06 1999 x-sender: agillies@pop.u.arizona.edu Date: Thu, 25 Mar 1999 11:53:57 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony S Gillies Subject: positive vs. negative conceivability To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO All, I've been thinking about some ways of making the positive/negative conceivability (PC/NC) distinction clearer. Here are some thoughts on the matter. First, NC is a bit trickier than we have been taking it to be. The official line is that something is NC iff there is no apparent contradiction in it. There aren't contradictions in *things*, so we have to be clearer about what we mean. One option is to say that a state of affairs S is NC iff the statement expressing S is apparently satisfiable. The trouble is that satisfiablitity requires us to picture a scenario (clearly and distinctly) which makes the statement expressing S come out true. But this is just what PC was supposed to be. So the satisfiability route won't work. The reason it won't work is instructive, though: any way of cashing out NC which is semantical will likely run into the same problem. Maybe we can think of NC along *formal* lines. Say that a statement P is NC iff P is not apparently a contradiction, where "contradiction" is understood purely formally (as a matter of syntax, if you like). This won't quite do either: we need to add a patch to the effect that P is not apparently a contradiction, nor is it apparent that on the assumption of only P we can derive a contradiction. But the patch is in the same spirit: derivability is not a semantical consideration. As for PC, we can treat that as a thorough-going semantical notion. A statement P is PC iff P is apparently satisfiable. Saying that P is (apparently) satisfiable requires us to picture a world in which P comes out as true. And this is what we want out of PC. Drawing the lines in this way has some nice features. For one thing, it explains why NC seems easier to do. It's pretty hard to reason formally without relying on semantic considerations, and so it's easy to overlook formal contradictions (especially when they are derived somewhere down the line). In the Goldbach case, for instance, it is easier to pump intuitions to the effect that it is NC either way. But you can't pump any PC intuitions at all here: they aren't both satisfiable, not even apparently. Also, PC and NC can still be combined with the other dimensions of conceivability we talked about the other night. Cheers, Thony "Curious green ideas sleep furiously." From owner-modality@LISTSERV.ARIZONA.EDU Thu Mar 25 13:18:29 1999 Date: Thu, 25 Mar 1999 13:09:57 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: positive and negative conceivability, ideal conceivability To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Re Thony's comments on positive and negative conceivability. The first thing to note is that as defined, the various sorts of conceivability are all predicates of *statements* S. So S is negatively conceivable if there is no apparent contradiction in the hypothesis that S, or if one cannot rule out S a priori. Then one gets prima facie or ideal negative conceivability depending on whether one goes with surface reasoning or ideal reasoning. I think it's right to say that this is a semantic notion. Syntactic considerations re formal logic don't really play a part here. Even the talk of "contradiction" is not really essential; it's the inability to rule S out a priori that's central. Re positive conceivability, "apparently satisfiable" is not a bad way to get at this, but of course that will come down to what "satisfiable" means here. If satisfiable means "possible", this comes down to something like van Cleve's definition of strong conceivability. One trouble there is that we want type-B materialists to come out saying zombies are conceivable but not possible; but on this definition, it's not clear that zombies will come out conceivable for a type-B materialist, as they will judge that zombies are impossible. Maybe there's a sense in which zombies are still "apparently possible" for the type-B materialist, but it's tricky to bring it out. That's why I go with clear and distinct conceptions instead (which is maybe more or less what Thony means by "picturing a world"). Re Josh's comment on ideal conceivability, I talked about this to some extent in our meeting yesterday. It's certainly true that defining "ideal reasoning" here as reasoning that tracks possibility will not be helpful, as it will render the conceivability/possibility thesis trivial. So we need some substantive notion of ideal reasoning. Or at least, we need a substantive notion of better and worse reasoning. What's really needed is some kind of partial ordering on the space of reasoners, such that for a pair X and Y, it can be the case that X is "less limited" than Y, because X does not suffer from some cognitive limitation that Y suffers from. E.g., X has more memory or attention capacity than Y, or more concepts, or makes better inferences, and so on. In addition to this, we also need the notion of "P is conceivable for X" (a sort of prima facie notion). Putting these two things together, we can say that P is ideally conceivable if there is some X such that P is conceivable for X, and for all Y that is less limited than X, P is conceivable for Y. This does require the notion of a cognitive limitation and the corresponding partial ordering. I think we at least have a good intuitive grasp on this notion, though I don't claim it's 100% well defined. And I don't think it's defined in terms of logical possibility. Still, maybe Josh's point that our own limitations limit what we can say about ideal rationality can be transferred here to the point that our limitations limit what we can say about cognitive limitations! (Maybe there are cognitive limitations we're unaware of, for example.) I'm not sure how much of a problem that is. As long as we have a good tacit grasp of the notion of a cognitive limitation, the notion of ideal conceivability will be well-defined; what's required for the definition is that for X and Y, our statement "X is more limited than Y" can be true or false, irrespective of whether we can determine the truth or falsity. Of course this puts certain limits on our epistemic access to ideal conceivability, but we knew that already. I suppose one might object that our notion of a cognitive limitation is so vague that for particularly advanced X and Y, our statement "X is more limited than Y" is neither true nor false, even though in some sense it "should" be true or false. I'm not sure exactly how coherent that is. At least, it seems that we can here defer to our smarter counterparts who can make the determination, so the truth of our statements of cognitive limitations is to some extent parasitic on their judgments. Or maybe one can just say that the very fact that the claim in question "should" be true or false suggests that our claim is true or false, it's just that we don't know it. (Like saying that certain mathematical statements are true/false even though we don't know it.) A tricky issue. The strategy mentioned above also helps with the potential circularity problem. We want to connect possibility to ideal conceivability, but ideal conceiv*ability* is defined in terms of possibility (what various possible smarter beings can conceive). The conc-poss claim wasn't really intended as a reductive definition of possiblity, so I'm not sure how much of a problem that is, but in any case, I'd like to think that any circle here is informative rather than vicious. Even if we defined ideal conceivability in terms of what certain conceivable beings could conceive, thus might be nontrivial, as long as we have facts about what we can conceive to get things off the ground. We can conceive of beings who lack some of our cognitive limitations, and they will be able to conceive of things that we can't conceive, among which things will be beings smarter than them, who will be able to conceive of things that they can't conceive, and so on. So maybe the circle here will be "expanding" and will eventually settle on a useful substantive notion of ideal conceivability. At least, it would be nice if it did! But I can't claim that anything here is crystal clear. All thoughts are welcome. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Fri Mar 26 13:59:11 1999 Date: Fri, 26 Mar 1999 14:58:00 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony T Lane Subject: weakly conceivable To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Van Cleve suggests that a proposition P is *weakly conceivable* for S iff S does not "see" that P is impossible. I realize that it is perhaps undesirabe to include possibility in ones definition of conceivability. Nevertheless, I thought that weak conceivablity sounds rather like negative conceivability. Van Cleve dismisses weak conceivability as a guide to possibility because of the Goldbach example: since we seem to be unable to see either that Goldbach's Conjecture is true or that it is false, we do not see either G or not-G as impossible, Thus, since we can weakly cnceive of both and they are contradictory, weak conceivability cannot be a guide to possibility. I do not realy see how negative conceivability is different weak conceivability. We could introduce the notion of ideal weak conceivability, which would presumably rule out the contradiction impied by the GC case-- presumably there is an ideal conceiver who can see that either G or not-G is impossible. Or, if by "seeing" we mean "clear and distinct perception", it does not seem clear that we can weakly conceive either G or not-G. I think the point I am trying to make here is somewhat similar to Josh's-- it seems that, even thonugh the definitions of the various types of conceivability do not explicitly refer to possibility, the notion of possibility is in some sense required to make sense of conceivability. Anthony From owner-modality@LISTSERV.ARIZONA.EDU Sun Mar 28 13:10:00 1999 Date: Sun, 28 Mar 1999 13:09:48 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: weakly conceivable To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Anthony writes: >Van Cleve suggests that a proposition P is *weakly conceivable* for S iff >S does not "see" that P is impossible. I realize that it is perhaps >undesirabe to include possibility in ones definition of conceivability. >Nevertheless, I thought that weak conceivablity sounds rather like >negative conceivability. You're right that negative conceivability is a lot like van Cleve's weak conceivability (and positive conceivability is a lot like can Cleve's strong conceivability). I prefer to define it without making direct reference to possibility and impossibility: P is prima facie negatively conceivable if one cannot rule out P a priori, and P is ideally negatively conceivable if not-P is not a priori. van Cleve's notion is obviously similar to this, at least if we restrict to the domain of primary possibility and conceivability. For P to be prima facie negatively conceivable (for S) is for S to rule P out a priori, which is for S to judge a priori that the actual world is a world where P is false, which is for S to judge a priori that all worlds considered as actual are worlds where P is false, which for S to judge a priori that P is 1-impossible, which is almost for P to be weakly conceivable for S. The main loophole is the chance that one might see that P is 1-impossible without judging this a priori; if so, P might be negatively but not weakly conceivable. That sort of loophole is common for 2-possibility, but rare for 1-possibility. >Van Cleve dismisses weak conceivability as a >guide to possibility because of the Goldbach example: since we seem to be >unable to see either that Goldbach's Conjecture is true or that it is >false, we do not see either G or not-G as impossible, Thus, since we can >weakly cnceive of both and they are contradictory, weak conceivability >cannot be a guide to possibility. > >I do not realy see how negative conceivability is different weak >conceivability. We could introduce the notion of ideal weak >conceivability, which would presumably rule out the contradiction impied >by the GC case-- presumably there is an ideal conceiver who can see that >either G or not-G is impossible. Or, if by "seeing" we mean "clear and >distinct perception", it does not seem clear that we can weakly conceive >either G or not-G. Right, weak conceivability isn't a good guide to possibility for the same reason that prima facie negative conceivability isn't (e.g. the cases you discuss above). Your two suggested corrections to the notion correspond to ideal negative conceivability and to prima facie positive conceivability respectively, both of which are much better guides to possibility. >I think the point I am trying to make here is somewhat similar to Josh's-- >it seems that, even thonugh the definitions of the various types of >conceivability do not explicitly refer to possibility, the notion of >possibility is in some sense required to make sense of conceivability. Well, I'd like to think my way of defining things doesn't import the modal notion as directly as van Cleve's. As usual, though, modal notions may be playing a subtle role, especially in the idealized version (which is in effect about what *can* be known a priori by some possible being). --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Sun Mar 28 21:38:53 1999 Date: Sun, 28 Mar 1999 22:37:34 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Angela J Burnette Subject: Re: positive and negative conceivability, ideal conceivability To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Regarding Josh's comments about ideal conceivability and Dave's responses, there seems to be some slippage between thinking of the ideal conceiver as being a better reasoner in terms of further reasoning/cognitive powers and being a better reasoner in terms of different reasoning/ cognitive powers...Sometimes it has been asserted that the ideal conceiver is merely faster, older, and the like, but basically of the same kind, and at other times the ideal conceiver has had capabilities predicated of it that would seem to suggest that its powers of conceivability are of a different kind altogether, for example Brad's speculation that the ideal conceiver could possess the concepts of what it's like to be a bat(shouldn't such phenomenal concepts be instrinsic to the bat by definition?), or the idea that the ideal conceiver could be aware of its own cognitive limitations (it seems like it could be aware of ours, but not its own if it merely possesses further reasoning powers, beings that are like us should be cognitively closed to their own cog. limitations)... it seems to me that we can only legitimately abstract away from our own case if we are considering further reasoning which is basically of the same kind as our own...i.e. something like faster reasoning by a being with infinite time and memory... but if we are considering a conceiver who is better in virtue of being a different kind of reasoner, then I'm not sure we are warranted in our abstractions...an additional worry for me is that in order to get the move from conceivability to possibility off the ground we have to posit a conceiver who would not need to make that jump...i.e., it is unclear to me that something like what Josh has argued for isn't the case because positing the ideal conceiver is like collapsing the explanatory gap rather than bridging it...So, I would like some clarity on either why nothing said of the ideal conceiver makes him of a different kind than us, or, why this isn't a problem for abstracting away from our own case... angela From owner-modality@LISTSERV.ARIZONA.EDU Sun Mar 28 22:48:20 1999 Date: Sun, 28 Mar 1999 23:45:35 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik J Larson Subject: Re: ideal conception To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I have a concern similar to Angela's re the ideal conceiver. We need a notion of ideal rational reflection, but it seems that it must be of a kind that we can plausibly extend from our own cognitive powers. This would seem to preclude an ideal conceiver capable of understanding its own cognitive limitations or anything as miraculous as that. We need a notion of "rational reflection" that is not different in kind than what we presently have (or any natural extension of that). Given this more modest extension of reasoning powers, however, there should be plausible counterexamples to the thesis from apriori reasoning to necessity. The Godel case, for one. This goes as follows. Given a minimum requirement for rationality--say, that any rational process could be described systematically or as a sequence of steps--we could not have an ideal conceiver solving any number-theoretic statements (Diophantine equations, say) without limit. This would place the notorious Godel formula outside the limits of ideal rational reflection. If, on the other hand, the ideal conceiver could have apriori access to the primary intensions of G statements, we would of course thwart these cases as real counterexamples, but at the expense of rendering the notion of "rational reflection" or more generally "thought" very vaguely defined, and of little help in the task of extending the scope of conceivability outward from our own case, for which, as Angela points out, we should have something of the same "kind". So I think we've got, on the one hand, some potent counterexamples to theses like strong aprioricity and negpos, and on the other we've got a notion of "thought" or "rational reflection" that pulls too far away from our own case. Erik "What our grammarian does is simple enough. He frames his formal reconstruction of K along the grammatically simplest lines he can, compatibly with inclusion of H, plausibility of the predicted inclusion of I, plausibility of the hypothesis of inclusion of J, and plausibility, further, of the exclusion of all sequences which ever actually do bring bizarreness reactions." -- W.V.O. Quine ---------------------- Erik J Larson erikl@U.Arizona.EDU From owner-modality@LISTSERV.ARIZONA.EDU Sun Mar 28 21:38:53 1999 Date: Sun, 28 Mar 1999 22:37:34 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Angela J Burnette Subject: Re: positive and negative conceivability, ideal conceivability To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Regarding Josh's comments about ideal conceivability and Dave's responses, there seems to be some slippage between thinking of the ideal conceiver as being a better reasoner in terms of further reasoning/cognitive powers and being a better reasoner in terms of different reasoning/ cognitive powers...Sometimes it has been asserted that the ideal conceiver is merely faster, older, and the like, but basically of the same kind, and at other times the ideal conceiver has had capabilities predicated of it that would seem to suggest that its powers of conceivability are of a different kind altogether, for example Brad's speculation that the ideal conceiver could possess the concepts of what it's like to be a bat(shouldn't such phenomenal concepts be instrinsic to the bat by definition?), or the idea that the ideal conceiver could be aware of its own cognitive limitations (it seems like it could be aware of ours, but not its own if it merely possesses further reasoning powers, beings that are like us should be cognitively closed to their own cog. limitations)... it seems to me that we can only legitimately abstract away from our own case if we are considering further reasoning which is basically of the same kind as our own...i.e. something like faster reasoning by a being with infinite time and memory... but if we are considering a conceiver who is better in virtue of being a different kind of reasoner, then I'm not sure we are warranted in our abstractions...an additional worry for me is that in order to get the move from conceivability to possibility off the ground we have to posit a conceiver who would not need to make that jump...i.e., it is unclear to me that something like what Josh has argued for isn't the case because positing the ideal conceiver is like collapsing the explanatory gap rather than bridging it...So, I would like some clarity on either why nothing said of the ideal conceiver makes him of a different kind than us, or, why this isn't a problem for abstracting away from our own case... angela From owner-modality@LISTSERV.ARIZONA.EDU Sun Mar 28 22:48:20 1999 Date: Sun, 28 Mar 1999 23:45:35 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik J Larson Subject: Re: ideal conception To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I have a concern similar to Angela's re the ideal conceiver. We need a notion of ideal rational reflection, but it seems that it must be of a kind that we can plausibly extend from our own cognitive powers. This would seem to preclude an ideal conceiver capable of understanding its own cognitive limitations or anything as miraculous as that. We need a notion of "rational reflection" that is not different in kind than what we presently have (or any natural extension of that). Given this more modest extension of reasoning powers, however, there should be plausible counterexamples to the thesis from apriori reasoning to necessity. The Godel case, for one. This goes as follows. Given a minimum requirement for rationality--say, that any rational process could be described systematically or as a sequence of steps--we could not have an ideal conceiver solving any number-theoretic statements (Diophantine equations, say) without limit. This would place the notorious Godel formula outside the limits of ideal rational reflection. If, on the other hand, the ideal conceiver could have apriori access to the primary intensions of G statements, we would of course thwart these cases as real counterexamples, but at the expense of rendering the notion of "rational reflection" or more generally "thought" very vaguely defined, and of little help in the task of extending the scope of conceivability outward from our own case, for which, as Angela points out, we should have something of the same "kind". So I think we've got, on the one hand, some potent counterexamples to theses like strong aprioricity and negpos, and on the other we've got a notion of "thought" or "rational reflection" that pulls too far away from our own case. Erik "What our grammarian does is simple enough. He frames his formal reconstruction of K along the grammatically simplest lines he can, compatibly with inclusion of H, plausibility of the predicted inclusion of I, plausibility of the hypothesis of inclusion of J, and plausibility, further, of the exclusion of all sequences which ever actually do bring bizarreness reactions." -- W.V.O. Quine ---------------------- Erik J Larson erikl@U.Arizona.EDU From owner-modality@LISTSERV.ARIZONA.EDU Mon Mar 29 19:46:51 1999 Date: Mon, 29 Mar 1999 19:43:11 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: ideal conception To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Re Angela's and Erik's comments on ideal conceivability: Angela writes: >Regarding Josh's comments about ideal conceivability and Dave's responses, >there seems to be some slippage between thinking of the ideal >conceiver as being a better reasoner in terms of further >reasoning/cognitive powers and being a better reasoner in terms of >different reasoning/ cognitive powers...Sometimes it has been asserted >that the ideal conceiver is merely faster, older, and the like, but >basically of the same kind, and at other times the ideal conceiver has had >capabilities predicated of it that would seem to suggest that its powers >of conceivability are of a different kind altogether, for example Brad's >speculation that the ideal conceiver could possess the concepts of what >it's like to be a bat(shouldn't such phenomenal concepts be instrinsic to >the bat by definition?), or the idea that the ideal conceiver could be >aware of its own cognitive limitations (it seems like it could be aware >of ours, but not its own if it merely possesses further reasoning >powers, beings that are like us should be cognitively closed to their >own cog. limitations)... Well, it's true that we don't have an exhaustive list of the relevant cognitive limitations. It would be nice if we did, but at least so far, it isn't in sight. Still, what matters is that we have a determinate enough concept of limitation such that there's a fact of the matter about whether A is less limited than B (whether or not we now explicitly include that limitation in any list). The case where B has concepts that A does not seems to be a straightforward case of that. So if B is just like us except that B has further concepts (further phenomenal concepts, say), then it seems reasonable to say that B is less limited than us in the relevant sense. So I'd like to think this isn't an unreasonable idealization. Indeed, in the standard definition of apriority, we say P is a priori if someone *who has the relevant concepts* can know P with experience-independent justification. On this definition it seems reasonable to say that there can be a priori relations among alien phenomenal concepts, even though we don't have those concepts. You raise the interesting issue of whether any non-bat could have a concepts that can describe a bat's experience. This is tricky, but I'm not sure why it should be ruled out by definition. I can conceive of what it is like to be certain other humans (at least in certain respects, e.g. visually) even though I'm not them. And arguably I might conceive roughly of what it's like visually to be a dog, even though I'm not a dog. There's nothing contradictory about the idea that my phenomenal concepts might apply to another being's experience. And if I can do it with dogs, there's no obvious reason why some other being could do it with bats. Problems would arise if it turns out that there are some phenomenal concepts that simply can't be had by other beings. Or perhaps, phenomenal concepts that simply can't be had by sophisticated beings. Just say experience E (perhaps the experience of stupidity?) could not be had by a sufficiently smart being, and further that such a being could not even form a direct concept of that sort of experience. If that were the case, then arguably the idealization here won't help in determining what's a priori or ideally conceivable re E (relatively ideal conceivers won't even have the concept!). That would be awfully tricky; but it's not obvious that such concepts exist. If such concepts exist, they don't raise problems for the inference from ideal conceivability to possibility, but they might raise difficulties for the inference from possibility to ideal conceivability. These are problems in the same vicinity as the "open inconceivabilities" I mentioned last week. I think such inconceivabilities are probably the biggest obstacle to "pure modal rationalism"; but even the "strong modal rationalism" that survives without them is pretty strong. And of course, it is not obvious that such inconceivabilities really exist. Re your second problem above: we are aware of some of our cognitive limitations, so why shouldn't a less limited being also be aware of theirs? And maybe they could even be aware of cognitive limitations they we haven't thought of yet? This is not to say that they can transcend those limitations, of course, any more than we can. In both cases, I'd like to think that we're bringing in idealizations that are of a sort familiar from our own cases. We know some people have concepts that others lack, and we know that we're aware of some cognitive limitations. So I doesn't seem too great a stretch to shift to beings with concepts that we lack, and beings that are aware of cognitive limitations that we haven't thought about. Once we have the latter, we then have the "bootstrapping" possibility of removing those limitations in turn. So I'd like to think that this sort of idealizing can take us a fair way. Exactly where it takes us, of course, is not entirely obvious. >it seems to me that we can only legitimately >abstract away from >our own case if we are considering further reasoning which is basically of >the same kind as our own...i.e. something like faster reasoning by a being >with infinite time and memory... but if >we are considering a conceiver who is better in virtue of being a >different kind of reasoner, then I'm not sure we are warranted in our >abstractions... The case of more concepts is like the cases of time and memory, I hope. Maybe the "second-order" idealizations (the ones we haven't thought of) go beyond this. And maybe that does raise a question about how well we can judge whether something is ideally conceivable or not. Of course there is a legitimate epistemic worry here: ideal conceivability is not epistemically transparent to us. But that's compatible with saying (i) there's a fact of the matter about whether P is ideally conceivable (whether or not we can know it), and (ii) in many cases, we can know whether or not P is ideally conceivable (e.g., we know it's ideally conceivable that something exists, and we know it's not ideally conceivable that 0+0=1). So there's a fair amount we can know, even if things get hazy around the edges. >an additional worry for me is that in order to get the move >from conceivability to possibility off the ground we have to posit a >conceiver who would not need to make that jump...i.e., it is unclear to me >that something like what Josh has argued for isn't the case because >positing the ideal conceiver is like collapsing the explanatory gap rather >than bridging it...So, I would like some clarity on either why nothing >said of the ideal conceiver makes him of a different kind than us, or, why >this isn't a problem for abstracting away from our own case... Hmm. I suppose the idea here is that (on the type-B position) a relatively ideal reasoner would just see that zombies are impossible, and would not be able to (positively or negatively) conceive of zombies? It's not obvious to me why this should be the case. Even the first half isn't obvious, but the second half is particularly unclear to me. What will rule out the clear and distinct conception of zombies for the ideal reasoner? Maybe something substantive in the a priori domain -- but then we just have type-A or type-C materialism, which is compatible with strong modal rationalism. If it's just that the reasoner knows that zombies are impossible, it seems to me that he/she would say, OK, zombies are ideally conceivable, they're just not possible. Presumably the ideal reasoner will be able to form just as clear and distinct a conception as we can, absent some substantive a priori obstacle. So anyway, if this is right, it seems to me that we're not buying the conceivability/possibility link trivially. Someone like a type-B materialist who denies the link will still deny it on this way of putting things. Of course *if* there is a deep a priori contradiction in the notion of a zombie just that a relatively ideal reasoner will find it, then zombies will come out both conceivable and impossible, but that's as it should be. (Maybe this does argue for not defining "positively conceivable" in terms of "seems possible", though. One could (could) argue that for the ideal reasoner, all and only the possible will seem possible, in which case the CP link will be trivial. This claim about the ideal reasoner is not obvious to me, but in any case, one can avoid the trivializing conclusion by going for a definition of positive conceivability that is not cast in terms of possibility. E.g., the clear and distinct conception idea.) Let me know if I've missed some of the point here. Erik writes: >I have a concern similar to Angela's re the ideal conceiver. We need a >notion of ideal rational reflection, but it seems that it must be of a >kind that we can plausibly extend from our own cognitive powers. This >would seem to preclude an ideal conceiver capable of understanding its own >cognitive limitations or anything as miraculous as that. Again, I'm not sure what's miraculous about understanding one's cognitive limitations. Even we can do that! >We need a notion of "rational reflection" that is not different in kind >than what we presently have (or any natural extension of that). Given >this more modest extension of reasoning powers, however, there should be >plausible counterexamples to the thesis from apriori reasoning to >necessity. The Godel case, for one. This goes as follows. Given a >minimum requirement for rationality--say, that any rational process could >be described systematically or as a sequence of steps--we could not have >an ideal conceiver solving any number-theoretic statements (Diophantine >equations, say) without limit. This would place the notorious Godel >formula outside the limits of ideal rational reflection. Well, even with the step-sequence restriction (a computability restriction, let's say), any given statement of mathematics (our Godel sentence G, say) could arguably still be knowable by *some* relatively ideal reasoner. (This arguably follows from the considerations about iterated Godelization that I previous attributed to Kleene, but which Shaughan tells me should be attributed to Feferman.) Of course no reasoner would know every truth, but every truth might be known by some sufficiently sophisticated reasoner, which is all we need. It's also not obvious why the computability restriction is essential. I'd argue that we can form a pretty decent conception of e.g. a being who can check out the truth of P(n) for all n at once. The idea is that when we can determine the truth of P(n) for any given n, but can't determine the truth of forall(n) P(n), this is due to a limitation on our part (e.g. seriality, or finite parallelness) that can be idealized away from. Of course the idealized being here won't be computable, but it's arguably still quite coherent. Though maybe it will still induce queasiness. >If, on the other >hand, the ideal conceiver could have apriori access to the primary >intensions of G statements, we would of course thwart these cases as real >counterexamples, but at the expense of rendering the notion of "rational >reflection" or more generally "thought" very vaguely defined, and of >little help in the task of extending the scope of conceivability outward >from our own case, for which, as Angela points out, we should have >something of the same "kind". Right. If we simply defined the ideal reasoner as being able to determine the truth of any mathematical statement M, that would by the conclusion too cheaply. I'd like to think that neither of the methods of idealization suggested above is quite as cheap as this, though. In each case we're abstracting away from a clear cognitive limitation on our part, such that we have at least a rough and ready conception of what a being that does not suffer from that limitation might be like. >So I think we've got, on the one hand, some potent counterexamples to >theses like strong aprioricity and negpos, and on the other we've got a >notion of "thought" or "rational reflection" that pulls too far away from >our own case. Right. I'd like to think that the idealizations suggested above give at least a potential "middle road". Of course the question is whether such a middle road will work across the board. If we characterize the reasoning powers too weakly, we get the counterexamples to scrutability, negpos, etc. If we characterize them too strongly, we buy the conclusion at cost of making it entirely trivial. I'd like to think the middle road above yields the desired conclusion while earning it nontrivially. The hope is that something like this will work across the board. But I don't have anything like a proof of that. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Mon Mar 29 14:09:14 1999 Date: Mon, 29 Mar 1999 15:01:27 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Brad J Thompson Subject: Re: Loar's mode of presentation argument To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO On Mon, 29 Mar 1999, Erik J Larson wrote: > Maybe we should add some antecedent clause, to the effect that "If we have > access to the correct primary intensions of phenomenal and physical > states, and if we are able to completely specify (without fear of > confusion or cognitive obscurities etc) these intensions...", then we get > necessarily distinct primary intensions and a postively conceivable zombie > scenario and a good foothold on the anti-materialist conclusion. But for > myself, all I can see is the former argument-- we cannot see any necessary > connection between physical and phenomenal states. And this is just a > fact about what we can and can't see, and not much more. So I'm looking > for someone to pull me out of the epistemic trap, so to speak (I'm > guessing it will be Dave). The first antecedent clause you mention seems to be required, but not the second one (the ability to completely specify the intensions). I assume that by "specicify the intensions" you mean something like to be able to give a description. But notice that we haven't really done that for water, but instead use the phrase "watery stuff" as shorthand or to get a grip on the intension. Yet, we aren't troubled with the identification of water and H20. What is required is the assumption that our considerations of actual and counterfactual scenarios (conceiving of worlds as actual or counterfactual) engages our primary intensions and can be used as evidence for the nature of those intensions. Your worry about getting out of an epistemic trap, though, is interesting. It seems to me that one can always raise worries about going from an epistemic conclusion to a metaphysical one--that follows straightforwardly from the fact that we can't rule out evil demons and the the like. But if we *assume* that mad-dog skepticism is false, then I'm not sure we should be all that worried about the move from conceivability to possibility (being careful in the ways we have discussed, that we are correctly describing the world that we are conceiving, etc.). ---------------------- Brad J Thompson bradt@U.Arizona.EDU From owner-modality@LISTSERV.ARIZONA.EDU Tue Mar 30 01:30:36 1999 Date: Tue, 30 Mar 1999 00:23:21 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: Loar's mode of presentation argument To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Re Loar, Angela writes: >is he claiming that the conceivability of zombie worlds doesn't warrant >the move from conceivability to possibility because we are fooled into >thinking that they can actually come apart just because we can conceive of >them coming apart (due to the fact that they are presented via different >modes?) In other words, is he offering evidence to the effect >that conc. doesn't nec. track possibility given the unique quality of >phenomenal states? becaus it seems to me that the ideal conceiver would >not be mistaken by different modes of presentation, but as this has not >come up as a response to Loar (to my knowledge) I wonder if I am >understanding the argument or not.... Right, I think that's what he's up to. He's trying to give reasons why even primary conceivability does not track primary possibility. He thinks that certain phenomenal and physical concepts can have different modes of presentation but still have the same primary intension across possible worlds, or as he puts it, "express the same property". If that were so -- i.e. if cognitively unconnected concepts could have the same primary intension -- he'd certainly have a way to block the epistemic/modal bridge. But I think his explanation ends up assuming what he needs to explain. (N.B. This harks back to an earlier post of Tim's re Hill, which I haven't answered. Tim suggested that Hill didn't want to rely on a distinction between property and mode of presentation; so that on his view two cognitive distinct concepts might have the same mode of presentation. I'd be surprised if this were Hill's view, though. I think he thinks that the concepts in question have different modes of presentation; it's just that those modes express (not just refer to) the same property. It's almost definitional in much of philosophy that a mode of presentation is precisely what account for the cognitive distinctness of two concepts. Tim's examples of tactile and visual experience of a property seem to involve a difference in both mode of presentation and in expressed properties. One concept fixes reference by "what causes this visual experience", one by "what causes this tactile experience"; different reference-fixing properties, even if the same property referred to. Of course if there are concepts like this which are a priori connected (a la Molyneux), then we'll have the same expressed property, but the thesis in question won't be endangered.) Re ideal conceivers, I'm not sure why the two modes of presentation might not still be distinct even for them. Maybe the ideal reasoner will know that the two concepts pick out the same property, just as Loar and others type-B materialists "know" this. But Loar and the others can still conceive of zombies, etc. So it seems to me that the ideal reasoner might equally still find zombies conceivable, even if he/she knows that they aren't metaphysically possible. To deny that would require an a priori connection between the concepts, and I don't think Loar thinks there is such a thing. Maybe the ideal reasoner could close the gap *if* the epistemic distinctness of the concepts was due to a cognitive limitation on our part. But I don't think that's really what Loar thinks. He says explicitly somewhere that even ideal reasoning won't connect the concepts a priori. As before, I think that's the central distinction between the type-B and type-C materialists. >The argument may be flawed in >its specific guise. But I think there is a general confusion here, about >exactly what the anti-materialist argument can sustain metaphysically >given the epistemic starting point. I am tempted to say that the >(anti-materialist) argument establishes this: Because we cannot see any >necessary connection between physical and phenomenal states--the >conditional P->Q is not apriori--we cannot positively establish the truth >of materialism. I take it that the argument is stronger: Because >physical and phenomenal states necessarily have different primary >intensions, materialism is false (add in "there is a logically possible >world where P holds but not Q and so on). The second formulation of the >argument gets to the conclusion, but it seems that it rests on much >shakier epistemic ground. Well, the second is closer. I'd rather say something like "because physical and phenomenal concepts are not a priori connected, they have distinct primary intensions, so [via a chain of reasoning] materialism is false". Now, your objection is that the argument establishes only that we can't see the connection between the intensions. I can see two ways this might do. (1) We can't see an a priori connection, but a smarter being could see an a priori connection. (2) There's no a priori connection, but the two primary intensions are identical all the same. The first is the type-C materialist position; the second is the type-B position. The first preserves a deep epistemic/modal link, so perhaps you're pointing to the second. What's wrong with the second? Well, my argument relies on one of the various epistemic/modal bridging principles (whether the strong a priority thesis, or the positive conceivability to possibility thesis, or ...), which seem to hold in all the usual cases. So denying them arguably would lead to a sui generis exception, which might be seen as ad hoc. And arguably this would make metaphysical modality brute and inexplicable. But the central reason against them I think is the argument against "modal dualism" that I outline toward the end of the PPR reply. We'll be talking about those considerations in more detail in a couple of weeks. >Maybe we should add some antecedent clause, to the effect that "If we have >access to the correct primary intensions of phenomenal and physical >states, and if we are able to completely specify (without fear of >confusion or cognitive obscurities etc) these intensions...", then we get >necessarily distinct primary intensions and a postively conceivable zombie >scenario and a good foothold on the anti-materialist conclusion. But for >myself, all I can see is the former argument-- we cannot see any necessary >connection between physical and phenomenal states. And this is just a >fact about what we can and can't see, and not much more. So I'm looking >for someone to pull me out of the epistemic trap, so to speak (I'm >guessing it will be Dave). Well, hopefully the above helps. Your point here is that if we don't have perfect a priori access to primary intensions, we can jump to false conclusions. And that seems right. Again, there seem to be two different sorts of gap. First, we haven't done enough (or good enough) a priori reasoning, so we think the two concepts have different PIs even though better reasoning would reveal they are the same. ("A=B" is a priori, we just haven't figured that out yet.) Second, the two concepts have the same PI even though that's not brought out by any amount of a priori reasoning ("A=B" is not a priori but is still 1-necessary). Type-C and type-B respectively, again. The former position keeps the epistemic/modal bridge but exploits our potential cognitive limitations. The second junks the bridge. I'm not sure which exactly you're pointing to here. On my view, the second is problematic because of modal dualism, and the first is problematic because it seems that the only remotely tenable way to support it is some sort of analytic functionalism, which seems to get the meaning of phenomenal concepts wrong. The general question of a priori access to primary intensions is very interesting and important, though. It connects to the scrutability issue. Basically, some parts of the 2-D framework assume that for a concept C and a given world W, given a specification of W, we can a priori determine what the reference of C in W (considered as actual) is. Same for truth-value of statements S. That seems to plausibly fit many cases, the armchair methodology in the theory of reference, etc. But it does presuppose something like the generalized scrutability thesis: that given a complete qualitative description of the actual world (and of actual-world candidates), one can know the reference of our terms and the truth-value of our statements. What is scrutability is false? E.g., just say the epistemic theory of vagueness is correct, so one can't know the truth-value of some "X is tall" given complete qualitative info. Or just say there are mathematical statements that are true and necessary but not a priori? Then in a certain sense, the primary intensions of the relevant terms and statements are not a priori accessible. E.g. the math statement might have its primary intension true in all worlds, but we couldn't know that. And something similar for the vague statement. In these cases, I think one can say that the notion of a primary intension splits into two notions. There is the *fixing intension*, which goes with what the reference or truth-value in a given world would really be (if that world is actual). And there is the *epistemic intension*, which goes with the best a priori judgment about what the reference or truth-value in a given world will be. In the vagueness case, the fixing intension of "X is tall" might be true or false in all worlds, while the epistemic intension is indeterminate in many of them. In the mathematical case, the fixing intension of P might be true in all worlds, while the epistemic intension is indeterminate. And so on. I'd like to think that the scrutability thesis is true, so these cases can't arise. If so, the fixing intension and the epistemic intension are always the same. If such cases do arise, though, things get tricky. What if an opponent suggests that the consciousness case is like this? In fact, Joe Levine and Andrew Melnyk have both suggested that a loophole may arise from the assumption of accessibility of primary intensions (see Levine's review). I think that inaccessibility of PI may not be enough for the materialist here. In the cases above, this inaccessibility delivers inscrutable truths (gaps between negative conceivability and possibility). But zombies are positively conceivable, so the opponent needs strong necessities, not just inscrutable truths. And it's not clear that inaccessibility of PIs can deliver strong necessities. At the very least, it would need a much more radical form of inaccessibility (e.g. one that positively misleads us about a PI, rather than just not telling us everything). --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Tue Mar 30 09:55:14 1999 Date: Tue, 30 Mar 1999 10:53:22 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Rachael J Parkinson Subject: Re: ideal conception To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I want to echo some of Angela's concerns here, and particularly take issue with point ii) that Chalmers makes: > The case of more concepts is like the cases of time and memory, I > hope. Maybe the "second-order" idealizations (the ones we haven't > thought of) go beyond this. And maybe that does raise a question > about how well we can judge whether something is ideally conceivable > or not. Of course there is a legitimate epistemic worry here: ideal > conceivability is not epistemically transparent to us. But that's > compatible with saying (i) there's a fact of the matter about whether > P is ideally conceivable (whether or not we can know it), and (ii) in > many cases, we can know whether or not P is ideally conceivable (e.g., > we know it's ideally conceivable that something exists, and we know > it's not ideally conceivable that 0+0=1). So there's a fair amount we > can know, even if things get hazy around the edges. Josh posed a concern about whether we have cognitive limitations that limit what we can say about our cognitive limitations. Dave responded by suggesting that we have a grasp of our own cognitive limitations which we can use as a springboard to get to an ideal conceiver. We can conceive of beings which lack our cognitive limitations, those beings can conceive of smarter beings which lack their cognitive limitations and so forth in an expanding circle. Hopefully, this will eventually give us a useful notion of an ideal conceiver. But this seems problematic to me. Given that we may not be aware of our own cognitive limitations, it seems that we do not really have access to ideal conceivability. (point ii) How do we know that it is not conceivable that 0+0=1 or that there are round squares? Perhaps we have a cognitive limitation that we do not come close to recognizing which prevents us from seeing that these things *are* ideally conceivable. If we do not have access to what an ideal conceiver can conceive of, this seems to seriously hamper our ability to do metaphysics, particularly when it comes to determining logically possible worlds. -Rachael From owner-modality@LISTSERV.ARIZONA.EDU Tue Mar 30 14:31:45 1999 Date: Tue, 30 Mar 1999 14:24:44 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: ideal conception To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Rachael writes: >But this seems problematic to me. Given that we may not be aware of our >own cognitive limitations, it seems that we do not really have access to >ideal conceivability. (point ii) How do we know that it is not conceivable >that 0+0=1 or that there are round squares? Perhaps we have a cognitive >limitation that we do not come close to recognizing which prevents us >from seeing that these things *are* ideally conceivable. If we do not have >access to what an ideal conceiver can conceive of, this seems to >seriously hamper our ability to do metaphysics, particularly when it comes >to determining logically possible worlds. It's true that our epistemic access to ideal conceivability is limited by not knowing exactly what a relatively ideal conceiver could and couldn't conceal. But I'd like to think it isn't this limited. I think we *know* that 0+0=1 and we know that this is a priori. Similarly, I think we know that it is a priori that there are no round squares. Given that we know that, we know that round squares are not (positively or negatively) conceivable even for a relatively ideal reasoner, and the same for 0+0=1. It seems to me that to be skeptical about these things is just to be skeptical about whether we really know a priori that 0+0 is not 1. After all, if a smart enough being can conceive of a world such that if that world is actual, 0+0=1, then we can't know a priori that 0+0 is not 1. Now maybe that's an interesting sort of skepticism, but I don't think talk of ideal conceivability makes it more or less plausible. I take it most of us think we have reasons to think "0+0=0" is true and a priori; so we have good reason to believe that even an ideal conceiver couldn't conceive otherwise. Similarly, I think we have reason to believe that "round squares do not exist" is a priori, so we have good reason to believe that even relatively ideal reasoners can't conceive of them. And we have good reason to believe that "there is a phone on this desk" is a posteriori, and that even a relatively ideal reasoner could conceive otherwise. That's to say, we have good reason to believe that any amount of further and better reasoning isn't going to change the status of the knowledge here. The moral, perhaps, is that even though we can't give an exhaustive characterization of what makes for relatively ideal reasoning, that doesn't mean we have no grip on what sort of things relatively ideal reasoning can and can't tell us. To be skeptical about that, I think, would be to be skeptical about our reasoning processes in general. --Dave.