From owner-modality@LISTSERV.ARIZONA.EDU Tue Mar 30 14:54:23 1999 Date: Tue, 30 Mar 1999 14:57:45 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Horgan on properties To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO A short note on Horgan and properties. . . Horgan asks what information the Demon needs in order to complete the task of cosmic hermeneutics. The demon knows: (1) All the a priori accessible truths; (2) All the microphysical facts. What else does it need? Horgan suggests that identity statements might do the trick. If all the macrophysical objects (states of affairs, etc.) are (ontologically, rather than conceptually) identical to various concatinations of microphysical objects, then macrophysical states of affairs necessarily supervene on microphysical states of affairs, *without the introduction of brute metaphysical relations,* which Horgan finds problematic. But, says Horgan, there's a problem. "It is most implausible, on the face of it, to suppose that every sortal and non-sortal *property* expressible in our higher-level vocabularly is also finitely expressible in microphysical vocubulary. Surely nobody seriously believes that that there is a single finitely-long microphysical predicate which expresses the property of being a chair, or being a coffee pot, or thinking of Vienna" (23). I think that the above is true, but I (very tentatively) suggest that there may be around this particular problem. My suggestion is that the materialist go nominalist on properties, (as many of them are wont to do anyway). There ain't no such thing as the property of being a chair, or being a coffee pot, and so on. There are, of course, chairs, but 'there are chairs' is true not because of a property of chairhood (conceived of as some unviersal over and above the microphysical structure of the world), but because certain concrete particulars (tropes) stand in certain relations to each other. Now surely (some) relations are included in the basic microphysical facts that the demon knows - at least facts about the relations that microphsical particles bear to one another. The nominalist about properties would suggest that those property facts such as there are logically fall out of (are entailed by) the microphysical facts (where these include microphysical relations). My materialistic nominalist might agree that she can't define in reductive terms, but she can reduce the (finite) chair tropes, such as there are, to (1) and (2). Insofar as there are properties, they are fully ontologically reducible. Now, it may be that there are independent problems with this suggestion - it may not get you anywhere near cosmic hermeneutics - but that's another story. PS. I don't find nominalism about properties attractive, but many do. Well, it's a thought. t. 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 30 23:22:20 1999 Date: Tue, 30 Mar 1999 23:18:27 -0800 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: Horgan on properties To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Tim writes: >Horgan asks what information the Demon needs in order to complete the task >of cosmic hermeneutics. The demon knows: >(1) All the a priori accessible truths; >(2) All the microphysical facts. > >What else does it need? Pardon me for using this to get into a digression from Tim's point, to address some general points re Horgan's framework and some things that came up in class today. Of course on my view, nothing else is needed. Or at least, (1) and (2) are enough if we add (a) an indexical truth (saying where the demon is now), (b) phenomenal truths, and (c) a that's-all truth. But leaving those "complications" aside, it's not clear why we need to add more. In the dialectic as Horgan sets it up, he first grants the demon knowledge of (2), then asks what else is needed to get to macro knowledge. Clearly, *some* sort of bridge is required to get to macro concepts. He canvases laws (no good), identities (no good), before settling on "meaning constraints". This is a little different from (1) above: as Horgan understands meaning constraints, they include a posteriori constraints such as the constraint that "water" picks out H2O. One might argue that this supplementation of the demon's physical knowledge base with ineliminably a posteriori knowledge weakens its claim to be engaging in "cosmic hermeneutics". But in any case I think that the demon doesn't need a posteriori meaning constraints. A priori meaning constraints (the sort that correspond to primary intensions) are enough for the demon to figure out the macro truths in these cases. I think Horgan now accepts this and regrets giving in to a posteriori meaning constraints here. Further, it's arguably best not to put things in terms of giving the demon "meaning constraints" at all (even the a priori sort), if these are understood a further class of truths. E.g. it's not clear what set of explicit a priori truths about "knowledge" will enable the demon to give the right verdict on all the Gettier-et-al cases. Rather, I think it's best just to give the demon the concept of knowledge, and argue that armed with microphysical facts, that concept, and sufficient a priori reasoning, the demon can figure out the knowledge facts. More generally, I think that one just needs to stipulate that the demon has the relevant concepts. Explicit "meaning constraints" aren't necessary here and are a bit of a distraction. Of course it may be that in giving the demon the concept we're giving it implicit meaning constraints, in some sense, but we needn't build that into the picture directly, and we certainly needn't give the demon any special truths directly in its inference base (as (1) above might suggest). Horgan's different way of doing things relates to his "naturalized semantics" approach that someone (Brad?) mentioned in class. In effect, he argues that meaning constraints aren't really problematic, as they're just another class of high-level fact. Eg. the demon can figure out English truths by looking at English speakers, figuring out what their terms mean, and seeing whether or not their statements are true. And can presumably figure out meaning constraints the same way, including explicit truths about what `water' in English refers to, etc.. I find this picture a little hard to follow, though. First, if we are really getting the meaning constraints "for free" via naturalized semantics, why bother building them in in the first place? The answer, presumably, is that even in the naturalized semantic case, we need *some* bridge to get us from microphysics to what is "true in English" and so on -- after all, we'll be stating such claims using such non-microphysical concepts as "true". So presumably, we'll at least need some bridge re the concepts of "truth", "reference", etc, otherwise we'll never be able even to get to claims about what is true in English, etc, even on the naturalized semantics approach. And presumably the bridge that is needed will be precisely something like meaning constraints for the concepts of "truth", "reference", etc. So it seems to me that Horgan can't really get the meaning constraints as cheaply as he'd like: he has to *start* with at least some of them. I don't think starting with these meaning constraints is problematic myself (one just has to suppose that the demon has the concepts of truth, reference, etc), but it does suggest that the naturalized semantics approach can't do all Horgan's work for him. In essence, we have to suppose that the demon *has* and *uses* some of the relevant concepts, and doesn't just *mention* them as part of a third-person interpretation. At least, it has to have the concepts of truth, reference, etc. And once we've done that, we might as well give the demon arbitrary concepts. That way, to see if any truth T is derivable by cosmic hermeneutics, we just give the demon the concepts in T, physical knowledge, and let things rip. No need to do everything resolutely in the third-person, restricting the demon to "X's utterance of T is true" and such. The demon can just tell us, T! The two approaches might even come apart in some cases. E.g. just say nonmental truths are a priori entailed by physical truths, but mental truths aren't. And now consider "water is wet". On the "concept use" approach to cosmic hermeneutics, where we give the demon (say, an idealized version of me) the concepts, then it will be able to figure out from microphysics etc that water is wet. On the "naturalized semantics" approach, though, the demon may not be able to figure out what "water" in English refers to, as that will probably depend on various mental facts about the intentions and beliefs, etc, of English speakers, which aren't accessible by hypothesis. If so, it will come out that cosmic hermeneutics doesn't even work for "water is wet", despite a priori derivability. I think that's the wrong result, and we want an understanding of CH so that CH is possible for "water is wet" in this case. If so, I think we need the concept-use approach. >What else does it need? Horgan suggests that identity statements might do >the trick. If all the macrophysical objects (states of affairs, etc.) are >(ontologically, rather than conceptually) identical to various >concatinations of microphysical objects, then macrophysical states of >affairs necessarily supervene on microphysical states of affairs, *without >the introduction of brute metaphysical relations,* which Horgan finds >problematic. Well, I'd argue that bringing in identities as basic here may well bring in something "brute" in some sense. Something like the type-B materialist bringing in a brute fact that pain = C-fibre firing, or some such. If there is no epistemic link from microphysics to that, then we have at least an epistemic bruteness, if not a metaphysical bruteness. And it would seem that as with a posteriori meaning constraints, we'd have then given up on pure cosmic hermeneutics. But set that aside. >But, says Horgan, there's a problem. "It is most implausible, on the face >of it, to suppose that every sortal and non-sortal *property* expressible >in our higher-level vocabularly is also finitely expressible in >microphysical vocubulary. Surely nobody seriously believes that that there >is a single finitely-long microphysical predicate which expresses the >property of being a chair, or being a coffee pot, or thinking of Vienna" >(23). > >I think that the above is true, but I (very tentatively) suggest that >there may be around this particular problem. My suggestion is that the >materialist go nominalist on properties, (as many of them are wont to do >anyway). There ain't no such thing as the property of being a chair, or >being a coffee pot, and so on. There are, of course, chairs, but 'there >are chairs' is true not because of a property of chairhood (conceived of >as some unviersal over and above the microphysical structure of the >world), but because certain concrete particulars (tropes) stand in certain >relations to each other. Now surely (some) relations are included in the >basic microphysical facts that the demon knows - at least facts about the >relations that microphsical particles bear to one another. The nominalist >about properties would suggest that those property facts such as there >are logically fall out of (are entailed by) the microphysical facts (where >these include microphysical relations). My materialistic nominalist >might agree that she can't define in reductive terms, but she can >reduce the (finite) chair tropes, such as there are, to (1) and (2). Insofar >as there are properties, they are fully ontologically reducible. Now, it >may be that there are independent problems with this suggestion - it may >not get you anywhere near cosmic hermeneutics - but that's another story. Hmm, interesting. I'm not sure I completely follow. What are the "bridges" that we are adding to get us from microphysics to macro truths? Horgan canvased the possibility of adding identities such as "water = H2O", but the trouble was that analogous identities don't seem available in many other cases (chairs, coffee pots, thinking). You're suggesting maybe adding token reductions? I presume that would be something like "this P1 is a chair", "this P2 is a chair", and so on for every chair in the world? I guess the trouble, as you suggest, is that this doesn't look much like cosmic hermeneutics. Rather, it's more like taking microphysics and tagging every chair explicitly as a chair, every coffee pot explicitly as a coffee pot, and so on. Not much surprise that we can get to the chair facts from there! I think your suggestion here might end up being relevant to a metaphysical reduction, as opposed to the sort of epistemic reduction Horgan is after. It might be that a token-token but not type-type identity is enough to save physicalism as a metaphysical doctrine (though there are many tricky issues here); and a tropewise reduction of properties might do much the same thing. But in building in so many references to specific instances, it seems to build in too much to qualify as cosmic hermeneutics. Rather, this might lead to a way in which physicalism could be true *without* cosmic hermeneutics being possible! That would be interesting, though of course at the end of the day I think such a combination isn't tenable. If there are these tropewise identifications, there ought to be some way of figuring out the identifications themselves a priori from the physical facts. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Mon Apr 5 10:49:02 1999 x-sender: agillies@pop.u.arizona.edu Date: Mon, 5 Apr 1999 11:01:48 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony S Gillies Subject: cosmic hermeneutics To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO All, Some thoughts on cosmic hermeneutics (CH) and reductive explanation (RE). There seems to be a bit of carelessness in talk of the possibility of CH with respect to what informaton we are to give a Laplacian-type knower: is it all the physical facts, or is it all he facts about physics? Byrne, e.g., seems to run the two together. But surely these are are conceptually distinct sets of facts, and their difference hinges on where we put natural laws. I think there is a dilemma lurking nearby, so I want to try to draw it out. (I'm not sure what force the dilemma might have for particular theories, over and above being a useful guide to taxonomy.) Let F be the set of facts we give to an ideal knower, and let P be the conjunction of all the microphysical facts, and let L be the conjunction of all the (true) physical laws. By hypothesis P is in F. The question is whether L is also in F. If L is *not* in F, then I think we are getting into Byrne's worry about conceivability. It is hard to see how one can get from low-level sentences describing mass, charge, spin and so on to high-level descriptions about juiciness or whatever without using th sort of bridge laws that are conjuncts of L. And if that's right, then it isn't very clear in what sense it is a priori that F implies a given macro-event M. (And, if I read Josh's worry re Aristotle correctly, this is similar in spirit to his point.) Now suppose that L *is* in F. So we give the ideal knower both all the micro-facts and all the causal laws. But then CH is possible iff the deductive-nomological theory of explanation is right. Right to Left: If the D-N model is right, then since explanation is just a deduction from conditions and laws to empirical facts, clearly CH is possible---CH amounts to "large scale" D-N prediction/explanation. Left to Right: If CH is possible, there is a deduction from F to M. But F= {P, L},where P is the description of all the physical conditions (antecedent circumstances) and L is the statement of all the general laws. M is just a macro empirical fact. But these are the sufficient ingredients for the D-N model of explanation. I don't think there are any particularly strong conclusions here: if L is not in F, then we might want a bit more said about what is clearly and distinctly conceivable about CH. If L is in F, then CH is possible iff D-N explanation is right. And, by extension, since advocates of CH seem to think that a RE of A in terms of B is possible iff CH with respect to B implies A, then CH is possible iff D-N model for RE is right. Thony "Curious green ideas sleep furiously." From owner-modality@LISTSERV.ARIZONA.EDU Tue Apr 6 11:39:08 1999 Date: Tue, 6 Apr 1999 11:36:51 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: Re: cosmic hermeneutics To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Here's why don't buy the D-N model: Why did Elvis not get pregnant? (1) Elvis took birth control pills; and (2) If you take birth control pills, then you won't get pregnant. It's a D-N explanation! On Tue, 6 Apr 1999, Anthony S Gillies wrote: > Well, I guess I mean *I* don't want to draw any strong conclusions, since > I'm not sure that I'm convinced that D-N explanation is so bad (Salmon > and his ilk not withstanding). I might be convinced, and just don't know > it, though. > > Thony > > >Thony, Just a point of clarification. I take it that at least part of > >your point about the D-N model is that, if we have committ ourselves to > >the D-N model to hold that CH is possible, this is potentially problematic > >(disastrous?) to the CH thesis, since there are well-known problems with > >D-N as an adequate model of scientific explanation (e.g., Salmon's point > >that D-N is neither necessary or sufficient for explanation). (I realize > >that this is an extrapolation, since you say that there *aren't any > >particularly strong conclusion here*, but I think it's clear that a > >committment to a D-N model of explanation would be bad, so I suppose I > >would want to add that, if you are correct, there IS a strong conclusion > >here; namely, there is something amiss with the CH thesis on this score.) > > > >Erik > >(p.s. there is no way out for you, if you do not answer me I will hunt you > >down for further clarification) > > > > > > > "Curious green ideas sleep furiously." > Timothy J. Bayne RM. 213 Social Science Department of Philosophy University of Arizona Tucson, AZ 85721 USA Hm ph. (520) 298 1930 From owner-modality@LISTSERV.ARIZONA.EDU Tue Apr 6 11:45:14 1999 x-sender: agillies@pop.u.arizona.edu Date: Tue, 6 Apr 1999 11:29:38 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony S Gillies Subject: Re: cosmic hermeneutics To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Well, I guess I mean *I* don't want to draw any strong conclusions, since I'm not sure that I'm convinced that D-N explanation is so bad (Salmon and his ilk not withstanding). I might be convinced, and just don't know it, though. Thony >Thony, Just a point of clarification. I take it that at least part of >your point about the D-N model is that, if we have committ ourselves to >the D-N model to hold that CH is possible, this is potentially problematic >(disastrous?) to the CH thesis, since there are well-known problems with >D-N as an adequate model of scientific explanation (e.g., Salmon's point >that D-N is neither necessary or sufficient for explanation). (I realize >that this is an extrapolation, since you say that there *aren't any >particularly strong conclusion here*, but I think it's clear that a >committment to a D-N model of explanation would be bad, so I suppose I >would want to add that, if you are correct, there IS a strong conclusion >here; namely, there is something amiss with the CH thesis on this score.) > >Erik >(p.s. there is no way out for you, if you do not answer me I will hunt you >down for further clarification) > > "Curious green ideas sleep furiously." From owner-modality@LISTSERV.ARIZONA.EDU Tue Apr 6 10:53:23 1999 Date: Tue, 6 Apr 1999 10:49:24 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Timothy J Bayne Subject: ON what Cosmic Hermeneutics involves To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Clarificatory point. I thought that the CH claim is that a being with (1) all the physical facts; and (2) all a priori knowledge; can (3) come to know all the facts, where (3) should be read as, 'can at one time know everything there is to know'. But in fact Byrne nor, as far as I can tell, Horgan, puts the CH conclusion in this way. Rather, they say that for *any* truth, one can come to know it, not that the Laplacean demon can come to know all the truths at once (although Horgan seems to gesture in this direction in at least one place). Call the former claim the existential claim, and the latter the universal claim. The problem is this: as far as I can see, the Demon might know everything, but can it know that it knows everything? We, as theorists, can stipulate the 'that's all' fact for any world, but how does one establish the that's all truth from *within* a world? One might believe that one knows everything, and it might even be true that one knows everything, but how could the belief that one knows everything be justified? (Why hasn't this objection been levelled against theistic omnipotence? Because God is said to *create* everything, and thus (it seems) God knows when God knows everything. I think that's the reason.) As I said, I think that Byrne (and perhaps) Horgan reject the universal claim in favor of the existential claim, and the existential claim isn't open to this problem. Or is it? Consider the claim: 'The demon knows everything'. The demon will (I guess) believe this, but how can it know it? But is this a truth? No. Is it a truth that the demon has true beliefs about everything? That may be true. Can it know that it has true beliefs about everything (i.e. can it know that it believes every truth). ? Again, I don't think so: how could you be justified that your set of beliefs is exhaustive? So, it seems that there is a particular truth that the demon cannot *know*. Here's an objection I tried to develop, but the wheels fell off. If someone can put them on again, go ahead! Suppose a weird world, in which p is true iff the demon doesn't believe it. The fact of believing p causes p to be true no longer, and the fact of not-believing p causes p to be true. Now, if you put the demon outside of the world (as Horgan does) you might be able to say that, necessarily, there are no such facts p. But - as far as I can see - putting the demon inside the world seems to open up the possibility of such facts. Problem is, fully developing this objection involves saying something about the relation of p and the physical facts. It needs to be the case that the demon can know the physical facts on which p supervenes, otherwise this isn't a counter-example to CH. And if looks like if the demon does know the physical facts on which p supervenes, and it knows all the a prior facts, then it will be able to work out whether or not p is true. In other words, the defender of CH seems to be within her rights in claiming that the notion of a p truth is incoherent. 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 Apr 6 14:46:54 1999 Date: Tue, 6 Apr 1999 14:42:27 -0700 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: cosmic hermeneutics To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Re cosmic hermeneutics and D-N explanation, Thony writes: >There seems to be a bit of carelessness in talk of the possibility of CH >with respect to what informaton we are to give a Laplacian-type knower: >is it all the physical facts, or is it all he facts about physics? >Byrne, e.g., seems to run the two together. But surely these are are >conceptually distinct sets of facts, and their difference hinges on where >we put natural laws. I think there is a dilemma lurking nearby, so I >want to try to draw it out. (I'm not sure what force the dilemma might >have for particular theories, over and above being a useful guide to >taxonomy.) I think we are supposed to give the demon (a) all the microphysical facts and (b) all the microphysical laws. Just (a) will be enough if one is a Humean about laws and causation (in that case (b) will follow from (a) a priori). On a non-Humean views, one will need both, to bring in facts about laws, causation, dispositions, counterfactuals, etc. > >Let F be the set of facts we give to an ideal knower, and let P be the >conjunction of all the microphysical facts, and let L be the conjunction >of all the (true) physical laws. By hypothesis P is in F. The question >is whether L is also in F. If L is *not* in F, then I think we are >getting into Byrne's worry about conceivability. It is hard to see how >one can get from low-level sentences describing mass, charge, spin and so >on to high-level descriptions about juiciness or whatever without using >th sort of bridge laws that are conjuncts of L. And if that's right, >then it isn't very clear in what sense it is a priori that F implies a >given macro-event M. It's not clear to me whether you intend L to be the collection of microphysical laws, or of physical laws more generally (including macrophysical laws). In the first case, L will be in F, as suggested above. In the second case, L will not be in F, but the advocate of cosmic hermeneutics will argue that L follows from F a priori. How do we get from microphysical facts and laws to facts about juiciness, etc? Presumably, by virtue of possessing the concept of juiciness. The CH claim is that armed with this concept, one grasps how this situation applies to various epistemic possibilities, and in particular, that one can know how it applies to a possibility microphysically described. So e.g. one will be able to get from a structural/functional characterization of an apple to knowledge of whether it is juicy. E.g., look to see how much water it contains! Where we've antecedently figured out what is water in the world, by the "watery stuff" method. And the CH advocate will argue that it is not a priori conceivable that something satisfies this physical characterization without being juicy. Etc. Of course whether this is really possible for concepts such a juiciness is just what the debate over CH is about. I am personally agnostic about whether it is possible, though I'm sympathetic with the idea. What I do think is that the juiciness facts follow a priori from the physical plus phenomenal facts. Once one allows the demon to have phenomenal facts too, e.g. about what the apple looks like, then its task is much easier. >Now suppose that L *is* in F. So we give the ideal knower both all the >micro-facts and all the causal laws. But then CH is possible iff the >deductive-nomological theory of explanation is right. Right to Left: If >the D-N model is right, then since explanation is just a deduction from >conditions and laws to empirical facts, clearly CH is possible---CH >amounts to "large scale" D-N prediction/explanation. Left to Right: If >CH is possible, there is a deduction from F to M. But F= {P, L},where P >is the description of all the physical conditions (antecedent >circumstances) and L is the statement of all the general laws. M is just >a macro empirical fact. But these are the sufficient ingredients for the >D-N model of explanation. Well, the laws most relevant to micro-macro D-N explanation are micro-macro bridge laws. And we certainly won't be giving the demon knowledge of micro-macro bridge laws. So L as construed here won't be in F. With knowledge of bridge laws, the demon's task would be much easier, but (as Horgan notes at one point) it trivializes things. It will turn out that CH is true even on a property dualist view of qualia, for example. (Even on my view of qualia, there are bridge laws!) So CH would certainly not imply reductive explation or any such thing. For CH to be interesting, the demon needs to be able to figure things out from microphysical facts and laws alone, in combination with its possession of the high-level concept. As Tim notes, there are various problems with the D-N model of explanation. But I think most of those can be set aside here. The most central point is that it is widely agreed now that D-N makes a terrible model of *reductive* explanation (although Ernest Nagel originally intended it this way). In helping itself to bridge laws, D-N makes the reductive project relatively trivial, not obviously reductive, and consistent with the falsity of materialism, as the qualia example suggests. So most people now think reductive explanation requires a stronger constraint. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Tue Apr 6 14:56:47 1999 Date: Tue, 6 Apr 1999 10:12:15 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik J Larson Subject: Re: cosmic hermeneutics To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Thony, Just a point of clarification. I take it that at least part of your point about the D-N model is that, if we have committ ourselves to the D-N model to hold that CH is possible, this is potentially problematic (disastrous?) to the CH thesis, since there are well-known problems with D-N as an adequate model of scientific explanation (e.g., Salmon's point that D-N is neither necessary or sufficient for explanation). (I realize that this is an extrapolation, since you say that there *aren't any particularly strong conclusion here*, but I think it's clear that a committment to a D-N model of explanation would be bad, so I suppose I would want to add that, if you are correct, there IS a strong conclusion here; namely, there is something amiss with the CH thesis on this score.) Erik (p.s. there is no way out for you, if you do not answer me I will hunt you down for further clarification) On Mon, 5 Apr 1999, Anthony S Gillies wrote: > All, > > Some thoughts on cosmic hermeneutics (CH) and reductive explanation (RE). > > There seems to be a bit of carelessness in talk of the possibility of CH > with respect to what informaton we are to give a Laplacian-type knower: > is it all the physical facts, or is it all he facts about physics? > Byrne, e.g., seems to run the two together. But surely these are are > conceptually distinct sets of facts, and their difference hinges on where > we put natural laws. I think there is a dilemma lurking nearby, so I > want to try to draw it out. (I'm not sure what force the dilemma might > have for particular theories, over and above being a useful guide to > taxonomy.) > > Let F be the set of facts we give to an ideal knower, and let P be the > conjunction of all the microphysical facts, and let L be the conjunction > of all the (true) physical laws. By hypothesis P is in F. The question > is whether L is also in F. If L is *not* in F, then I think we are > getting into Byrne's worry about conceivability. It is hard to see how > one can get from low-level sentences describing mass, charge, spin and so > on to high-level descriptions about juiciness or whatever without using > th sort of bridge laws that are conjuncts of L. And if that's right, > then it isn't very clear in what sense it is a priori that F implies a > given macro-event M. (And, if I read Josh's worry re Aristotle > correctly, this is similar in spirit to his point.) > > Now suppose that L *is* in F. So we give the ideal knower both all the > micro-facts and all the causal laws. But then CH is possible iff the > deductive-nomological theory of explanation is right. Right to Left: If > the D-N model is right, then since explanation is just a deduction from > conditions and laws to empirical facts, clearly CH is possible---CH > amounts to "large scale" D-N prediction/explanation. Left to Right: If > CH is possible, there is a deduction from F to M. But F= {P, L},where P > is the description of all the physical conditions (antecedent > circumstances) and L is the statement of all the general laws. M is just > a macro empirical fact. But these are the sufficient ingredients for the > D-N model of explanation. > > I don't think there are any particularly strong conclusions here: if L is > not in F, then we might want a bit more said about what is clearly and > distinctly conceivable about CH. If L is in F, then CH is possible iff > D-N explanation is right. And, by extension, since advocates of CH seem > to think that a RE of A in terms of B is possible iff CH with respect to > B implies A, then CH is possible iff D-N model for RE is right. > > Thony > > > > > "Curious green ideas sleep furiously." > "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 Tue Apr 6 15:25:27 1999 Date: Tue, 6 Apr 1999 15:23:54 -0700 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: ON what Cosmic Hermeneutics involves To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Tim writes: >Clarificatory point. I thought that the CH claim is that a being with (1) >all the physical facts; and (2) all a priori knowledge; can (3) come to >know all the facts, where (3) should be read as, 'can at one time know >everything there is to know'. But in fact Byrne nor, as far as I can tell, >Horgan, puts the CH conclusion in this way. Rather, they say that for >*any* truth, one can come to know it, not that the Laplacean demon can >come to know all the truths at once (although Horgan seems to gesture in >this direction in at least one place). Call the former claim the >existential claim, and the latter the universal claim. I think it's probably better to interpret CH as making at most the existential claim. The universal claim may have problems. E.g. there is a "paradox of the knower" doe to Montague and Kaplan that raises Godel-style problems with the idea of a being who knows all truths. One might argue that the existential claim entails some version of the universal claim, since the conjunction of all truths is itself a truth, so the existential claim applies to it. In response, a CH advocate might suggest (1) that they are only dealing with finite claims (it's not clear whether or not this would be ad hoc), or at least (2) that the notion of the conjunction of all truths is not well defined. Alternatively, they could put forward a slightly different version of CH (see below). >The problem is this: as far as I can see, the Demon might know everything, >but can it know that it knows everything? We, as theorists, can stipulate >the 'that's all' fact for any world, but how does one establish the that's >all truth from *within* a world? One might believe that one knows >everything, and it might even be true that one knows everything, but how >could the belief that one knows everything be justified? (Why hasn't this >objection been levelled against theistic omnipotence? Because God is said >to *create* everything, and thus (it seems) God knows when God knows >everything. I think that's the reason.) Well, I take it that the "that's all" claim, like microphysics, is simply stipulated to be in the demon's knowledge base. One could argue about just how a demon could really know that, just as one could argue about just how a demon could really know all the microphysical truths, but that's not really the central issue here. It's probably better to put the CH claim in terms of knowledge of conditionals. For any truth Q, "P+T+I -> Q" is knowable a priori, where P is the conjunction of the microphysical truths, T is the that's-all truth, and I is the indexical truth. >As I said, I think that Byrne (and perhaps) Horgan reject the universal >claim in favor of the existential claim, and the existential claim isn't >open to this problem. Or is it? Consider the claim: 'The demon knows >everything'. The demon will (I guess) believe this, but how can it know >it? But is this a truth? No. Is it a truth that the demon has true beliefs >about everything? That may be true. Can it know that it has true beliefs >about everything (i.e. can it know that it believes every truth). ? Again, >I don't think so: how could you be justified that your set of beliefs is >exhaustive? So, it seems that there is a particular truth that the demon >cannot *know*. Well, *if* the demon has the that's-all fact in the antecedent part of its knowledge base, this obstacle will be removed. Again, you can raise questions about whether a demon could really have that antecedent knowledge, but we are just stipulating it here. I also note that "the demon knows everything" will likely actually be *false*, at least for the existential version of CH, as here the demon knows truths one at a time, not all at once. Maybe the demon will believe "everything is knowable", although there are tricky issues even there. >Here's an objection I tried to develop, but the wheels fell off. If >someone can put them on again, go ahead! Suppose a weird world, in which >p is true iff the demon doesn't believe it. The fact of believing p >causes p to be true no longer, and the fact of not-believing p causes p to >be true. Now, if you put the demon outside of the world (as Horgan does) >you might be able to say that, necessarily, there are no such facts p. But >- as far as I can see - putting the demon inside the world seems to open >up the possibility of such facts. Nice objection. In fact there is a perfectly straightforward such P. Let's assume that the demon doesn't in fact know all truths (we're going with the existential rather than the universal version, after all). So there is some truth Q that the demon doesn't know. So let R = "Q and the demon does not know that Q". R is clearly true. But if the demon were to come to know that R, he would thereby come to know that Q, so P would be false. Hence the demon cannot know that R! This is precisely the "paradox of knowability" that we discussed in class. There is a bit of literature on it. It clearly poses a problem for some formulations of CH. It can't be the case that all truths are knowable for the demon, and it can't even be the case that given full physical knowledge, all truths are knowable. The best response, I think, is to go back to the conditional formulation, which says that the *conditional* from microphysics to any given truth (i.e. "P+T+I -> Q" for any Q) is knowable a priori. In the case where Q is the paradoxical R above, the demon can't know that R, but it can still know a priori that *if* P+T+I, then R. Of course, on the hypothesis of P+T+I, the demon will be able to figure out both that Q above is true and that he doesn't know it, so he'll be able to figure out R. Because the antecedent here is merely hypothetical, this doesn't imply that he actually knows R, but he can still figure out the conditional a priori. >Problem is, fully developing this objection involves saying something >about the relation of p and the physical facts. It needs to be the case >that the demon can know the physical facts on which p supervenes, >otherwise this isn't a counter-example to CH. And if looks like if the >demon does know the physical facts on which p supervenes, and it knows all >the a prior facts, then it will be able to work out whether or not p is >true. In other words, the defender of CH seems to be within her rights in >claiming that the notion of a p truth is incoherent. I think your P is actually coherent, as above. In the case where the demon *knows* all the microphysical facts, it will be able to figure out Q, and "I didn't know that Q", but it coming to know these things it will render "Q and I don't know that Q" false, so the original truth was still arguably unknowable. But if we put things in the conditional format, there doesn't seem to be a problem with the demon knowing a priori that *if* microphysics is like so, then P is true, even though P itself is unknowable. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Sat Apr 3 17:11:19 1999 Date: Sat, 3 Apr 1999 18:10:14 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Josh Cowley Subject: reducitve concepts To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Cosmic Hermenudics. I want to push the Aristotle problem I was suggesting in class. The basic idea is that we have a person who has all the concepts of Aristotle (ie. no modern science) and is an ideal reasoner. We then give him all the micro-physical truths and the "that's all" fact. Now consider Aristotle's favorite wooden chair. The question I have is whether Aristotle can get from the micro-physical truths to "There are macro objects in the world." Without any further concepts. In particular whether he needs reductive concepts linking macro objects to micro-physical stuff. Many psychologists have suggested that we have an innate concept of object and that the concept doesn't really change as we age. Spelke has suggested four principles which characterize our object concept. 1) An object is a connected and bounded region of matter that maintains itself when in motion. 2) Principle of continuity: Roughly this says that objects do not teleport. 3) Principle of contact: Objects don't move unless touched by something that is moving. 4) Principle of solidity: One object cannot pass through another object (without damage). Suppose that something like this is right. Presumably, Aristotle had this concept of object. Now the question is whether he can find anything in the micro-physical description of the world that matches this concept of object. The problem is that the modern view of quantum mechanics has things violating every one of these principles. Principle 1): From the micro-physics standpoint "matter" is all over the place. The only boundaries we can draw are based on densities. We might be able to come up with a sense of "connected" which involves the forces which make one chunk of matter more dense than another. But nothing in QM is connected in the macro sense of "touching." Finally, matter does not maintain its connectedness while in motion. Electrons leave my body every time I stand up from my couch and touch something metal. In fact the micro-physical elements in every object are constantly being exchanged with elements from other objects as well as being in constant motion within an object. Principle 2): There is a phenomena in QM known as quantum tunneling. Basically elements at the QM level simply move from one place to another without existing anywhere in between. At the level of individual QM elements this is happening all the time. Furthermore, there is a finite probability that this could happen to a macro sized object as well. But if our concept of macro sized objects doesn't allow for this then how is our ideal Aristilian reasoner going to figure out that some particular portion of space time currently has an object in it. I'll skip 3 and 4 as I think the point has been made. We get from everyday objects to particle physics by reductive analysis. Objects are collections of molecules. Molecules are composed of atoms. Atoms are composed of some elementary particles. An ideal reasoner with only Aristotle's concepts and a theory of elementary particle physics won't find objects in a quantum mechanical description of the world. Nothing behaves the way objects do when described by quantum mechanics. In a molecular description of the world we get collections of things that do behave like objects. In an atomic description of the world we get things that behave like molecules. Finally, in a QM description of the world we get things that behave like atoms. But I think you really do need the intervening concepts to get from QM to objects. From owner-modality@LISTSERV.ARIZONA.EDU Mon Apr 5 16:22:28 1999 Date: Mon, 5 Apr 1999 16:13:33 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik J Larson Subject: Re: CH and Quantum Mechanics To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Hi everyone, I've been thinking that Horgan may be too dismissive of the possibility that quantum indeterminacy could spell trouble for the supervenience thesis. He points out that DM and S are "logically independent", and apparently thinks that any quantum indeterminacy in physics would apply only to DH. I want to attempt to develop a quantum counterexample to CH and the supervenience thesis. Consider a conditional whose antecedent contains a complete specification of microphysical facts for some time slice, and whose consequent contains the set of macrophysical facts entailed by the antecedent. Apply this to the Shrodinger's cat thought experiment, where a photon impinges a set of half-silvered mirrors with probability one-half of deflection or penetration and thereby either triggers or fails to trigger a mechanism which breaks a flask of poison in a sealed container containing a live cat. The idea is that the cat's life status is indeterminate while the photon is superposed over possible states in the deflective mirror apparatus, until someone opens the container, collapsed the superposed "wave-function" and either observes a healthy cat or a dead one. Now, given the interpretation of quantum theory that posits a real indeterminacy (takes a realist interpretation on the wave-function and its collapse), it would seem that there would be one perfectly well-formed macro-fact--the cat being either alive or poisoned to death--that cannot be entailed by the antedent of the CH conditional. The cat example seems contrived (it is) but the breakdown of the CH thesis (S) would seem to follow antime we want to know macro-facts about the future that have to be entailed from prior microphysical facts that enter into quantum superpositions, generally. I'm thinking that Horgan dismisses quantum indeterminacy because he considers only cases of entailment from some set of microphysical facts to some other set, such as the conditional from rates of decay for the particles of a radioactive element to a specification of a particular particle's emission, and other standard examples of quantum indeterminacy. That being said, the anticipated objection is that some interpretations of quantum theory do not take a realist line on the wave-function collapse. (I don't remember all the details of the various interpretations, so I'll say more generally that some interpretations apparently can accomodate the collapse without conceding fundamental indeterminacy). Still, if it is at least plausible that quantum theory may turn out to be fundamental to a completed physics, then in that conceivable scenario the supervenience thesis of CH will fail to hold, and so the argument against materialism--predicated on the distinction between failed supervenience only in the phenomenal case and not for standard physical macro objects--would seem to be in big trouble. This strikes me as odd, since the cogency of the argument would then hinge on the contingencies of our physical theory qua quantum indeterminacy. It seems strange to say that IF our physical theory turns out such and such, then we have a good argument agaist materialism, but possibly not. It also seems strange that the anti-materialist argument should be put forth so assertively given our epistemic confusion about QM in general, since it seems that the failure of CH and S is at least possible on some interpretations, and in that case the argument wouldn't be very compelling. 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 Tue Apr 6 15:15:32 1999 Date: Tue, 6 Apr 1999 15:10:59 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony T Lane Subject: Re: reducitve concepts To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Some thoughts on Josh's email. I'm not sure that the features of quantm mechanics Josh notes are necessarily all that problematic for cosmic hermeneutics. Although I do not know all that much about QM, I think that, when we say that an atom is a certain size, we specify a sphere within which there is a fairly good chance of finding the appropriate electrons. Although this certainly allows that there is some amount of fuzziness in the boundaries of objects, this does not necessarily undermine the possible success of cosmic hermeneutics-- given the exceedingly high probability of electrons and so forth occupying particular regions of space, the 'macro' properties of solidity and so forth arise. I think the root of Josh's worries seems to be the quantum indeterminacy stuff. The Heisenberg Uncertainty Principle asserts that we cannotknow both the location of a particle and its velocity beyond a certain level of accuracy. This arises because, in the process of measuring, we must interact with the particle-- if we measure the velocity of a particle, we alter its position, and so forth. A LaPlacian Demon, I presume, is not constrained by this as it simply knows all of the mcirophysicl facts of the universe at a particular instant. I wonder whether this bypasses the problems of indeterminacy with which we are faced when we study quatum phenomena. Anthony > Principle 1): From the micro-physics standpoint "matter" is all over the > place. The only boundaries we can draw are based on densities. > We might be able to come up with a sense of "connected" which involves > the forces which make one chunk of matter more dense than another. > But nothing in QM is connected in the macro sense of "touching." > Finally, matter does not maintain its connectedness while in motion. > Electrons leave my body every time I stand up from my couch and touch > something metal. In fact the micro-physical elements in every object > are constantly being exchanged with elements from other objects as > well as being in constant motion within an object. > Principle 2): There is a phenomena in QM known as quantum tunneling. > Basically elements at the QM level simply move from one place to > another without existing anywhere in between. At the level of > individual QM elements this is happening all the time. Furthermore, > there is a finite probability that this could happen to a macro sized > object as well. But if our concept of macro sized objects doesn't > allow for this then how is our ideal Aristilian reasoner going to > figure out that some particular portion of space time currently has an > object in it. > > I'll skip 3 and 4 as I think the point has been made. We get from > everyday objects to particle physics by reductive analysis. Objects > are collections of molecules. Molecules are composed of atoms. Atoms > are composed of some elementary particles. An ideal reasoner with > only Aristotle's concepts and a theory of elementary particle physics > won't find objects in a quantum mechanical description of the world. > Nothing behaves the way objects do when described by quantum > mechanics. In a molecular description of the world we get collections > of things that do behave like objects. In an atomic description of > the world we get things that behave like molecules. Finally, in a QM > description of the world we get things that behave like atoms. But I > think you really do need the intervening concepts to get from QM to > objects. > From owner-modality@LISTSERV.ARIZONA.EDU Tue Apr 6 21:56:29 1999 Date: Tue, 6 Apr 1999 21:53:57 -0700 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: CH and Quantum Mechanics To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Lots of interesting issues about QM and cosmic hermeneutics. Josh suggests that one cannot get from QM to "there are macro objects" a priori, as the concept of an object involves various principles (connectedness, continuity, contact, solidity) that are false in QM. I think it's certainly right that QM poses problems here. When philosophers talk about microphysics in this context they usually assume some sort of classical mechanics. This makes things easier, as one has definite entities with definite position, velocity, mass, etc, right at the bottom level. From there, it's easy to get to macro position, velocity, shape, size, mass density, etc. But from QM, this isn't so easy. Of course one thing one might do is cheat, and say: in the base, we'll allow not just microphysics but chemistry. Then we'll be allowed facts about definite molecules with definite positions, masses, etc. And from there we can get the macro stuff out. We still have an interesting and substantive thesis, which will still probably capture what's at issue between someone like me (or Jackson, Lewis, etc), and someone like Byrne, Block, or Stalnaker. But still, this leaves open the interesting question of just what follows from microphysics alone, whether chemistry follows from microphysics, etc. On Josh's principles: of course violations at the micro level alone may not matter too much, as it's at the macro level that we expect to find objects. But serious violations at the macro level will matter. One strategy would be to argue that such violations won't happen, or will happen only rarely. E.g., quantum tunnelling discontinuities are presumably rare at the macro level, so they may not rule out objects there. And maybe (maybe) there is some way to make sense of connectedness using the QM notion of decoherence, etc. So I think there are strategies here. A more detailed answer depends really on what interpretation of QM one gives. Without such an interpretation, one can't really know what QM is saying about microphysics. With such an interpretation, I think one gets different results re cosmic hermeneutics. (1) Bohm interpretation (with particles governed by the wave function): Here there isn't too much trouble for cosmic hermeneutics. One has definite particles all the way down, just as on a classical view. So no more problems deriving macro objects here than on a classical view. (2) Everett interpretation (with a giant superposed wave function with all possibilities at once): Here, one certainly can't derive macro "truths" from the wave function. E.g., the "truth" that the cat is alive or the point is pointing up won't follow from the superposed wave function. But arguably, on the Everett view these aren't really truths. The cat is really a superposition of both alive and dead in reality, and so on. And of course we don't expect indeterminate matters to be settled by cosmic hermeneutics on microphysics, any more than questions of tallness are. Alternatively, one can say that the sense in which the cat is "really" alive is relative to my own mind -- my mind splits into a component, one of which sees a dead cat and one (this one) an alive cat. If so, the cat status won't follow from microphysics, but it will follow from microphysics plus facts about my mind. What to say here depends on whether mental facts are derivable from physical facts. If they are derivable, then the cat status will follow from physical facts plus an indexical about where *I* am. If they are not derivable, then we need mental (e.g. phenomenal) truths in the inference base. I take the latter option, of course. Basically, on this version of the Everett view, cosmic hermeneutics is radically false, as the apparent "determinacy" in the world stems from the minds of observers and not from the world. But one still has derivability from the microphysical plus mental combined. (3) Collapse interpretation (on which there is just a wave function which "collapses" into a relatively definite state every now and then): This is probably the hardest case. The collapse will be in microphysics, so one will be able to know that the pointer or cat wavefunction has collapsed. But there is still a question of how one gets from even a collapsed wavefunction to macro truths -- e.g. from a macro object's wave function to facts about its position -- and whether this is a priori. The position etc of an object or particles will correspond to a certain eigenvalue, but it's not clear in what sense the mapping from microphysics to position is a priori, as it's not obvious in what sense taking the eigenvalue is part of the concept. Perhaps the easiest thing is to build this matter into the interpretation itself, so that the wavefunction is to be counted as specifying these positions etc via eigenvalues (that's part of what the wavefunction means). If one does that, then arguably one can get from the wave function to positions, velocities, etc, of various entities, and then eventually to macro determinacies, at least in some cases (cases where the wave function has collapsed appropriately, so that things actually have a determinate position, etc). The alternative is to say that we can't get this from the wave function alone, and that we need to introduce substantive "bridge principles" to get from wave function to macro position, etc. On the previous view, these bridge principles are a priori, but if that's denied, then they are substantive and a posteriori. On such a view cosmic hermeneutics will need to be denied. But arguably that's appropriate. The "thin" nature that we are building into microphysics on this interpretation arguably needs to be supplemented with further laws (these bridge laws) to determine a macro world. So the bridge laws should be taken as a further part of the theory. I have heard some people (e.g. Shelly Goldstein at Rutgers) argue for this explicitly, and some others seem to be implicitly committed to such a view. Personally, I prefer the former view, but I think it is a difficult and somewhat open question. The ontology of these collapse interpretations has not been worked out as well as it might be, I think. Overall: Clearly the answer depends on the interpretation. But I think we find a pattern: in the cases where the microphysics gives us a *complete* theory of the fundamental laws and such (e.g. Bohm, some collapse interpretations), cosmic hermeneutics is possible. In cases where microphysics is an *incomplete* theory, and where further fundamental laws are needed (e.g. some collapse and Everett interpretations), cosmic hermeneutics is impossible. But in each case, the macro truths (insofar as there are such truths) follow a priori from the fundamentals, whatever those are. Presumably, given an interpretation of QM, with all the relevant concepts, and with all the relevant facts, even Aristotle would be able to figure out the truths about macro objects (insofar as there are such truths) from the relevant facts. In the Bohm case it's easy; in the Everett case he may need facts about minds; in the collapse case he may or may not need further principles, depending on just what we give him as "built into" the microphysical concepts. But in each case some sort of macro picture should emerge, insofar as there is such a picture (of course, in some version of the Everett interpretation, there simply is no macro picture in the world as opposed to in the mind). Erik L. considers a case involving Schrodinger's cat, where an observation triggers a collapse in the cat's wave function at time t, so it goes from a superposition of alive and dead to being alive (let's say). So we're assuming a collapse interpretation here. Erik suggests that the facts that the cat is alive isn't entailed by the microphysics here. I think it's right that the cat's status isn't entailed by microphysics before time t. But that's not a problem for cosmic hermeneutics; rather, it's a problem for "cosmic number-crunching". In this case, Laplace's number-crunching demon can't even predict the future evolution of the wave function from the previous state. But the hermeneutic demon is given *all* the facts about the wavefunction across space and time, so this isn't a problem for him. Given all the microphysical facts (including facts after t), presumably examination of the post-t wave function will reveal that the cat is alive rather than dead. One will find particles with definite locations constituting a cat walking around, etc, with a beating heart, etc, rather than a cat lying down with no heartbeat, etc. At least, that's so modulo the issues about interpretation of collapse formalisms discussed above. But any such problems won't stem merely from indeterminate evolution between past and future states of the wave function, which the interpreting demon can take for granted. Erik suggests that given the possibility that CH fails for QM, one should be careful putting forward anti-materialist arguments given the failure of CH in the phenomenal case. I do agree that QM should make us cautious here. But I'd like to think that on a careful examination of QM, the interpretations on which CH fails are precisely those in which the relevant microphysics is an incomplete theory. So one still has an inference from failure of CH to expansion of ontology. Anthony notes another way of resisting the conclusion that QM falsifies CH. If in QM, particles have "fuzzy" location, that may be enough to give enough macro determinacy for our purposes. I think that's more or less right, though the details of this fuzziness depend a lot on the interpretation of QM. On the Bohm interpretation, the particles have definite interpretations and the wave function corresponds to our epistemic probabilities of finding them in certain definite places, so one has probabilistic access to the truth of the sort Anthony suggests. Of course in this case the demon will strictly speaking have access to the positions of the underlying particles, not just to the wave function, so he won't have problems. As Anthony notes, the demon is not constrained by our limitations of measurement -- he knows all the determinate facts! On other interpretations, things are a bit more complicated, as there really aren't particles with determinate positions in the general case. Here it's not just a problem of access to the facts, but of there not being facts. So one can't really even talk of "the probability that the particle is in position X". The wave function gives the full state of the particle, and that's that. But still, some of the time, the wave function will specify a definite position (as eigenvalue of position operator), which will help. And other times, it will specify a determinate range of positions (as a range of eigenvalues), which will presumably correspond to the sort of "fuzzy" location Anthony mentions. And maybe such micro fuzziness could add up to a sort of macro determinacy. There is a worry that some such fuzziness can be amplified at the macro level rather than cancelled out. That gets quite tricky, and again depends on the QM interpretation. On collapse interpretations, it will be the case that at least sometimes, the amplified fuzziness will be collapsed out, so in those cases we'll have a relatively determinate state. On Everett interpretations, the fuzziness may amplify indefinitely and forever (!), so there really won't be any determinacy to be found in the physical state alone; any determinacy will then have to be found in the mind. Anyway, plenty of issues here. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Thu Apr 8 01:53:46 1999 X-Priority: 3 X-MSMail-Priority: Normal X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Date: Thu, 8 Apr 1999 01:48:47 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Brad Thompson Subject: Does the CH-thesis have metaphysical significance? To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO In looking back over the Horgan and Byrne readings on cosmic hermeneutics, I am somewhat puzzled about the significance of the CH-thesis. It seems that Horgan is interested in the possibility of CH because it provides a way of capturing the physicalist claim that all facts are necessitated by the microphysical facts. Now, there is no *immediately* obvious contradiction in claiming that all the facts are necessitated (this is a metaphysical notion) by the microphysical facts and yet for a demon (an ideal reasoner) who knows all these microphysical facts to be unable to deductively infer from this knowledge all true propositions. I say there is no obvious contradiction in holding these two views because the former is about metaphysics and the latter is about epistemology. Nonetheless, this physicalism minus CH seems to create a gap between reality and our knowledge of reality (or at least its intelligibility) which is extremely unpalatable. The problem is that those who deny CH have to claim that we arrive at micro to macrophysical identity statements (I hope this is an ok way of characterizing "water = H20") in some other way than the analysis approach suggested by Chalmers, Jackson, and even Loar. The analysis approach renders the identity intelligible. But I suspect that any other way of arriving at such identities falls short of intelligibility. This came out of discussion of the Byrne article a couple weeks ago--the denial of an a priori entailment of the macro from the micro leads to a new skeptical hypothesis, such as the "problem of other chairs"! If I know all the microphysical facts of the thing I'm currently sitting on, I can still wonder "Is it a chair?" Here is a more specific way of bringing out this point. Suppose, as I think Block, Stalnaker, and Byrne would probably say, we arrive at identities through the observation of correlations. Chemists observe a correlation between "being a water sample" and "being an H20 sample" and conclude on grounds of parsimony perhaps that "water = H20". But if our reasons for believing in the identity claim is merely that they are correlated, then we can certainly raise the worry that the property of this stuff in virtue of which it falls within the (secondary) intension of "water" is a different property from the property in virtue of which it falls within the intension of "H20". I believe that Block and Stalnaker claim that identities don't need explanation. But they also seem committed to the view that identities can't be understood--they just have to be taken as brute. Physicalism minus CH seems to lead to the view that there are strong necessities everywhere, and that our hermeneutic demon would have to know the microphysical facts PLUS these strong necessities in order to deduce all the facts. This really seems outside the spirit of physicalism. Brad ____________________ Brad Thompson Department of Philosophy University of Arizona Tucson, AZ 85721-0027 From owner-modality@LISTSERV.ARIZONA.EDU Thu Apr 8 12:41:02 1999 Date: Thu, 8 Apr 1999 12:38:58 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Erik J Larson Subject: Re: Does the CH-thesis have metaphysical significance? To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO Brad, I'm going to pick on just a specific part of your message here; namely, your point that micro to macro identity statements might be had by taking note of correlations between properties. I would argue that mere correlations cannot possibly constitute sufficient grounds for identity. For example, there is a high correlation between the occurrence of a storm and a rapid fall in barometric pressure, but claiming an identity here would be absurd. However we arrive at identities, it can't be from just observed correlations, even if this constitutes an initial heuristic for focusing attention on certain properties over others. Erik On Thu, 8 Apr 1999, Brad Thompson wrote: > In looking back over the Horgan and Byrne readings on cosmic hermeneutics, I > am somewhat puzzled about the significance of the CH-thesis. It seems that > Horgan is interested in the possibility of CH because it provides a way of > capturing the physicalist claim that all facts are necessitated by the > microphysical facts. Now, there is no *immediately* obvious contradiction > in claiming that all the facts are necessitated (this is a metaphysical > notion) by the microphysical facts and yet for a demon (an ideal reasoner) > who knows all these microphysical facts to be unable to deductively infer > from this knowledge all true propositions. I say there is no obvious > contradiction in holding these two views because the former is about > metaphysics and the latter is about epistemology. > > Nonetheless, this physicalism minus CH seems to create a gap between reality > and our knowledge of reality (or at least its intelligibility) which is > extremely unpalatable. The problem is that those who deny CH have to claim > that we arrive at micro to macrophysical identity statements (I hope this is > an ok way of characterizing "water = H20") in some other way than the > analysis approach suggested by Chalmers, Jackson, and even Loar. The > analysis approach renders the identity intelligible. But I suspect that any > other way of arriving at such identities falls short of intelligibility. > > This came out of discussion of the Byrne article a couple weeks ago--the > denial of an a priori entailment of the macro from the micro leads to a new > skeptical hypothesis, such as the "problem of other chairs"! If I know all > the microphysical facts of the thing I'm currently sitting on, I can still > wonder "Is it a chair?" Here is a more specific way of bringing out this > point. Suppose, as I think Block, Stalnaker, and Byrne would probably say, > we arrive at identities through the observation of correlations. Chemists > observe a correlation between "being a water sample" and "being an H20 > sample" and conclude on grounds of parsimony perhaps that "water = H20". > But if our reasons for believing in the identity claim is merely that they > are correlated, then we can certainly raise the worry that the property of > this stuff in virtue of which it falls within the (secondary) intension of > "water" is a different property from the property in virtue of which it > falls within the intension of "H20". > > I believe that Block and Stalnaker claim that identities don't need > explanation. But they also seem committed to the view that identities can't > be understood--they just have to be taken as brute. Physicalism minus CH > seems to lead to the view that there are strong necessities everywhere, and > that our hermeneutic demon would have to know the microphysical facts PLUS > these strong necessities in order to deduce all the facts. This really > seems outside the spirit of physicalism. > > Brad > > ____________________ > Brad Thompson > Department of Philosophy > University of Arizona > Tucson, AZ 85721-0027 > "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 Thu Apr 8 16:37:33 1999 Date: Thu, 8 Apr 1999 16:35:08 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Brad J Thompson Subject: Re: Does the CH-thesis have metaphysical significance? To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I agree with what Erik says below, concerning the shortcomings of arriving at identities via observed correlations. That really was my point. Correlations seemed to be the best that Block, Stalnaker, and Byrne can do to establish identites, and that seems unsatisfactory. So how else can we arrive at identities if not via a priori conceptual analysis which connects the two terms? Brad On Thu, 8 Apr 1999, Erik J Larson wrote: > Brad, > > I'm going to pick on just a specific part of your message here; namely, > your point that micro to macro identity statements might be had by taking > note of correlations between properties. I would argue that mere > correlations cannot possibly constitute sufficient grounds for identity. > For example, there is a high correlation between the occurrence of a storm > and a rapid fall in barometric pressure, but claiming an identity here > would be absurd. However we arrive at identities, it can't be from just > observed correlations, even if this constitutes an initial heuristic for > focusing attention on certain properties over others. > > Erik > > > On Thu, 8 Apr 1999, Brad Thompson wrote: > > > In looking back over the Horgan and Byrne readings on cosmic hermeneutics, I > > am somewhat puzzled about the significance of the CH-thesis. It seems that > > Horgan is interested in the possibility of CH because it provides a way of > > capturing the physicalist claim that all facts are necessitated by the > > microphysical facts. Now, there is no *immediately* obvious contradiction > > in claiming that all the facts are necessitated (this is a metaphysical > > notion) by the microphysical facts and yet for a demon (an ideal reasoner) > > who knows all these microphysical facts to be unable to deductively infer > > from this knowledge all true propositions. I say there is no obvious > > contradiction in holding these two views because the former is about > > metaphysics and the latter is about epistemology. > > > > Nonetheless, this physicalism minus CH seems to create a gap between reality > > and our knowledge of reality (or at least its intelligibility) which is > > extremely unpalatable. The problem is that those who deny CH have to claim > > that we arrive at micro to macrophysical identity statements (I hope this is > > an ok way of characterizing "water = H20") in some other way than the > > analysis approach suggested by Chalmers, Jackson, and even Loar. The > > analysis approach renders the identity intelligible. But I suspect that any > > other way of arriving at such identities falls short of intelligibility. > > > > This came out of discussion of the Byrne article a couple weeks ago--the > > denial of an a priori entailment of the macro from the micro leads to a new > > skeptical hypothesis, such as the "problem of other chairs"! If I know all > > the microphysical facts of the thing I'm currently sitting on, I can still > > wonder "Is it a chair?" Here is a more specific way of bringing out this > > point. Suppose, as I think Block, Stalnaker, and Byrne would probably say, > > we arrive at identities through the observation of correlations. Chemists > > observe a correlation between "being a water sample" and "being an H20 > > sample" and conclude on grounds of parsimony perhaps that "water = H20". > > But if our reasons for believing in the identity claim is merely that they > > are correlated, then we can certainly raise the worry that the property of > > this stuff in virtue of which it falls within the (secondary) intension of > > "water" is a different property from the property in virtue of which it > > falls within the intension of "H20". > > > > I believe that Block and Stalnaker claim that identities don't need > > explanation. But they also seem committed to the view that identities can't > > be understood--they just have to be taken as brute. Physicalism minus CH > > seems to lead to the view that there are strong necessities everywhere, and > > that our hermeneutic demon would have to know the microphysical facts PLUS > > these strong necessities in order to deduce all the facts. This really > > seems outside the spirit of physicalism. > > > > Brad > > > > ____________________ > > Brad Thompson > > Department of Philosophy > > University of Arizona > > Tucson, AZ 85721-0027 > > > > > "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 > ---------------------- Brad J Thompson bradt@U.Arizona.EDU From owner-modality@LISTSERV.ARIZONA.EDU Sat Apr 10 16:35:29 1999 Date: Sat, 10 Apr 1999 16:35:17 -0700 Sender: "Philosophy 596B: Mind and Modality" From: David Chalmers Subject: Re: Does the CH-thesis have metaphysical significance? To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO I'm obviously in sympathy with what Brad wrote re cosmic hermeneutics. One of the problems with denying CH, I'd argue, is that it creates apparent epistemological problems where we had none before. If there are two different epistemic possibilities compatible with all the physical (and phenomenal) facts, it's not clear how we are to choose between them. At least we have a substantive skeptical problem. But on the face of it there is no such problem for chairs, tables, etc. (Of course there is such a problem relative to the phenomenal facts alone, but not relative to the physical and phenomenal facts.) There is an interesting question about how one gets to micro-macro identity without CH. One thing Block/Stalnaker et al might say is that one gets such an identity when one has a reductive explanation, but deny that reductive explanation is to be cashed out in terms of a priori entailment. Another thing would be to deny that reductive explanation is required at all, as that's an epistemic matter where identity is ontological. I would tend to agree with Brad that either way, we are left with "brute" identities that are very difficult to integrate with our view of the world and that border on being unintelligible. Block/Stalnaker et al would argue that these identities are perfectly intelligible once we get out of our a priori straightjacket. Combinations of empirical evidence plus considerations of simplicity can make such identities very plausible even where there is no transparent epistemic connection from micro to macro. One might say that the identity is the "best explanation" of the empirical evidence, including correlation. And they would argue that it is OK for an identity to be "brute" in this way, as identities don't need to be explained. They use "Samuel Clemens = Mark Twain" as an example here -- how would one explain that? But they'd say that nevertheless, such identities are perfectly intelligible. Now, I'd argue that for standard identities, some sort of epistemic entailment from a more fundamental base still holds (e.g. one could deduce that Mark Twain is Samuel Clemens from all the physical plus phenomenal facts), and this is what renders the identity there intelligible and "non-brute". Where the identity needed in the mind-brain case would seem quite "brute". But clearly B&S don't agree with either of these points. I don't think B&S need say that every correlation implies an underlying identity. They could deal with some cases (e.g. the ones Erik L. raises) by putting some conditions on them. E.g., one would at least require the correlated events to occur at the same point in space and time. And they would argue that it can't be the case that there is an obviously causal process mediating the relation between the two (as with storm and pressure). And so on. So maybe they could motivate identities in the water-H2O cases and mind-brain cases without being committed to every correlation implying identity. N.B. One note: Physicalism minus CH need not quite require strong necessities as defined (these involve positive conceivability without possibility), but they will at least lead to inscrutable truths (which come down to negative conceivability without possibility). I don't think inscrutable truths are quite as problematic as strong necessities, but they still have lots of problems, and in particular seem unsuited to play much of a role in epistemic matters such as as explanation. --Dave. From owner-modality@LISTSERV.ARIZONA.EDU Sat Apr 3 12:55:28 1999 Date: Sat, 3 Apr 1999 13:53:59 -0700 Sender: "Philosophy 596B: Mind and Modality" From: Anthony T Lane Subject: dogs To: MODALITY@LISTSERV.ARIZONA.EDU Status: RO In Dave's response to Angela re the capabilities of the "ideal conceiver", he suggests that, "...Arguably, I can conceive roughly of what it's like visually to be a dog, even though I'm not a dog." This seems to be accurate. Whereas we have rod and cone structures in our eyes, dogs only have rod structures (I think..). Given that we have good reasons for thinking that cones are responsible for our experience of color, we think that dogs see in black and white. It seems quite straightforward to imagine the visual experiences of a dog-- they are probably somewhat like the experience of seeing a black and white movie. It seems to me that it is somewhat more difficult to conceive of a dogs olfactory experiences. Avalanche rescue dogs are capable of quickly detecting the odor of people buried beneath considerable amounts of very hard snow (3 meters or more, sometimes). They are further able to distinguish the smell of the buried victim from the smell of the person handling them and other rescuers on the scene. I was once buried in a snowcave some six feet below the surface, and the rescue dog found me and had dug down to me in about 2 minutes. I don't have the faintest idea what it would be like to be able to make the kinds of olfactory discriminations that dogs are capable of making. It does not seem to be me to be particularly likely that even an ideal conceiver would be able to do this. I sems fairly straightforward to imagine a what a dog's visula experiences are like because we assume that their visual experiences are somehow less rich than our own. But, given that dogs have olfactory capabilities far superior to ours, itjust doesn't seem plausible to suggest that any non-canine being, however idealized and with whatever instrumentation you please, could conceive of what a dogs olfactory experiences are like. I'm not sure if all of this amounts to anything (except praise for the virtues of dogs). Anthony