In Defense of the Phenomenal Concept Strategy

Chair: Richard Brown

Presenter: Katalin Balog, Yale University

Kati’s paper

 

Commentator 1: István Aranyosi, Bilkent University -Ankara, Turkey

István’s paper & powerpoint slides. For a larger version of the video click here

 

Commentator 2: Esa Diaz-Leon, University of Manitoba -Winnipeg, Canada

 A larger version of Esa’s video

Advertisements

54 Comments

  1. Interesting reply, Istvan.
    I have a question about something you say in section III: “Since it is part of PCS that phenomenal concepts are not physical-functional concepts, the PCStrategist is well aware that there is no physical explicability of phenomenal concepts; otherwise they would be physical-functional concepts. ”
    I thought that perhaps we should distinguish these two issues: the PCS’s claim that phenomenal concepts and physical-functional concepts play different roles and therefore are not a priori connected, and the question of whether phenomenal concepts can be physically explained or not (that is, whether P entails C a priori).
    As I understand it, all versions of the PCS are committed to the first claim, but this leaves open what position they take with respect to the second issue. As Chalmers (2007) explains, some versions of the PCS claim P entails C a priori, and some others claim it doesn’t. But they all say that physical and phenomenal concepts come apart (this is what explains the conceivability of zombies).

  2. Hi Esa, and thanks for the kind words.

    What Chalmers says in that 2007 paper is that the account of Papineau, Balog, Hill, etc. can be interpreted either as (a) involving thick phenomenal concepts (i.e. when P does not a priori entail C), or as (b) involving thin ones (i.e. when P a priori entails C).

    I disagree about (b). Given that these people, like the ones I mention above, all think that phenomenal concepts are not conceptually (non-empirically) related to any physical concept, and given that the content of C involves reference to phenomenal concepts, C cannot a priori follow from P.

    Furthermore, there are people (Yablo, Block & Stalnaker) who argue that not even macro truths can be a prirori deducible from microphysical truths. They would certainly not like the idea that P a priori entails C.

    Thirdly, some of these people have a general lack of sympathy for apriorism, or even the very notion of apriority (e.g. they would say that there is no non-trivial a priori knowledge whatsoever), and certainly they wouldn’t want their theory to be interpreted as implying anything like a priori deducibility of C from P.

    This being said, you are right that, nevertheless, all these people think that their theory, contained in C, does actually EXPLAİN in a physicalistically acceptable way our epistemic situation. Whether they are right or not is another question, but what we are interested in is whether their view makes sense, and I find it perfectly coherent to say that P does not a priori entail C (or, for that matter, even that there is no a apriori knowledge whatsoever, except knowledge of propositions like [a = a]), but C explains in a physicalistically acceptable way some other phenomenon P.

    This is possible because part of C is that phenomenal concepts and physical ones corefer (to the extent that it is correct at all to say that consepts “refer”). That P and Q corefer cannot follow a priori from P, if Q is a phenomenal concept, because phenomenal concepts are inferentially disconnected from physical ones; yet “for any Q, if Q is a phenomenal concept, then there is a physical concept P, such that P and Q corefer” is simply a way to put the materialist thesis (for instance, Papineau in his 2002 book puts it this way, i.e. conceptual dualism plus ontological physical monism, or dualism of sense plus physical monism of reference), so it must be compatible with materialism, right?

    Chalmers’ problem here is that he wants to force this whole “type B” materialism into his own views about apriority and the “tight connection between explanation and apriority”. There is no such tight connection.

    Chalmers uses “explanation” and “explication” interchangeably in the 2007 article (he starts by using “explanation”, then without any warning switches to “explication”). An obvious counterexample to the thesis of explanation as explication is causal explanation: NO causal explanation is an explication, if causal laws are contingent. Explication, on the other hand, is a technical notion we owe to Carnap (1950). So what I want to say is that Chalmers’ “master argument” against the Phenomenal Concept Strategy is toothless, because: (i) C explains E, without P explicating C, (ii) C contains the information equivalent to the thesis of materialism (“for any Q, if Q is a phenomenal concept, then there is a physical concept P, such that P and Q corefer”), and so C is physicalistically acceptable.

    In my reply to Kati I don’t say anything about these issues, so it might seem that I also think that explanation is the same as explication, but that’s not true. The only bit I agree about with Chalmers is that C explains our epistemic situation E, only if C is not a priori deducible from P. But this is not to say that when C is not a priori deducible from P, C is not a physicalistically acceptable explanation of E.

    (By the way, there is a negation missing in my paper: (AC) should be read: “Either psi is not explanatory regarding our epistemic situation, or physicalism is false.”)

  3. Istvan, your commentary is very interesting and intricate. Here is what I think the crux of the matter is. Istvan’s argument, simply stated, is that physicalism is incompatible with the existence of phenomenal concepts of the sort I claim exist. Here is the argument in his own words:

    CPhen expresses in phenomenal vocabulary the fact that there are phenomenal concepts that have some features. Let G stand for the fact that there are phenomenal concepts that have some features. Let F stand for the fact that CPhen expresses G.

    …following Balog’s recipe…, there is a physical conceptualization of F. But F contains phenomenal vocabulary, and on the assumption of PCS that phenomenal and physical concepts are distinct and not a priori related, there can’t be a physical conceptualization of F. That is, there can’t be a physical conceptualization of a phenomenal conceptualization – that would destroy the phenomenality of that conceptualization. But given Balog’s implicit assumption that according to physicalism there is a physical conceptualization of any actual fact, the impossibility of a physical conceptualization of F entails the falsity of physicalism.

    There is a use-mention issue throughout these paragraphs. It gets confusing in the second paragraph: F doesn’t contain phenomenal vocabulary, “F” does. Or rather, on the physicalist view, a certain conceptualization of F, i.e. Fphen does. But according to the physicalist, there is also a physical conceptualization of the fact F, namely Fphen. Just because both Fphen and Fphys both express facts about phenomenal concepts, it doesn’t mean that they both have to employ phenomenal vocabulary to do so. If physicalism is true, and if the constitutional account of phenomenal concepts is true, then phenomenal concepts are constituted by phenomenal states, themselves physical or functional. There is no mystery then what types of states are or realize phenomenal concepts (i.e., physical states) and we can talk about these states in both physical and phenomenal vocabulary. So there is no difficulty for the physicalist here.

  4. Thanks for yor comment, Istvan. You say:
    “I disagree about (b). Given that these people, like the ones I mention above, all think that phenomenal concepts are not conceptually (non-empirically) related to any physical concept, and given that the content of C involves reference to phenomenal concepts, C cannot a priori follow from P.”
    I still don’t see why. I agree that those people might endorse Block and Stalnaker’s contention that microphysical truths don’t entail macroscopic truths a priori. They might even reject the notion of apriority. But if they don’t endorse B&S’s claim, and if they don’t reject apriority, then it is an open question for them whether P entails C a priori.
    The fact that C talks about phenomenal concepts does not entail that C is not a priori entailed by P, because C could describe phenomenal concepts in purely physical-functional terms. What C must say is that physical and phenomenal concepts are cognitively disconnected, but this can be expressed in purely physical-functional terms. As Chalmers says, even zombies could instantiate these versions of C: they could have corresponding physical structures which play the role of our phenomenal concepts (even if they don’t pick out genuine phenomenal kinds).
    I think a similar issue raises in connection to your reply to Balog (and in her response to you, as I understand it). You say:
    “CPhen expresses in phenomenal vocabulary the fact that there are phenomenal concepts that have some features. Let G stand for the fact that there are phenomenal concepts that have some features. Let F stand for the fact that CPhen expresses G. …following Balog’s recipe…, there is a physical conceptualization of F. But F contains phenomenal vocabulary, and on the assumption of PCS that phenomenal and physical concepts are distinct and not a priori related, there can’t be a physical conceptualization of F.”
    If I understand it, F contains phenomenal vocabulary in the sense that F is about phenomenal concepts. But as I say above, we can express facts about the special features of phenomenal concepts in purely physical-functional terms. So I don’t see why there cannot be a physical conceptualization of F. As Katalin says in her comment above:
    “Just because both Fphen and Fphys both express facts about phenomenal concepts, it doesn’t mean that they both have to employ phenomenal vocabulary to do so.”

  5. Hi Kati and Esa

    Short answer to your worry about my equating “containing phenomenal vocabulary” with “using phenomenal vocabulary”: in my reply, in the paper I offer two independent longish arguments why, given PCS, the use/mention distinction becomes irrelevant. If I’m right in those arguments, then there can’t be a physical conceptualization of the fact, F, that some conceptualization, C, is a phenomenal conceptualization of the fact, F*, that there are some phenomenal concepts with some features.

    Take, for instance, the following phrases: “Obama”, “The President”, and “That man”. Now take the sentences:

    (1) “Obama is tall”
    (2) “The President is tall”
    (3) “That man is tall”

    Given some fixed context, none of them a priori entails any other. Suppose Obama is tall. We can then say that (2) and (3) are a descriptive and a demonstrative conceptualization, respectively, of the fact that Obama is tall.

    Let’s denote by A the fact that Obama is tall.
    Let’s denote by B the fact that “The President is tall” is a descriptive conceptualization of the fact that Obama is tall.

    Note that by saying that “The President is tall” is a descriptive conceptualization of the fact that Obama is tall, we say that our words disquote, so “tall” means tall, “president” means president. Mutatis mutandis for (1) and (3).

    What we ask now is: can “That man is tall” a priori entail B?

    Suppose that “That man is tall” a priori entails B. How is this possible given that we agreed above that (2) doesn’t a priori entail either (1) or (3)? It’s pretty clear that these are incompatible.

    Suppose now that “That man is tall” does not a priori entail B. Following Kati’s recipe, this is compatible with “That man is tall” nevertheless modally (non-apriori) entailing B, only if there is a demonstrative conceptualization of B.

    What would a demonstrative conceptualization of B look like? I have no idea. It doesn’t seem to make sense for me. Is it something like this:

    “That conceptualization (pointing to: “The President is tall” is a descriptive conceptualization of the fact that Obama is tall.)”

    But this seems to a priori entail that the president is tall, which we don’t want from a DEMONSTRATİVE conceptualization; remember: “That man is tall” does not entail anything more than that some man is tall.

    I have to admit that I think, maybe naively, that conceptualizations of conceptualizations have to be of the same kind -there can’t be, for instance, a physical conceptualization of a chemical conceptualization of some fact.

    Returning to Esa’s worry about my being committed to not letting the PCS person have the option of her theory of phenomenal concepts be a priori entailed by physical truths, in a nutshell, I would say this. If we can talk about phenomenal concepts without using them, and therefore without getting out of the realm of physical and functional concepts, then zombies are conceived to do the same. If that is the case, then zombies are conceived as not being different from us at all when it comes to their explaining phenomenal concepts. But zombies are conceived as radically different from us when it comes to their explaining phenomenal concepts. Therefore, we can’t talk about phenomenal concepts in purely physical and functional terms.

    Further proof for this is Kati’s own (1999), and Papineau’s (2002) idea, that zombies are conceived as having something like “quasi-phenomenal concepts”. My question is: why the prefix “quasi”, if the theory of phenomenal concepts is to a priori follow from the physical and functional truths?

  6. Ok, a nice analogy came to my mind to convince you that there can’t be a physical conceptualization of a phenomenal conceptualization of some fact:

    There can’t be a photograph P that depicts in black-and-white the fact that another photograph P* depicts, in color, a purple sky. P destroys the colorfulness of P*.

    There can’t be a black-and-white TV screen S, on which I can see another TV screen, S*, which depicts, in colors, the purple sky. S destroys the colorfulness of S*.

    I hope this is convincing enough…

  7. Thanks Istvan. In response:
    “If we can talk about phenomenal concepts without using them, and therefore without getting out of the realm of physical and functional concepts, then zombies are conceived to do the same. If that is the case, then zombies are conceived as not being different from us at all when it comes to their explaining phenomenal concepts. But zombies are conceived as radically different from us when it comes to their explaining phenomenal concepts. Therefore, we can’t talk about phenomenal concepts in purely physical and functional terms.”
    I would deny that “zombies are conceived as radically different from us when it comes to their explaining phenomenal concepts.” If we are focusing on purely physical-functional accounts of phenomenal concepts, then zombies are not different from us regarding phenomenal concepts. They are different from us regarding phenomenal properties.

    “Further proof for this is Kati’s own (1999), and Papineau’s (2002) idea, that zombies are conceived as having something like “quasi-phenomenal concepts”. My question is: why the prefix “quasi”, if the theory of phenomenal concepts is to a priori follow from the physical and functional truths?”
    Answer: They are called quasi-phenomenal concepts because they don’t pick out genuine phenomenal kinds (since zombies lack these), but not because they lack any special features of phenomenal concepts as stated in C. These quasi-phenomenal concepts can possess all the special features stated in C (given that C is a physical-functional account).

  8. Istvan: I am not sure about your analogies. Isn’t the view defended by Balog suggesting something more like this? Suppose humans use images to represent features in the world (as Kosslyn suggests). It seems possible that one could completely describe the image, how it functions, etc., without using images in the description. This would be a non-imagistic “conceptualization” of an imagistic “conceptualization” of some property (e.g., shape).

  9. Esa:

    As I try to argue in the second part of section 5 of my paper, the problem with these theories, especially with Kati’s, is that they assert the existence of a super-tight connection between phenomenal concept and phenomenal property. Kati says the concept is constituted by the property.

    Now how can it be the case that, as you say: “zombies are not different from us regarding phenomenal concepts. They are different from us regarding phenomenal properties.”?

    How can zombies be conceived as different from us with respect to phenomenal properties, but without there being any difference between them and us regarding phenomenal concepts, if phenomenal concepts are simply constituted by the corresponding phenomenal properties? I would understand that zombies are conceived as exactly like us when it comes to phenomenal concepts if phenomenal concepts were picking out phenomenal properties via some other properties, so that those properties stay constant across our case and their case.

    Martin and Esa:

    I should make clear the following:

    – that I agree with Kati that there is a physical conceptualization of the story about phenomenal concepts
    – but that conceptualization does not a priori entail the existence of our phenomenal concepts (as opposed to zombie ‘phenomenal concepts’); it only entails them via a separate premise, Kati’s own premise that that conceptualization is a conceptualization of one and the same fact as what some phenomenal conceptualization conceptualizes.

    Here is an example. Take the phenomenal concept of PAIN, and say that it has feature F (F is assumed as topic-neutral). Call the story that there is a phenomenal concept of PAIN with feature F, S.

    Physical conceptualization of S: “there is a neural property N, such that F(N).”

    Phenomenal conceptualization of S: “there is a concept whose canonical deployment involves a feeling of pain, such that F(N).”

    The first one is a priori entailed by physical and functional truths, but it itself does not a priori entail anything that wouldn’t be present in a zombie, so it can’t serve as a story about our phenomenal concepts exclusively. It can only entail the existence of our phenomenal concepts, rather than the zombies’, if we add the proposition:

    “There is a neural property N, such that F(N)” and “there is a concept whose canonical deployment involves a feeling of pain, such that F(N)” conceptualize the same fact.

    Now we can deduce from the physical conceptualization that our phenomenal concepts exist.

    What about whether the physical truths a priori entail the fact that “there is a concept whose canonical deployment involves a feeling of pain, such that F(N)” is a phenomenal conceptualization of S?

    I say that they can’t entail it, because of the phenomenal vocabulary present in that conceptualization, and because we know from a few lines above that the physical conceptualization of S does not entail the existence of our phenomenal properties.

    If Kati is right, this lack of entailment could be made compatible with physicalism and explanatoriness if there was a physical conceptualization of the fact that “there is a concept whose canonical deployment involves a feeling of pain, such that F(N)” is a phenomenal conceptualization of S?

    (By the way, now I observe that Kati, by saying, implicitly, that whenever we have lack of a priori entailment from the physical truths to some other truth, we can only make that compatible with physicalism and explanatoriness if we find a physical conceptualization of the former, she is committed to the necessity of having a priori entailment in order to physically explain something, which brings me to wondering then how is her account different from an a priori materialist account; a true a posteriorist should say that there is no need to explain anything if by explanation you mean a priori account of something).

    But there can’t be a physical conceptualization of a phenomenal conceptualization of S (there can only be a physical conceptualization of S and a phenomenal conceptualization of S)

    This brings me to my analogy with pictures. Martin says that his example would be “a non-imagistic “conceptualization” of an imagistic “conceptualization” of some property (e.g., shape)”. I doubt it. I think it would just be a non-imagistic conceptualization of shape. Suppose we try to conceptualize the fact that some color image represents a green forest with packs of white goats. We conceptualize this by some sounds, or by zeros and ones, or whatever. What I say is that the result may well be that we perfectly conceptualize the green forest with the packs of white goats, but we have no idea that we conceptualized the COLOR image represented the forest the way it represented. To see this consider that there is also an infrared image representing the forest. How do you know that your zeros and ones, or your sounds, represent the color image rather than the infra-red image of the forest.

    Suppose you reply that your zeros and ones would be differently arranged if conceptualizing the infra-red image’s depicting the forest. But this is not a good reply for Kati’s paper, because then it shows , by analogy, that the phenomenal conceptualization of S is a priori entailed by physical conceptualization of the phenomenal conceptualization of S (whether something is a phenomenal or not conceptualization of S supervenes, if the reply is correct, on the physical conceptualization of these conceptualizations), and it simply shouldn’t: a physical conceptualization cannot a priori entail a phenomenal one, by the very theory of phenomenal concepts.

  10. Kati, Istvan & Esa,

    Kudos to all of you on a nice exchange. I have some questions and comments on Kati’s ‘constitutional account’ of the phenomenal concept strategy (PCS). I realize this isn’t addressing the main theme of her paper (i.e. Chalmers’ Master Argument) but I assume it’s fair game anyway. So what I have to say is mainly addressed to Kati, but needless to say I would have no objection if the others of you chimed in.

    So, Kati: It seems to me that on your constitutional version of PCS, a non-physicalist account of conscious experience straightaway follows (I’ll qualify that later). And given what I see as the similarities between your version of PCS and David Papineaus’ ‘quotational’ version, I would say the same of it too.

    Here are some passages from your paper that are relevant to my case. You say that, when one employs a phenomenal concept in refering to a consious experience, “the experience serves as its own mode of presentation” (p. 3). A little before that, you say of the relation of acquaintance that is implicated in the use of phenomenal concepts, “it is generally taken to be a unique epistemological relation that relates a person to her own mental states directly and, according to some, in a substantial way. I will understand ‘substantial’ grasp of a property as requiring the possession of a vivid mode of presentation of that property that strikes the thinker as essential to the property” (pp. 2-3) And I take the passage I quoted just before this to be saying that, in so striking the thinker, it is striking him/her right.

    Now why do I say this leads to non-physicalism? Well, a few years ago, Terence Horgan and John Tienson (H&T) wrote a very nice article entitled “Deconstructing New Wave Materialism” (Gillett & Loewer (ed.) “Physicalism and Its Discontents”) that contains the following short, sweet and to the point argument:

    1. When a phenomenal property is conceived under a phenomenal concept, this property is conceived otherwise than as a physical-functional concept.

    2. When a phenomenal property is conceived under a phenomenal concept, this property is conceived directly, as it is in itself.

    3. If (i) a property P is conceived under a concept C, otherwise than as a physical-functional property, and (ii) P is conceived under C, as it is in itself, then P is not a physical-functional property.

    4. Therefore, phenomenal properties are not physical-functional properties.

    I said this argument is “to the point”. But actually, not quite. H & T use it to target Brian Loar’s version of PCS (the version favored by Esa) but there is a way of reading Loar so that he can answer it. (For a defense of this way of reading him see Janet Levin’s article in the recent anthology by Alter and Walter). On this reading, H & T are making a mistake in premise 2 in assuming that ‘conceiving a property directly’ means ‘conceiving it as it is in itself’. But this isn’t so. Conceiving a property directly is just NOT conceiving it thru the mediation of some contingently related property–and that is all there need be to it. So on this reading, phenomenal concepts have NO MODE OF PRESENTATION at all. We are just wired to respond differently to, say, experiences of red, orange and blue; and make similarity and difference comparisons among them–i.e., in effect locate them in abstract color space. But we don’t do this in virtue of conceiving them ‘as they are in themselves’, or conceiving them via any mode of presentation. We just do it, we know not why (as it were); we conceive them as those properties (that are so related by similarity and difference) we know not what. On this reading, Loar also has the resources to answer the objection that he can’t distinguish between phenomenal color experiences and blindsight. (We are just wired to respond differently to instantiations of introspectively accessed properties that are phenomenal and those that are not, and maybe locate them in some kind of abstract space too. But that’s all)

    So Loar can reject premise 2 of the foregoing argument. But given the passages I earlier quoted from your paper, I don’t see how you can, and hence it looks like you are committed to non-physicalism. And I think the same is true of David Papineau’s quotational account.

    Interestingly, tho, it does look like your version of PCS leads naturally to a Russellian view: Physical theory just tells us about the causal, functional and other structural relations in which the fundamental properties of nature stand to each other. It doesn’t tell us about their intrinsic natures. Hence, the indirect and insubstantial nature of the concepts of physical theory. To know what the intrinsic properties are with which physical theory deals, and to know those properties in their essential nature, one has to have experience–which gives us phenomenal concepts with their directness and substantiality. Given H & T’s argument, it may seem that you are committed to a Russellian non-physicalism; even pan-psychism. But Daniel Stoljar has proposed (in a couple of 2001 papers) a way in which maybe the day can be saved for a kind of Russellian physicalism. (This is the qualification promised earlier).

    I have all this in the forefront of my mind right now, since I’m participating in the session on Russellian physicalism with Barbara Montero and Stoljar. (But no, this is not a sly way of getting all of you to rush over to that session and read all the papers; tho, of course, none of us would object if you did)

    So I guess that about covers it for now. I’m wondering what you (all of you) think.

    Emmett

  11. Istvan:

    You write:

    “Here is an example. Take the phenomenal concept of PAIN, and say that it has feature F (F is assumed as topic-neutral). Call the story that there is a phenomenal concept of PAIN with feature F, S.

    Physical conceptualization of S: “there is a neural property N, such that F(N).”

    Phenomenal conceptualization of S: “there is a concept whose canonical deployment involves a feeling of pain, such that F(N).”

    And then later:

    “But there can’t be a physical conceptualization of a phenomenal conceptualization of S (there can only be a physical conceptualization of S and a phenomenal conceptualization of S)”

    Isn’t the situation like this? The physicalist is going to say that the phenomenal conceptualization you discuss is something physical As such, there is a physical conceptualization of a phenomenal conceptualization of S (where ‘of a phenomenal conceptualization of S’ is read transparently). But the physical conceptualization will not a priori entail (3) that there are phenomenal concepts (read opaquely). If it did, then given Balog’s account of phenomenal concepts, it would turn out that zombies were not conceivable after all.

    The point, I take it, is that IF Balog’s account of phenomenal concepts is plausible, physicalists can explain why P does not a priori entail the existence of phenomenal concepts. So it can hardly be a knock against physicalism that P does not a priori entail phenomenal concepts.

    As applied to the picture and TV analogies, I think you are right: consider a color photo of David Chalmers. The photo is a physical thing and one can (in principle) describe it in wholly physical terms. Furthermore, what the photo depicts (Chalmers) is wholly physical (assume for the sake of argument). It seems possible that one could completely describe the photo in physical terms without using the *concept* of a color photo, or without being able to a priori deduce that the thing described was a color photo of David Chalmers. It does not follow that the physical description does not pick out the same thing that “a color photo of David Chalmers” picks out. The “gap” here entails no ontological gap.

    Am I missing something?

  12. Hi Emmett,

    Interesting question. I accept Chalmers’ characterization according to which the substantiality of phenomenal beliefs comes from the fact that via phenomenal concepts we are acquainted with phenomenal properties: we grasp them directly and in a way that appears to reveal their essence. The constitutional account explains this by suggesting that an experience instantiating the phenomenal property q provides a direct and essential mode of presentation for the phenomenal concept Cq by being constitutive of the relevant token of Cq.
    There is a delicate issue regarding the nature of acquaintance here. If phenomenal properties are, as the physicalist claims, physical or functional properties, then there is a clear sense in which acquaintance doesn’t reveal their nature. According to the constitutional account of phenomenal concepts, phenomenal judgments don’t reveal their referent as physical or functional, for the reason that they don’t analyze their referent in physical/functional terms. In this sense, contrary to the dualist view, they don’t reveal the nature of their referent. In another sense they do, since, in the canonical, introspective applications of phenomenal concepts, the very phenomenal (i.e., physical or functional) property that is being introspected serves as its own phenomenal mode of presentation. To avoid this equivocation, perhaps it would be better for the physicalist to analyze acquaintance and the substantiality of phenomenal beliefs in terms of the phenomenal *presence* of the introspected properties in phenomenal judgments; and not in terms of our *direct grasp of the essence* of phenomenal properties. This is a characterization of acquaintance that physicalists and dualists can agree about.

    Kati

  13. Esa,

    Thanks for your comments. I agree with most of the things you say. You are right to bring up up the issue of how the degree of abstraction is fixed in phenomenal concepts. There are several ways to go about them; Loar’s recognitional account is one, but there are other ways to appeal to the functional roles of phenomenal concepts to play a role in determining reference.

    However, these approaches are not at odds with the constitutional account, they are *part* of such an account. Btw, Loar’s account can be interpreted as the Ur constitutional account. He talks about the mode of presentation being the referent itself…

    K

  14. Martin says: “The point, I take it, is that IF Balog’s account of phenomenal concepts is plausible, physicalists can explain why P does not a priori entail the existence of phenomenal concepts. So it can hardly be a knock against physicalism that P does not a priori entail phenomenal concepts.”

    And Kati agrees and says: “Your answer to Istvan’s concerns is exactly how I would have put it – (well, perhaps a bit more eloquent). I have nothing to add at this point.”

    I reply: you could read my paper, and see that i never say that physicalism comes out as false because P doesn’t a priori entail Q. I’m not going to repeat the arguments I put forward in sections 4 and 5 of the paper, or in the comments above, as they are pretty clear, and so far no one really addressed them…

  15. OK, as a last ditch effort, here is a simplified version of the main argument in my paper:

    1. Given the way Kati replies to Chalmers’ “master argument”, she believes that something X can be both physicalistically acceptable and explanatory of our epistemic situation only if there is a physical conceptualization of X.
    2. A physical conceptualization of the fact that there are phenomenal conceptualizations of some facts is impossible. (see my picture and TV analogies)
    3. According to Kati, if physicalism is true, then there is a physical conceptualization of every actual fact.
    4. Therefore, Kati should admit that physicalism is false.

    So, as you see, everything depends on my premise 2; if the analogy with pictures and TV sets I offer in my comment is right, then 2 is true. The objection appealing to use/mention is a red herring, or at most it proves again that 2 is true (thats what try to prove in section 5).

  16. Kati and Emmett:

    Have you had the chance to look at Michael Tye’s new book (Consciousness Revisited)? It just came out, and while I have not read it yet, I had a conversation with Tye about it last year. From what I can remember (and understand!), Tye’s new view abandons the appeal to phenomenal concepts to explain, e.g., Mary’s cognitive gain post-release. He argues that her new knowledge is knowledge by acquaintance (a la Russell), and that such knowledge does not require deploying concepts of any sort. Tye’s twist, however, is that such acquaintance does not “reveal” the essential nature of properties like redness, and thus it is perfectly compatible with his acquaintance view that the property one is acquainted with is physical.

    Given that Tye previously championed phenomenal concepts, one may wonder why he gives them up (I did not get into that part with him, though he told me that much of the book consists in attacking the appeal to phenomenal concepts). It may turn out that Kati’s account of phenomenal concepts survives his objections, however.

  17. Hi, and excuse me for jumping in mid-discussion here. This is directed to Kati’s reply to Emmet.
    I want to hear more about this passage:

    ‘phenomenal property q provides a direct and essential mode of presentation for the phenomenal concept Cq by being constitutive of the relevant token of Cq.
    There is a delicate issue regarding the nature of acquaintance here. If phenomenal properties are, as the physicalist claims, physical or functional properties, then there is a clear sense in which acquaintance doesn’t reveal their nature.’

    What’s the precise force of ‘essential’ in ‘essential mode of presentation’ here? From the context, it doesn’t mean that the essence of the property is involved in picking it out. But what does it mean then?

    Relatedly: If Cq doesn’t pick q out via an essential feature of q, what is the status of the (for example) experiential redness that one tokens when thinking about experiential redness? This is certainly some kind of property or nature in its own right. If it’s just a non essence-disclosing feature of a physical/functional property, then it would seem that we have something new to worry about: how is the non-essential feature of the physical/functional property (that Cq really refers to) to be conceptualised and accommodated within a physicalist framework?

    (btw Emmet I’m in the middle of writing a paper to the effect that the phenomenal concept strategy implies panpsychism!)

  18. Istvan:

    At this point I am almost willing to concede that I am massively misunderstanding your argument; if so, I apologize for my obtuseness and thank you for your patience.

    If you will bear with me, however, perhaps you can point out the specific flaw in the following argument. I will admit up front that I am putting matters in a slightly different way that you have, and this difference may make all the difference. If so, I’d like to know why.

    (1) Phenomenal conceptualizations are cognitive states
    (2) Cognitive states are functional/physical states (physicalism)
    (3) If cognitive states are functional/physical states, then there exist functional/physical conceptualizations of cognitive states
    (4) Therefore, there exist functional/physical conceptualizations of cognitive states
    (5) Therefore, there exist functional/physical conceptualizations of phenomenal conceptualizations

  19. Martin:

    I don’t see how (5) follows.

    Second, I was a bit sloppy before, when formulating the simplified version of my argument. Here is the real version:

    1. Given the way Kati replies to Chalmers’ “master argument”, she believes that PCS can be both physicalistically acceptable and explanatory of our epistemic situation only if there is a physical conceptualization of X.

    2. A physical conceptualization of the fact that there are phenomenal conceptualizations of some facts is impossible. (see my picture and TV analogies)

    3. Therefore, if Kati accepts PCS, she should admit that Either physicalism is false, Or PCS is not explanatory of our epistemic situation.

  20. Istvan:

    I tried to make the argument I gave more explicit

    (1) If C is a phenomenal conceptualization, then C is a cognitive state (premise)
    (2) If C is a cognitive state, then C is a functional/physical state (premise)
    (3) Therefore, if C is a phenomenal conceptualization, then C is a functional/physical state (from (1) and (2), hypothetical syllogism)
    (4) If C is a functional/physical state, then there exists a functional/physical conceptualization of C (premise)
    (5) Therefore, if C is a phenomenal conceptualization, then there exists a functional/physical conceptualization of C (from (3) and (4), hypothetical syllogism)

  21. Thanks, Martin

    I would reply to your argument as follows.

    If all cognitive states are physical/functional, then I have reason to doubt that phenomenal conceptualizations are cognitive states, because I accept that all physical/functional states have physical/functional conceptualizations, and I have independent reasons to think that physical/functional conceptualizations of phenomenal conceptualizations are a contradiction in terms.

    If phenomenal conceptualizations are cognitive states, then I have reason to doubt that all cognitive states are physical/functional states, because I accept that all physical/functional states have physical/functional conceptualizations, and I have independent reasons to think that physical/functional conceptualizations of phenomenal conceptualizations are a contradiction in terms.

    The issue is, again, whether my independent reasons for thinking that physical/functional conceptualizations of phenomenal conceptualizations are a contradiction in terms are acceptable.

    The reasons are:

    The analogy with descriptive and demonstrative phrases: if some phrase is demonstrative, then you can’t express the fact that it is demonstrative by non-demonstrative, descriptive phrases, even though you might be able to descriptively express what the demonstrative expresses. If you look at the literature on indexicals, even this much –that you might be able to descriptively express what the demonstrative expresses– is dubious for many theorists (by the way, that’s why Chalmers introduces indexical information besides P, and argues that everything follows from P and Indexical info, except consciousness), so my thesis (that you can’t express the fact that some phrase is demonstrative by non-demonstrative vocabulary) should really be obvious.

    The TV or photo analogy: if the thesis (that physical/functional conceptualizations of phenomenal conceptualizations are possible) were right, all people in the USA should be able to enjoy High Definition TV experience on their normal TV screens. All that is needed would be that one person in the USA has a HD TV, a video camera is recording what is going on in its screen, and the rest of the people are watching the screen of that HD TV on their own non-HD screens. I hope it is absurd enough to say now that these people would be enjoying a HD experience.

  22. Hi Istvan,

    Thank you for your comments; they have been quite thought provoking.

    Here is why I am still not convinced. Your analogies suggest that one can give a physical conceptualization of a phenomenal conceptualization only if the physical conceptualization includes or entails the phenomenal conceptualization. Since it doesn’t, one can’t. To motivate this, you point to demonstratives. You write:

    “if some phrase is demonstrative, then you can’t express the fact that it is demonstrative by non-demonstrative, descriptive phrases, even though you might be able to descriptively express what the demonstrative expresses.”

    But now consider the following example and line of response:

    A: “That is tall”

    1. ‘that’ is a demonstrative
    2. A demonstrative is a pronoun or adjective which points out which item is being referred to.
    3. Therefore, ‘that’ is a pronoun or adjective which points out which item is being referred to.
    4. ‘that’ is the first word in A
    5. Therefore, the first word in A is a pronoun or adjective which points out which item is being referred to.

    I am tempted to say that “The first word in A is a pronoun or adjective which points out which item is being referred to” is a demonstrative-free description that expresses the fact that ‘that’ is demonstrative.

    Now, in an earlier post, you wrote: “I have to admit that I think, maybe naively, that conceptualizations of conceptualizations have to be of the same kind -there can’t be, for instance, a physical conceptualization of a chemical conceptualization of some fact.”

    But consider this:

    A: “there is H2O in my glass”

    A is a ‘chemical conceptualization,’ if you will

    B: ‘Sentence A uses a concept that is important in science”

    B is a conceptualization of a chemical conceptualization, but B is not a chemical conceptualization.

    So it looks like you are thinking that we can physically conceptualize phenomenal conceptualizations only if the physical conceptualization conceptualizes the phenomenal conceptualization qua phenomenal conceptualization.

    Generalizing, this says that we can physically conceptualize X only if the physical conceptualization conceptualizes X qua X.

    This implies that we can physically conceptualize water only if the physical conceptualization conceptualizes water qua water.

    Does not this imply that in conceptualizing something as H2O, one fails to conceptualize water? And isn’t this false?

  23. Hi Martin and thanks for addressing the real arguments,

    Your examples might point to a misunderstanding regarding the concept of conceptualization. In Kati’s paper to conceptualize something means to express the same thing in a different way.

    Let’s start with your second example:

    ——————————————————-
    But consider this:

    A: “there is H2O in my glass”

    A is a ‘chemical conceptualization,’ if you will

    B: ‘Sentence A uses a concept that is important in science”

    B is a conceptualization of a chemical conceptualization, but B is not a chemical conceptualization.
    ——————————————————–

    It is clear that B is not a conceptualization of A, because it does not express the same thing as A.

    Let’s take your first example:

    ——————————————————–
    A: “That is tall”

    1. ‘that’ is a demonstrative
    2. A demonstrative is a pronoun or adjective which points out which item is being referred to.
    3. Therefore, ‘that’ is a pronoun or adjective which points out which item is being referred to.
    4. ‘that’ is the first word in A
    5. Therefore, the first word in A is a pronoun or adjective which points out which item is being referred to.

    I am tempted to say that “The first word in A is a pronoun or adjective which points out which item is being referred to” is a demonstrative-free description that expresses the fact that ‘that’ is demonstrative.
    ———————————————–

    First, your 5. does not express A. Second, maybe the following would express A (but in fact it doesn’t, see below):

    “The first word in A is a pronoun or adjective which points out which item is being referred to and is the subject in A of a sentence which expresses the fact that the item referred to by the first word is tall”

    This could express A on the condition that “The first word in A is a pronoun or adjective which points out which item is being referred to” is a demonstrative-free description that expresses the fact that ‘that’ is demonstrative, as you say.

    But the latter could only be the case if “the first word in A” is interchangeable –that is, always interchangeable– salva veritate with “that”, which is not the case. Conversely, if (for the sake of reductio) we assume that “the first word in A” is interchangeable salva veritate with “that”, then it follows that “the first word in A” is a demonstrative phrase. We know that the latter is false, so “the first word in A” is not interchangeable salva veritate with “that”.

    To see what would follow if your examples were to show what you propose them to be taken to show, consider, as you seem to suggest, that “the derivation of the solution to one of the longest equations in Hartle-Hawking quantum cosmology” could be taken as a conceptualization of the set of sentences which constitute the derivation of the solution to one of the longest equations in Hartle-Hawking quantum cosmology. Then it follows that all of us are just the brightest physicists in the world.

  24. Hi everyone, really interesting discussion going on in here!

    István , I wonder if you are interpreting the conceptualization claim too strictly. So, for C to be a conceptualization of P;

    C must express the same thing as P in a different way.

    You take this to mean that C and P must be interchangeable salva veritate in all extensional contexts but can’t one interpret the ‘in a different way’ as introducing a kind of intensional context? So, ‘superman can fly’ and ‘Clark Kent can fly’ express the same thing, but in different ways; i.e. by using different modes of presentation.

    If this is right then there is nothing wrong with Kati’s response to the master argument. there is something wrong with her alligience to the view that phenomenal properties serve as their own modes of presentation. If one dropped that they could help themselves to the basic strategy that Kati here develops, no?

  25. For Martin:

    Finally, since it is ambiguous whether A is the sentence “That is tall” or just names the sentence, here is an argument for how in both cases it leads to trouble:

    1. “The first word in A” is either interchangeable salva veritate with “that” (in extensional contexts), or not.

    2. If it is, then “The first word in A” is a demonstrative conceptualization.

    3. If it is not, then “The first word in A” is not a conceptualization of the fact that “that” is a demonstrative.

    4. Either “The first word in A” is demonstrative or it not a conceptualization of the fact that “that” is a demonstrative.

    5. Gneralizing, there is no non-demonstrative conceptualization of the fact that “that” is a demonstrative.

  26. Hi Richard,

    By saying interchangeable salva veritate in all extensional contexts, I intended to say something mild; it is much stronger to say interchangeable salva veritate in all intensional contexts, and I don’t want to say that, because that would mean, for instance, that phenomenal concepts are deducible from physical ones –whenerver someone believes P (physical truth), we can replace P with Q (phenomenal truth) in the belief report. By limiting the thesis to extewnsional contexts, we make it possible for P to be conceptually independent of Q.

    To take your example, ’Superman can fly’ and ‘Clark Kent can fly’ are interchangeable salva veritate in all extensional contexts, but not in intensional contexts.

  27. Thanks for the quick response István!

    Right, I see that your argument only works if the context is purely extensional but in the superman/Clark Kent example there are non-extensional contexts were we cannot substitue salva veritate even though they express the same thing. And isn’t it the case that what forces us to take the extensional view in Kati’s case is her claim that phenomenal properties are their own modes of presentation? So if we give that up then your argument against her fails. Thus your argument can be interpreted as targeting the claim that phenomenal properties are their own modes of presentation and *not* Kati’s general strategy for defending the phenomenal concept strategy.

  28. Hi all. Interesting discussion. I can’t address everything that has come up here, but it’s worth pointing out two relevant aspects of my paper that don’t seem to have been taken into account in the discussion above.

    One point (relevant to Kati’s paper and some of Istvan’s comments above) is the crucial requirement that C, the account of phenomenal concepts, be given in topic-neutral terms. Of course, as Kati and Istvan observe (and as I observe in the paper), if C is given in phenomenal terms (so that it implies phenomenal truths), the explanatory gap between P and Q will immediately yield an explanatory gap between P and C. I take it that here, what theorists such as Loar and Papineau do is appeal to general structural features of phenomenal concepts that can be characterized topic-neutrally: epistemological and semantic properties for Loar, syntactic/realizational properties for Papineau. Once C is characterized this way, it’s no longer trivial that there’s an explanatory gap between P and C. I take it that the promise of such a topic-neutral account with no explanatory gap either between P and C or between C and E is the source of much of the appeal of the phenomenal concept strategy. But this is also what I argue we can’t have.

    Kati doesn’t address the issue of a topic-neutral account in her paper, preferring to focus on a phenomenal version of C. It’s not clear to me whether Kati thinks that there is a topic-neutral version C* of her preferred quotational account of phenomenal concepts (I don’t see why there shouldn’t be), and if so whether she thinks that there is an explanatory gap between P and C*, between C* and E, or both or neither. If there is an explanatory gap between P and C*, for example, I think that this is a much more serious lacuna for the PCS proponent than the trivially predictable explanatory gap between P and (phenomenally conceived) C, as now it appears that the key features of the explanation (quotationality, etc), which may have appeared physicalistically unproblematic, are now just as problematic as the original explanandum. And if there is an explanatory gap between C* and E, then these features don’t seem to be doing much of a job of explaining.

    It may be that Kati doesn’t care about a topic-neutral version of C and thinks that the important version is the phenomenal version. Here I’ll just say that I think that this version of the PCS is much less powerful than the topic-neutral version (which I find in Loar and Papineau), precisely because of the appeal to unexplained phenomenal features in the explanation. The dialectic at this point (regarding circularity and the like) gets complicated, though, so this is enough for now.

    The other point (relevant to some of Esa’s and Istvan’s points) is my discussion of the link between conceivability and explanation, and of option 4 which rejects this link. Esa at the end her comments appears to take a version of this strategy (applied to the C/E gap), but doesn’t discuss my reply. One relevant point is that if one takes this strategy for the P/C or C/E explanatory gaps, the strategy is equally available in response to the P/Q explanatory gap, thereby rendering the PCS somewhat redundant.

  29. Thanks, Istvan, for the clarification.

    First point. In my demonstratives example, I believe extensional equivalence is preserved. The relevant phrases are ” ‘that’ ” and “The first word in A”, not “that” and “The first word in A.”

    1. ‘that’ is a demonstrative phrase
    2. The first word in A is a demonstrative phrase

    These are extensionally equivalent, since both ” ‘that’ ” and “The first word in A” refer to ‘that’.

    Second point. It seems to me that we’ve been vacillating between two ways to understand what is required to physically conceptualize a phenomenal conceptualization. In light of your latest posts, I understand that what is needed is extensional equivalence. Let’s see if I follow.

    Consider these two sentences:

    (a) “there is water in my glass”
    (b) “there is H2O in my glass”

    (a) and (b) are extensionally equivalent, and extensional equivalence is necessary and sufficient for (b) to be a conceptualization of (a).

    Earlier, you gave the following example of physical and phenomenal conceptualizations:

    Physical conceptualization of S: “there is a neural property N, such that F(N).”

    Phenomenal conceptualization of S: “there is a concept whose canonical deployment involves a feeling of pain, such that F(N).”

    But if physicalism is true, then it very well may be that ‘a neural property N’ and ‘a concept whose canonical deployment involves a feeling of pain’ are extensionally equivalent and thus the two conceptualizations are extensionally equivalent. But then it turns out that we can give a physical conceptualization of a phenomenal conceptualization. The only way to deny it is to beg the question against physicalism.

    Note, too, that if extensional equivalence is all that is needed, then your analogies are not correct, for those analogies require intensional equivalence. The correct analogy would be that we can use both “there is a color photo of David Chalmers” and a color photo of a color photo of David Chalmers to express the fact that there is a color photo of David Chalmers.

    Finally, you write:

    “To see what would follow if your examples were to show what you propose them to be taken to show, consider, as you seem to suggest, that “the derivation of the solution to one of the longest equations in Hartle-Hawking quantum cosmology” could be taken as a conceptualization of the set of sentences which constitute the derivation of the solution to one of the longest equations in Hartle-Hawking quantum cosmology. Then it follows that all of us are just the brightest physicists in the world.”

    Yes, it is a conceptualization of the set of sentences, just as uttering “there is a way to play great cello” while pointing at Yo-Yo Ma is a way to conceptualize a way to play great cello. It is not, perhaps, a very useful way to conceptualize playing cello; it certainly won’t make me a great cello player. So, too, being able to conceptualize the derivation in this way may not entail that I can give the solution, or that I would even recognize the derivation if it was presented to me.

  30. Speaking of David’s paper, I’ve recently wondered whether the defender of the PCS strategy could respond like this:

    In explaining our epistemic situation, we do not mean every aspect of our situation. We are explaining why there is a gap between physical truths and phenomenal truths. The answer has to do with the special nature of phenomenal concepts, most importantly, the fact that they do not have a priori associated functional/physical reference fixers. If P–>C is a priori, then of course it is not conceivable that zombies fail to share the relevant epistemic situation with us–there will be an epistemic gap between physical and phenomenal concepts for them, too (where phenomenal concepts are understood in topic-neutral terms). David’s reply is that there will be significant epistemic differences between us and zombies still, so the strategy doesn’t work. But will there be a difference vis-a-vis the epistemic gap between physical and phenomenal concepts? I don’t see how. A zombie will possess a phenomenal concept C that lacks an a priori associated functional/physical reference fixer. So, e.g., ‘if P, then C = neural state N’ will not be a priori for the zombie, even though C picks out neural state N. The reply to this appears to be that, among other things, if C makes the zombie directly acquainted with neural state N, it will be trivial to determine that C = neural state N. But why think that direct acquaintance with neural state N would make determining the identity trivial? The thought seems to be that if I am directly acquainted with X, then I know all the properties of X, and so direct acquaintance with neural state N would reveal all the neural properties. But why think this? It seems plausible to me to suppose that if direct acquaintance of a neural state did reveal its neural properties to a zombie, the zombie would know them as neural properties only if it had the relevant neuro-physiological concepts. But are we supposed to imagine that all zombies have mastery of neuroscience? And if zombies can be acquainted with neural state N without knowing the neural properties (using neural concepts), then why would determining identity be trivial?

  31. Kati & Martin,

    Things can get confusing with at least two different lines of discussion going on at once, so to orient you and anyone else who is tuning in, I’m resuming the line I initiated with entry 10 and that you responded to in entry 12 (Kati) and entry 17 (Martin). (Tho I now see that the numbers of the entries only show up if you print this out. Well, all the entires came in on Feb. 23) This concerns the issue of why Kati’s constitutional account of PCS doesn’t lead to non-physicalism.

    To Martin:

    No, I haven’t read Michael Tye’s new book, but thanks for alerting me (us) to it.

    To Kati:

    Well, your proposal walks a fine line. ‘Presence’ but not ‘direct grasp of essence’. What do we make of that?

    Let me review the issue as I see it: The problem with Brian Loar’s version of PCS is that, in order to answer objections like those raised by H & T, it has to render phenomenal concepts phenomenologically vacuous. (See entry 10 for meaning of abbreviations). Or, as you put it, the phenomenal concepts lack substantiality. Daniel Dennett might not see anything wrong with that, but I assume you do and that your constitutive version of PCS is designed to remedy that situation. But (to give Dennett his due) it is not easy to explain exactly WHAT it is that phenomenal concepts ought to have that Loar’s phenomenal concepts lack. Or at least it is hard to do so from the phenomenological perspective. From the neurological or functional perspective, your constitutive account probably captures the intuitions about as good as any neurological of functional account could. But what is it about the phenomenology that this account is supposed to explain?

    Well, whatever it is, one assumes it is something that has epistemic import. After all, focusing on the Knowledge Argument as an example, it’s the fact that post-release Mary acquires new knowledge (albeit, so the argument runs, not knowledge of new facts) that PCS is trying to explain. So what is it that phenomenal concepts are supposed to have? Is it that, in virtue of having phenomenal concepts, one can recognize, or come to know that, one is having a sense experience of red (or blue or orange) upon having one? But Loar’s version can account for that. So then what does Loar’s version leave out? The idea that phenomenal concepts allow us to have a direct grasp of the ESSENTIAL NATURE of sense experiences surely strikes one as intuitively right. But then H & T’s argument for non-physicalism goes thru.

    You suggest that we understand phenomenal concepts so that they yield the PRESENCE of introspected phenomenal properties in phenomenal judgments, but not a DIRECT GRASP OF THEIR ESSENTIAL NATURE. But what is the epistemic import of such presence?

    Suppose we try grafting the idea of an epistemically irrelevant presence onto Loar’s account and see if this helps. In virtue of having phenomenal concepts, one can recognize, identify, locate in abstract color space, etc. color experiences (e.g.). In addition, there is this epistemically irrelevant ‘presence’. (I suppose we could allow it to have some causal power) But then, as far as the content of what we know is concerned, the presence may as well not be there. (The presence need not be present, as it were) So then we are effectively back to Loar’s original view.

    Now I don’t (of course) assume that this settles the issue. Indeed, other possible physicalist responses are forming in my head even as I write this. But I want to give these time to form more fully, and besides, we are not supposed to allow these entries to get too long. So I’ll save more of what I have to say for later, and in the meantime see it you (or Martin, or anyone else who is reading this) have anything to say.

    Emmett

  32. Sam,

    I just now read your recent entry. I’ll be looking forward to seeing how your argument for the panpsychist implications of the PCS goes: If it’s a more detailed argument of what I briefly sketched in my earlier entry. And stay tuned for my next entry where, depending on how my thinking plays out on this, I may bring in a Rusellian perspective.

    Also, you may want to read my article in the May, 2008 issue of “Journal of Consciousness Studies” (vol. 15, # 5) in which I argue that the way you deal with the subject combination problem (in the article you wrote in response to Galen Strawson) better situates you as a Russellian physicalist than a Russellian panpsychist. The argument comes in a paragraph on p. 63, but to understand it, you need to have gone thru the argument that leads up to it.

    Emmett

  33. Martin:

    OK, in the paper I make it clear that the fact that C conceptualizes F is not to be confused with C itslef. Of course, when I say there are no physical conceptualizations of phenomanal conceptualizations that is shorthand for:

    There is no conceptualization C , having kind K, of the fact that some C* is a conceptualization, having kind K*, of some fact F, if K is not identical to K*.

    Of course, I don’t say:
    There is no conceptualization C , having kind K, of some C*, which is itself a conceptualization, having kind K*, of some fact F, if K is not identical to K*.

    That I don’t say this you can see in my paper, where I accept that are both phenomenal and physical conceptualizations of the same story S about phenomenal concepts. All I say is that it is contradictory to give a story using only physical concepts, which story also uses phenomenal concepts, just like it is contradictory to draw a circle by using only straight lines in normal plane. What can at most be done is to have the phenomenal vocabulary quoted (mentioned) in such a physical conceptualization, but that is not about phenomenal stuff any more, because, as we know already from Frege, the rererence of the quoted terms is the term, not the customary reference, and the sense is “the sense of “T””. Such a physical conceptualization is about words, and could be equally offered to us by a zombie, so it is not a conceptualization that would physically express that something phenomenal can be used to express some truth – there is nothing phenomenal about the word “pain”, but there is a lot about the PAIN (concept).

    So phenomenal vocabulary has to be used, not mentioned. But if it is used, then the conceptualization is not wholly physical.

    Here is by way of argument:

    1. If phenomenal vocabulary is used in S, then S is a phenomenal conceptualization.
    2. If phenomenal vocabulary is mentioned in S, then S is not a phenomenal conceptualization.
    3. A physical conceptualization is not a phenomenal one.
    4. A conceptualization C of the fact that S is a phenomenal conceptualization of F either uses or mentions phenomenal vocabulary.
    5. If it uses, then C is phenomenal.
    6. If it mentions, then C is not about S being a phenomenal conceptualization of F.
    7. Therefore, there can’t be a physical conceptualization of the fact that S is a phenomenal conceptualization of F.

    This answers your examples. You say:

    [1. ‘that’ is a demonstrative phrase
    2. The first word in A is a demonstrative phrase
    These are extensionally equivalent, since both ” ‘that’ ” and “The first word in A” refer to ‘that’.]
    Reply: “That” occurs in 1. not as a demonstrative, so 2. is not the relevant type of conceptualization we are after.
    Then you say:
    [(a) “there is water in my glass”
    (b) “there is H2O in my glass”
    (a) and (b) are extensionally equivalent, and extensional equivalence is necessary and sufficient for (b) to be a conceptualization of (a).]

    Reply: This is true, but irrelevant, because I accept it, just like I accept (I assume it in my paper, though I don’t myself believe it or disbelieve it, as PCS assumes it): “I feel pain” can be conceptualized as “P”, where “P” is physical/functional. To repeat, what I don’t accept is, supposing WATER and H2O are not a priori connected, that you could give a story in water terms about the fact that a chemist expresses truths about the world in terms of the concept H2O. That is contradictory, but nevertheless it is needed to be true, if PCS, and especially Kati’s reply to Chalmers, is to be coherent (remember that Kati thinks that if physicalism is true, then every fact can be physically conceptualize – I say, hey, here is a fact that is impossible to physically conceptualize: that PAIN is a phenomenal conceptualization of some fact about the world, or, if you want, that a thought deploying PAIN is a phenomenal conceptualization of some fact about the world)

  34. Emmett thanks for that, I’ll get on to that article asap.

    I have to confess though, I’ve never properly grasped the difference (distinction due to Stoljar right?) between a Russellian physicalist and a Russellian panpsychist…

  35. Sam,

    Ah, that (the difficulty you allude to about types of Russellianism) turns out to be my point. Well, sort of. I argue (both in my paper and also more briefly in my comments on Barbara Montero’s paper in this conference) that distinguishing between Russellian physicalism and Russellian non-physcialism is a tricky business, especially since the non-physicalists don’t want to say that neutrinos or strings or whatever turn out to be fundamental are conscious in the full fledged sense that, say, we humans are. So they fudge and say that the fundamental entities are conscious in an attenuated sense, or are ‘proto-conscious’ or whatever. But then one wonders how they differ from physicalists, especially since one finds self-styled physicalists using the ‘proto’ language as well. I do then go on and propose a way of distinguishing them, and according to that way you turn out to be a physicalist.

    Emmett

  36. Hi Guys,

    Just a quick note. I had a minor eye operation this morning. I was hoping that I’ll be OK by this evening but unfortunately it is very hard for me to read at the moment. I am sorry to be missing out on the fun. I hope I’ll be back to normal on Friday and can address all the interesting issues you guys have brought up.

    K

  37. Hi Istvan,

    I am trying to put various pieces of your responses together to see if I get it. Here are some things you’ve said:

    “There is no conceptualization C , having kind K, of the fact that some C* is a conceptualization, having kind K*, of some fact F, if K is not identical to K*”

    “Here is an example. Take the phenomenal concept of PAIN, and say that it has feature F (F is assumed as topic-neutral). Call the story that there is a phenomenal concept of PAIN with feature F, S.

    Physical conceptualization of S: “there is a neural property N, such that F(N).”

    Phenomenal conceptualization of S: “there is a concept whose canonical deployment involves a feeling of pain, such that F(N).”

    So is the fact you are claiming cannot be physically conceptualized the following?

    “there is a concept whose canonical deployment involves a feeling of pain, such that F(N)” is a phenomenal conceptualization of the fact that the concept PAIN has feature F

    Let’s name this fact ‘T’.

    Let ‘P’ be a name for the conceptualization “there is a concept whose canonical deployment involves a feeling of pain, such that F(N)”

    Let us assume, for the sake of argument, that physicalism is true. If it is, then the property of being a phenomenal conceptualization of the fact that the concept PAIN has feature F is some complex functional/physical property. As such, it has a physical conceptualization. Let’s pretend that the physical conceptualization of that property is ‘Z’.

    Here is the physical conceptualization of T:

    “P is Z”

    This conceptualization contains no phenomenal concepts. And it clearly a conceptualization of T.

    No good?

  38. C = the story of deploying the concept of PAIN according to PCS.

    CPhys = the physical way of putting that story = there is a brain state with such and such chemical or physical or neural properties.

    CPhen = the phenomenal way of putting that story = there is a state characterized by a repulsive, disturbing, indeed, painful feel.

    Is there a physical way of putting it that CPhen is a way to put C? Suppose there is. Will it use or mention the concepts contained in CPhen?

    If it will use them, then it will be a phenomenal way of putting CPhen, just like CPhen itself. If it is a phenomenal way of putting it, then it is not a physical way of putting it (compare: a evening-star-way of putting it that Venus is bright is never a morning-star-way of putting it that Venus is bright, though both ways are ways of putting it that Venus is bright).

    If it will mention them, then a conceived zombie could equally offer such a way of putting it that CPhen is a way to put C, unless mentioning the concepts used in CPhen is the same as using them. If a conceived zombie could equally put it, then the way of putting it that CPhen is a way of putting C has nothing to do with the phenomenality of CPhen’s way of putting C. If mentioning them is the same as using them, then zombies can’t offer it, but then it is, again, a phenomenal way, not a physical way.

    I really thought that my premise is obvious, because I don’t see any difference between what I say and saying that a round square is impossible. With this I put an end to further explanations.

    Kati has to either give up on the idea that there is a physical conceptualization of any physical fact, if physicalism is true (so there are physical facts that can’t be put in physical terms), or, as Richard Brown suggested earlier, to give up on any tight, constitution-like connection between concept and property when it comes to phenomenal ones. Both of them are bad, in my opinion, at least for the PCS. The first makes it doubtful that the resulting theory is physicalist, the second makes it doubtful that the resulting theory is about phenomenal concepts at all (given that these concepts are supposed to explain our epistemic situation, to use a phrase of Dave Chalmers’).

    My view, in general, is that PCS is incoherent. Type A materialism looks false if you think there are non-trivial facts about what is conceivable, but it is coherent at least. My real view is that one should be agnostic about all views, because there is no non-trivial fact about what is conceivable and whether conceivability entails possibility, and so any reasoned proof or disproof of any one of them is impossible, or meaningless if you want.

    So, I believe that there isn’t a real debate at all, because there can’t be: all there is is reducible to some people saying P and some people saying non-P. But this is the topic of another paper…

  39. We can give a physical conceptualization C of a phenomenal conceptualization C*. C* has the property of being a phenomenal conceptualization. We cannot give a physical conceptualization of the fact that C* has the property of being a phenomenal conceptualization, because such a conceptualization would not a priori entail “C* has the property of being a phenomenal conceptualization”. And if it does not a priori entail that, then the PCS fails (where we are not understanding phenomenal conceptualizations in topic neutral terms).

    I *think* I agree with this. I am just not sure whether Kati needs the a priori entailment. Trying to remember off the top of my head, she says that the anti-PCS argument is on par with the original argument that says that physicalism is false because P–>Q is not a priori. The physicalist (Type-B PCS-er) says that the reason P–>Q is not a priori has to do with the nature of phenomenal concepts. Kati rejects the topic neutral account of phenomenal concepts, and we worry whether P—>C is a priori. If it is, then it looks like P–>Q should be, too. If it isn’t, then the physicalist is in trouble. Kati says no, she isn’t. Her account of phenomenal concepts explains how it can be both possible that physicalism is true and that P—>C is not a priori. This requires that there be a physical conceptualization of the fact that C* is a phenomenal conceptualization. But the physical conceptualization need not a priori entail a phenomenal conceptualization–and of course it won’t if her account is right.

    Here is my quick take on what’s going on. Chalmers challenges the PCS: either P & ~C is conceivable, or it isn’t. If it is, the PCS strategy fails. If it isn’t, then since C explains our epistemic situation, it explains the zombies situation. But it doesn’t. The typical type-B materialist sees the failure of P–>Q being a priori as a challenge that needs answering. Chalmer’s argument is that she cannot answer it (the challenge simply reappears with respect to P—>C). Kati says that P & ~C is conceivable, but there is a perfectly respectable explanation available to the materialist for why. It is perfectly respectable because all that is required is that there be a physical conceptualization of the FACT that C* is a phenomenal conceptualization. And, of course, if physicalism is true, there will be such a conceptualization. In this context, it is not question begging to assume physicalism, since she is just trying to show that there is a coherent explanation of why P & ~C is conceivable that is compatible with materialism.

  40. Thanks a lot for your comments, Dave.
    Regarding the response to your objections to the PCS that I (briefly) suggest at the end of my commentary, I was not directly advocating a denial of the link between conceivability and explanation, but rather I was suggesting a new way of understanding the task of the PCS (although perhaps my line of response is committed to the denial of such a link after all). I was focusing on the a priori version of the conceivability argument that Kati discusses in her paper, and I noticed that the aim of the PCS is to refute the following premise:

    (2) It is a priori that if P&~Q is conceivable, it is metaphysically possible

    Put in these terms, the aim of the PCS is to show that we can conceive of a scenario where conceivability does not entail possibility, that is, an entirely physical scenario where individuals are conscious and still P&~Q is conceivable, due to the fact that their phenomenal concepts (physically realized) are not a priori connected to physical-concepts.
    Of course, the crucial question is what else is required in such a scenario, once we try to fill in the details to make it fully conceivable. I guess you would claim that since in that scenario phenomenal concepts are physically realized, there should be a physicalist explanation of them, and therefore P&~C cannot be conceivable. And at the same time, since the conceivability of P&~Q is to be explained in terms of the special features of phenomenal concepts, then C&~E cannot be conceivable either. But as you argue, these two conditions cannot be met at the same time.
    One response, as you say, would be to break the link between explanation and conceivability (option 4 in your paper), and say that C can both be physically explainable, and explain the epistemic gap.
    But I was trying to suggest a different response, one according to which explanation is not even required. My suggestion is that all we need to refute (2) above is to conceive of a scenario in which (i) conscious beings are entirely physical, and (ii) they possess phenomenal concepts that are not connected a priori to their physical concepts, so that P&~Q is conceivable. It is not clear to me why these two conditions couldn’t (conceivably) be satisfied at the same time, or why explanation needs to enter the picture at all.

  41. Hi Emmett,
    Very interesting comments. I would like to say a couple of things in response, regarding the PCS vis-a-vis the knowledge argument, and about the epistemic import of “presence”.
    Regarding the knowledge argument, I think the PCS has a more straightforward explanation of why Mary acquires new knowledge after being released. It is part of most accounts of phenomenal concepts that, for most phenomenal concepts, it is necessary to have had the relevant experience in order to possess the corresponding concept. If so, then it is clear why Mary acquires new knowledge: she acquires new concepts that she could not get before, so she can form new thoughts and therefore new knowledge (if we individuate knowledge in terms of thoughts, not in terms of facts). But the main point of the PCS is not only that possessing phenomenal concepts requires having had the relevant experiences, but also and more crucially, that phenomenal concepts are not connected a priori with physical concepts. This claim is needed in order to account for more sophisticated versions of the knowledge argument, such as the ‘Experienced Mary’ version discussed by Stoljar’s “Physicalism and Phenomenal concepts” (M&L, 2005) (in which Mary possesses phenomenal concepts but still is not able to infer Q a priori from P), and in order to explain the conceivability of zombies. (I elaborate on this in my paper “Defending the phenomenal concept strategy”, AJP 2008).
    -RE the “presence” relation: I don’t see why ‘presence’ has to be epistemically irrelevant. Let me elaborate a bit. As Chalmers has pointed out here and elsewhere, we could understand the constitutional account as either topic-neutral (and therefore physically explainable), or in phenomenal terms (that is, entailing phenomenal truths). If we take the latter option, I think the epistemic import is clear. If we take the former, things are a bit more complicated, but I think we can still argue phenomenal concepts are epistemically significant, along the lines of Carruthers and Veillet’s “The phenomenal concept strategy” (JCS, 2007). If we characterize phenomenal concepts in purely physical-functional terms, then zombies will also have phenomenal concepts. But zombies lack phenomenal properties, so does this mean that phenomenal concepts are epistemically insignificant? Carruthers and Veillet argue that, in fact, zombies are no different from us, concerning our epistemic situation: we are directly acquainted with phenomenal properties, thanks to phenomenal concepts, and zombies are directly acquainted with pseudo-phenomenal properties, thanks to their pseudo-phenomenal concepts. There is a difference concerning the referents of the concepts, but not the justification of phenomenal beliefs. They compare this to the case of Oscar and Twin-Oscar: they have analogous epistemic situations, but one is thinking about water and the other about twater.

  42. Hi Martin,
    Great comments. I am very sympathetic to your suggestion that what needs to be explained by the PCS is not our epistemic situation regarding consciousness as a whole (and certainly not our knowledge of consciousness), but rather something more specific, namely, the epistemic gap between physical and phenomenal truths, that is, that P does not entail Q a priori. If we take the second horn of Chalmers’ dilemma, according to which P&~C is not conceivable, then zombies will also instantiate C, and therefore they will not be able to infer Q from P a priori either. So C can explain the epistemic gap between P and Q, which is everything that needs to be explained!
    Chalmers discusses this type of response in his paper, under option 3 (accepting that C does not explain E). He argues that E (the epistemic situation) cannot be just the inferential disconnection between P and Q (which he agrees zombies also satisfy), because the epistemic gap that gives rise to the conceivability argument is richer: we are able to conceive physical duplicates of ourselves that lack the phenomenal states that we do possess.
    But this is tricky: Chalmers also claims that E has to be characterized in topic-neutral terms, so it cannot be a priori that E entails having phenomenal states. However, Chalmers’ argument requires that E is satisfied only by conscious beings, not by zombies. How is this possible, if E is topic-neutral? Can Chalmers have his cake and eat it too? It’s not clear to me why a topic-neutral characterization of E would have the consequence that zombies don’t instantiate E.

  43. I don’t see a problem giving a topic-neutral characterization of E here. It will require that zombies have a concept Q* (a quasi-phenomenal concept, as characterized in the paper) such that “I have Q*” is true for them (this part captures the “we do possess” requiremnent), P&~”I have Q*” is conceivable for them (this part captures the inferential disconnection requirement), and so on (including a clause about the cognitive significance of their knowledge of “I have Q*” if necessary). This E is topic-neutral, and I argue in the paper that zombies don’t satisfy E (and more importantly, that it’s conceivable that zombies don’t satisfy E).

  44. Kati & Esa,

    Kati: I’m sorry to hear that you’re temporarily out of commission. I hope you’ll be better soon. In fact, I was going to hold off a bit with a followup to my most recent (Feb. 25) missive to you and Martin to see if you had a response. But since you may not be able to respond, I’ll do the followup now before this conference shuts down.

    So, the issue, recall, is how, given your constitututive account of phenomenal concepts (and David Papineau’s similar quotational account of phenomenal concepts) you (or he) can withstand H & T’s argument for non-physicalism. And in my most recent missive to you, I expressed some doubt that your distinction between a ‘direct grasp of essence’ and ‘presence’ can do the job. Now Esa has some interesting responses to this that I might respond to when I process them and if I get time. But in the meantime, let me offer some ideas that have been going thru my head.

    First, go Russellian. This is an idea I raised earlier. On a Russellian view, physical theory just tells us about the causal, functional and other structural relations in which the fundamental properties of nature stand to each other; but it doesn’t tell us about their intrinsic natures.So when we are conceving of P & not-Q our conception of P just covers those causal, sturctural and functional relations. What experience does is acquaint us with the intrinsic properties and that is what constitutes the content of phenomenal concepts. But one could take the position that the intrinsic properties with which experience confronts us have an underlying essence that constitutes a supervenience base for them, but with which experience cannot acquaint us. So this essential nature cannot be incorporated into the contents of phenomenal concepts. Furthermore, one could take the position that the essential supervenience base is physical. So the intrinsic phenomenal properties with which one is acquainted have a nonphenomenal base with which one is not acquainted. This is more or less a position that Stoljar has proposed.

    But there is another idea that occurred to me. Might it be possible to directly grasp the essence of experience but not, thereby, know whether it is mental or physical? This is an example of a more general question: Might it be possible to directly grasp the essnece of a type but not, thereby, be able to know that it fits under a higher order type. I decided to try this with water. Granted the analogy may not be quite right here, since we don’t directly grasp the essence of water in the way we do (or maybe do) the essence of experience. But let’s go with it anyway. Let’s say the essence of water is H2O. So to know the essence of water we just need to know that it is composed of H2O molecules and what this incolves in terms of the exchange of valence electrons, the binding strength among H2O molecules under various conditions and so forth–but we need not go to the level of quarks, strings, or whatever. (Tho actually, I don’t think it would hurt my argument if we did) Now on the basis of this, can we know that water is liquid at room temperature, turns solid at 0 degrees C, etc.? Well, it may seem obvious that we can. And I agree–EXCEPT that arriving at these conclusions requires REASONING. So certainly it seems that there could be some cases where we know, even directly grasp, the essential nature of a type but the reasoning involved in deciding if that type is of a higher order type is beyond us. And maybe that is how things stand with conscious experience and its physical or non-physical nature. And so maybe, in turn, our phenomenal concepts have their contents thus limited. In fact, if the reasoning in question is beyond us for reasons of genetic hardwiring, then we have a form of mysterianism–mysterianism with respect to physical or non-physical nature, but not with respect to essential nature. Derek Bell has an interesting paper in this conference on a form of mysterianism that trades on genetically hard wired limitations on reasoning power, and maybe some of his ideas could be enlisted to flesh this out.

    The foregoing is, of course highly speculative, and I’m not sure if it can be made to fly. But I offer it for your (and others’, of course) consideration.

    So, that’s about it. Thanks for your input Esa, and I’ll mull over your remarks. And get well soon Kati.

  45. Hi Esa, Martin, David,

    I’m also a bit puzzled by E, in the following way: It seems that we can conceive of the zombies lacking E precisely because we conceive of them lacking consciousness. That’s what distinguishes our epistemic position from theirs, because it’s the factor that makes them wrong about some beliefs, unjustified about others etc.
    But then mightn’t the PC strategist say that it is OUR having C – the ground of our ability to conceive of consciousness-less physical duplicates – that is also the ground of our ability to imagine these zombies lacking E when they have C. For it is C, in us, that (to repeat) allows us to refrain from imagining the zombies as conscious, and on the strength of that allows us to imagine that they differ regarding E.
    So, such a PC strategist might claim, C – whatever it is – DOES in fact explain E, as in necessitates E in a perspicuous a priori way. But we simply can’t see that, for the same reason that we can’t see why the physical necessitates the conscious in an equally bona fide way: it’s C – and so our phenomenal concepts – messing the whole thing up for us cognitively. C is the source of the conceivability of zombies, and of the conceivability C without E in the zombies. It underlies both gaps.
    The PC strategist ought to hold, therefore, that no explanation (for us) of E in terms of C could be hoped for, precisely because our having C precludes this. But that doesn’t count against the possibility that some factor C exists which plays these roles.
    What about that for a story? I don’t like it, but is it inconsistent?

  46. I think it’s inconistent to say that C necessitates E in a perpicuous a priori way and that C without E is conceivable, as it’s a constraint on the relevant notion of conceivability if S is conceivable, S cannot be ruled out a priori. I suppose the suggestion might be rephrased as the suggestion that C without E is prima facie but not ideally conceivable (but then this would be very different from the way the standard PCS works), or else that C without E is conceivable but impossible (but then we rub up against issues concerning the link between conceivability and explanation).

  47. Yes I think in your terms it’s prima facie conceivable but not ideally; i.e., it’s inconceivable for beings whose imaginings aren’t messed up by their having C, as ours are.
    As I say, I don’t warm to this story. But it does seem to be a remaining hiding place in logical space for the PC strategist.

  48. Esa,

    Greetings. Thanks for your comments and sorry for dragging my feet in responding.

    Granted everything you say about the Knowledge Argument and the role phenomenal concepts play in debating it, but I don’t think it answers the question. In particular, I don’t think it answers H & T’s argument against physicalism. It still all hinges on the content of the phenomenal concepts post-release Mary acquires, and hence on what exactly it is she comes to know.

    Of course, H & T’s argument highlights a difficulty only if we don’t take the view that phenomenal concepts are phenomenologically vacuous, as I think Loar, on the best reading of him, takes them to be. (You’ve mentioned that you like Loar. I’m not sure if you read him that way) Kati (how’s the old eye doin’, by the way?) doesn’t take that view. For her, phenomenal concepts involve a mode of presentation with substantial content (as she puts it), and given the dialectic of the debate, it looks like that content has to reveal the essential nature of color (or whatever) experience. Since it doesn’t reveal it as physical of functional, H & T’s argument goes thru. Or at least that is how I argued earlier.

    As to the dichotomy you cite: Either topic neutral or in phenomenal terms. Keeping in mind that the issue here is concerning the mode of presentation of phenomenal concepts, the former seems to me to be tantamount to phenomenological vacuousness. The problem with this isn’t that it doesn’t allow for epistemic import. It does. The problem is that it is phenomenologically implausible. That’s not an argument, of course, it’s a confession of faith. …Well, of intuition. (Someone like Dennett wouldn’t see the difference, of course) But I take it to be Kati’s point of departure, and it’s from that that I developed my argument. So we are left with understanding the constitutional account in phenomenal terms. But that opens it up to H & T’s attack.

    Enter Kati’s suggestion of a ‘presence’ that does not involve a ‘direct grasp of essence’, as a third alternative. That brings us back to my question in an earlier entry. What does this presence do? It looks like it is epistemically irrelevant.

    Now my most recent entry allows that I may have set up a false dichotomy here (or trichotomy if we count the epistemically irrelevant presnece) since there are a couple of other possibilities. So I’ll leave those for folks (myself included) to ponder and bid adieu, except to say thanks for alerting me to two articles I have not yet read: yours and Stoljar, 2005.

  49. Hi Emmett,
    Thanks for your interesting response. I am not sure about the phenomenological point you mention, but I thought of another way of responding to Horgan & Tienson’s argument which does not rely on the rejection of your intuitions concerning the phenomenology of phenomenal judgements.
    This new response basically claims that there is an equivocation over the expression “physical”:

    1. When a phenomenal property is conceived under a phenomenal concept, this property is conceived otherwise than as a physical-functional property.

    2. When a phenomenal property is conceived under a phenomenal concept, this property is conceived directly, as it is in itself.

    3. If (i) a property P is conceived under a concept C, otherwise than as a physical-functional property, and (ii) P is conceived under C, as it is in itself, then P is not a physical-functional property.

    4. Therefore, phenomenal properties are not physical-functional properties.

    In the debates concerning the characterization of physicalism, it is common to distinguish between a ‘narrow’ notion of physical (that is, entities posited by physics), and a broad notion of physical (that is, entities that depend on or are determined by the former). Therefore, physicalism would claim that all entities in the actual world are physical-broad, that is, determined by the physical-narrow. (The proper characterizations of ‘determination’ and ‘physical-narrow’ are controversial matters, of course, but we don’t have to go into that now.)
    If we follow this distinction, it can be argued that the notion of physical in premise 1. should be physical-narrow, because otherwise 1 would beg the question: what we are justified in asserting is that our phenomenal properties are conceived as other than physical-narrow. It would be question-begging, at this stage, to assume that phenomenal properties are conceived as other than physical-broad, because if we grant that the mode of presentation of phenomenal concepts are the very same phenomenal properties, then to assume that would amount to saying that phenomenal properties are not physical-broad, which clearly begs the question at issue here.
    So premise 1 should go as follows:
    1*. When a phenomenal property is conceived under a phenomenal concept, this property is conceived otherwise than as a physical-narrow property.

    But then, if we want the conclusion to follow, it should invoke the same notion of physical:

    4*. Therefore, phenomenal properties are not physical-narrow properties.

    But it is clear that 4* is not damaging to physicalism. What is damaging to physicalism is the following:

    4**. Therefore, phenomenal properties are not physical-broad properties.

    But 4** does not follow from 1*. So H&T argument does not work.

  50. Esa,

    Well, the beat goes on. I appreciate your attempt to answer H & T’s argument, but I don’t think it works. Or at least it doesn’t work without abandoning PCS.

    One key point to keep im mind is that premise 1 is not about what phenomenal properties ARE (physical-broad, or whatever) but how they are CONCEIVED under phenomenal concepts. Another key point revolves around something you allude to: The sense in which the physical-narrow determines the physical-broad. Granted there is controversey about this, but one assumes that it is such that there is no metaphysical gap between the one and the other (so something like metaphysical supervenience would do the trick). And this in turn means that when the phys-narrow is conceived under physical-narrow concepts, and similarly for the phys-broad, there is no epistemic or conceptual or explanatory gap between them either. So from all the phys-narrow facts conceived as such, the phys-broad facts conceived as such follow a priori.

    But if all that is so, then premise 1 of H & T’s argument has to be true even when ‘physical’ is construed as ‘phys-broad’. Indeed, to deny this would be to abandon PCS (and Type B physicalism in general, I think). After all, the whole point of PCS is that there IS a conceptual (or epistemic or…) gap between phenomenal properties conceived under physical (including phys-narrow) concepts and phenomenal properties conceived under phenomenal concepts. So, so far from begging the question, maintaining premise 1, no matter how one reads ‘physical’, is just being true to PCS. If there is any question begging going on here, it would most likely be with premise 2. But I have already argued that abandoning that has problems of its own.

    There is one answer to this that it has occurred to me someone might give–namely, that we are in no position to know that premise 1 (with the ‘phys-broad’ reading) is true, because we don’t yet know all the relevant phys-narrow facts. Indeed, we are not even close. It just might be that, once we did, we could infer a priori all the facts we know about phenomenal properties under phenomenal concepts. So retrospectively we would see that premise 1 is wrong (for the ‘phys-broad reading) and that upon conceiving phenomenal properties under phenomenal concepts we were all along conceving them under phys-broad concepts.

    This is a possible position to take, but it too amounts to abandoning PCS, since conceptual/a priori connections are possible between nar-phys concepts/facts and phenomenal concepts/facts. The latter thus lose the special status that PCS gives them. Indeed, you might recognize this as Daniel Dennett’s Type A physicalism. Addressing the Knowledge Argument, we are in no position, he holds, to say that pre-release Mary cannot know what it’s like to have color experiences, because we have no way of conceiving the vast amount of information she is supposed to have mastered. Still, assuming, per impossible, that she could do this, his money is on her knowing pre-release what it’s like to experience color. So special phenomenal concepts need not be brought into the picture to save the day for physicalism.

    Actually, this view seems to kind of phase into a kind of mysterianism. If our inability to master all the relevant phys-narrow facts is due to genetically hard wired cognitive limitations, as Dennett seems to think, then it turns out that phenomenal properties are phys-broad properties, and phenomenal concepts are really phys-broad concepts, but we will never quite see how.

    So, yes, one can reject premise 1 of H & T’s argument given a ‘phys-broad’ reading of ‘physical’. But doing so amounts to abandoning PCS and rendering phenomenal concepts otiose.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s