Why Consciousness Can’t Just be in the Head: A New Argument against Biological Theories

Presenter: Adam Pautz, University of Texas –Austin

Commentator 1: Sydney Shoemaker, Cornell University

Commentator 2: Thomas Polger & Doug Keaton, University of Cincinnati

Advertisements

35 Comments

  1. I am very grateful to Sydney Shoemaker for taking the time to write his comments, which will be invaluable to me in revising the paper. Shoemaker helpfully suggests that my “externally-directed” argument against biological theories be slightly modified in the light of brains in vats (BIVs). I was pleased to find that he “leans towards” accepting the modified version. (Although, given his intentionalism, I would have thought Shoemaker is outright committed to something in the vicinity of my “externally-directed intuitions” – I say more on this below.) My version relies on a number of intuitions, but here I will simplify, and say it rests on the *roundness claim*: R necessarily represents roundness at p. (Contrary to what Shoemaker implies, I do not say R necessarily represents redness.) Shoemaker replaces this with the *determinable claim* (1 below) and the rest of the argument proceeds as before:

    1 Experience property R necessarily represents d-roundness at p, where d-roundness is a determinable with ordinary roundness and roundness* (the relevant computer state in the BIV scenario) as determinates
    2 Biological property N doesn’t necessarily represent d-roundness at p
    C So R is not N

    Shoemaker accepts 2 on the basis of the same sort of considerations I develop in the paper and says he “leans towards” 1. Now if either my original version or Shoemaker’s helpful alternative version is correct, then my general strategy for refuting biological theories (which is distinct from the usual multiple realization argument) is vindicated. But I think the issue of which is to be preferred raises interesting questions about shape representation (about which Shoemaker has not said much in his work, having focused mainly on color representation) and the physical basis of consciousness. For this reason, in a future version of the paper, I would like to provide some reasons to think the roundness claim is to preferred to the determinable claim. I will try them out here. I would be grateful for any feedback.

    First, a clarification. Contrary to Shoemaker, my main response to the BIV argument against the roundness claim was not that in every possible BIV scenario R’s representing roundness is constituted by the truth of a counterfactual along the lines of: if the BIV were “appropriately” connected to its environment, then the relevant state would be caused by round objects and cause roundness-appropriate behavior. Although I should have been more clear, I only meant that this is *one* possible response to *some* elaborations of the BIV. In fact, in the paper, I express skepticism about all such reductive theories of sensory intentionality. And I agree with Shoemaker that in any case this response may not apply to some particularly bad-off BIVs (e. g. ones with no evolutionary history, no human input-output systems, etc.). In what follows, for the sake of argument, I will assume that we are dealing with such a bad off BIV (= a brain in a vat to which the counterfactual strategy does not apply).

    As I said in the paper, my response to the BIV argument against the roundness claim is that rejecting one of its premises is more reasonable than rejecting the roundness claim. To see this, let me first reiterate how plausible the shape claim is. Then I will turn to the premises of the BIV argument.

    Imagine looking at a tomato. On Shoemaker’s suggestion, if a BIV has that very experience-type, R, then its experience might be wholly veridical, even though no round, tomato-like thing is before the BIV. On his suggestion, in the BIV, R merely accurately represents the computer state, roundness*. (Or perhaps the suggestion is that in the first instance R represents d-roundness, which in this case is realized by roundness*.) Against this, it is intuitively necessary that, if an individual (even a BIV) has R, then that individual is in a state that matches the world only if a round (not merely d-round, not merely round*) object is before one: an object whose edges are in the ordinary sense roughly equidistant from a common point. (Such an individual could learn Euclidean Geometry, not Euclidean Geometry*.) In accordance with this, most everyone would agree that, *if* a BIV might have R, then its experience is at some level non-veridical, since no such object is before him.

    Now, Shoemaker grants the pull of the intuition but objects that it makes a questionable use of the first person perspective, and that it at best shows that the roundness claim applies to human beings. I have two replies. (i) The intuition is that any individual x whatever (whether human or no) that has R represents roundness at p. (ii) Everyone must grant that the first person perspective can be used to justify *some* unrestricted modal claims about experience properties: for instance, necessarily, if x has a bluish experience, a purplish one, and a greenish one, then x’s first experience resembles his second more than his third. Indeed, Shoemaker himself seems to grant this, since, on the basis of the transparency observation, he has long held that (as he puts it in his recent ‘A Case for Qualia, at p. 325) “each quale has, constitutively and independently of context, a representational content of its own” (and presumably not just in humans but in all possible individuals). He also “leans towards” the determinable claim (understood as a thesis applying to all individuals, not just humans), so he must think it enjoys some justification. Presumably, the first person perspective has something to do with that justification. Why then cannot the first person perspective be used to provide an initial, defeasible justification for my shape claim?

    It seems to me that, by contrast to the shape claim, the premises of Shoemaker’s BIV argument against it do not enjoy very strong support. The first (implicit) premise is that a bad off BIV might have R. The second is that in the BIV R represents the computer state roundness*, not roundness. Accordingly, a BIV who has R might enjoy a perfectly accurate experience, even if no round thing is before him.

    An aside: Shoemaker’s main point is that the BIV’s *beliefs* represent roundness*, not roundness. Even if this is right, it does not without some argument undermine my shape claim that in the BIV the *experience property* R represents roundness, so that it inaccurate at some level. Likewise, it does not undermine my externality and matching intuitions, since these only concern the representational properties of the experience property R. In fact, in his ‘Matrix’ paper and elsewhere, David Chalmers combines something like Shoemaker’s claim about belief with my roundness claim about (the Edenic content of) experience. For reasons I would be interested to hear, Shoemaker himself says he is “not sure whether [R in the BIV] would count as having the matching property [and hence count as inaccurately representing roundness]”. So it looks like he is actually *open to* something in the vicinity of my roundness claim. However, since at the start of his comments Shoemaker says BIVs undermine all my claims, I will provisionally assume that Shoemaker’s suggestion is that in the BIV R (not just the BIV’s beliefs) accurately represents the computer state roundness*, rather than inaccurately representing roundness.

    What might be the argument for the first premise, that a bad off BIV might have R? After all, many externalists-functionalists would deny that this is metaphysically possible. Shoemaker does not provide an argument in his comments. One argument is that it is conceivable, and therefore possible. But no *physicalist* could comfortably put forward this argument, since Zombies, etc. are equally conceivable. A second argument is based on empirical investigation into the neural basis of experience. But empirical investigation at most shows that the neural plays a role in determining the phenomenal; this falls short of showing that a bad off BIV might have R (for instance, maybe there is a general externalist condition on experience that is not met by such a BIV). A third argument is this follows from the right theory of R – e. g. the biological theory, or some non-standard ‘narrow’ functionalist theory. But of course this is hard to establish. Indeed, I myself am confused about how it is consistent with Shoemaker’s own (partially) functionalist theory. On the only plausible way of elaborating that theory, phenomenal similarity among shape experiences is explained in terms of their being apt to cause beliefs about the objective similarities of things in respect of shape, and their being apt to cause sorting and other behavior. But, by Shoemaker’s own lights, the bad off BIV does not have shape beliefs; it also does not (even potentially) engage in sorting and other behavior. (For more here, see below.)

    Granting the first premise for the sake of argument, what might be the argument for second premise? The second premise was that in the BIV R represents the computer state roundness*, not roundness, and accordingly that a BIV who has R might enjoy a perfectly accurate experience, this despite the fact that no round thing is before him. Again, Shoemaker does not give an argument. One argument might rest on the assumption that some kind of causal constraint applies to all representation, including the sensory representation of roundness. But (as David Lewis pointed out in his response to Putnam), even if a causal constraint is right for names and natural kind terms, it is very implausible that it is right for all representation. Indeed, against the second premise, Horgan, Tienson, Chalmers and others would say that even in a BIV R represents roundness; they therefore deny that a causal constraint applies here. (Horgan-Tienson would also say the BIV’s *beliefs* represent roundness, in accordance with my ‘grounding intuition’, despite the absence of a causal connection with round objects.) And Colin McGinn’s much less extreme Percy-Twin Percy case (from his *Mental Content*) strongly suggests that a causal constraint does not apply to the representation of shape. I am not here defending a kind of radical internalism about sensory representation on which the BIV might sensorily represent roundness. I am only saying that we have not been presented with any strong reason why the BIV could not sensorily represent roundness. Indeed, it seems to me equally hard to explain how a BIV might represent d-roundness, as on the determinable claim suggested by Shoemaker (the last section of my paper is relevant here).

    So, the shape claim is enjoys stronger support than either one of the premises of Shoemaker’s BIV argument against it. I conclude that rejecting one or another of these premises is more plausible than rejecting the shape claim.

    Let us now look at Shoemaker’s determinable claim, which he says he ‘leans towards’, and so thinks enjoys some justification. Presumably, that justification is that from the first person perspective R intuitively represents some shape-like property, and the BIV argument means that it cannot be roundness and suggests that it might be something like d-roundness. (Further, at the end of his comments, Shoemaker suggests that the determinable claim is compatible with something very close to the internalism he favors.) Since I do not myself find the BIV argument very persuasive, as I have just explained, I do not find this justification for the determinable claim very persuasive. Further, I have questions about the determinable claim. Is it possible to describe in English the relevant property, d-roundness, of which the radically different properties roundness and roundness* are determinates? Is it something like a Shoemakerian appearance property or qualitative character, a disposition to produce certain responses in individuals? What psychosemantic theory explains how it is represented (an issue briefly addressed in the final section of my paper)? I am very curious about how the determinable claim might be developed, since Shoemaker has not said much about the content of spatial experience, having focused mostly on the content of color experience. I worry that perhaps it cannot be adequately developed (for reasons related to what I say in my paper’s final section). So, not only is there no strong argument for the determinable claim; there may also be an argument against it. By contrast, as we saw above, the overall case for the roundness claim is very strong, notwithstanding BIVs. This concludes my discussion of my reasons for preferring my original roundness claim to the determinable claim that Shoemaker offered in its place.

    A final point. The roundness claim (and likewise all my externally-directed intuitions) is consistent with a variety of different views about the metaphysics of the shape property that R necessarily represents: it might necessarily represent round-from-here (Tye), a Thouless property (Hill-Bennett), ‘perfect roundness’ (Chalmers). There are also difficult and fascinating questions arising out of Brad Thompson’s ‘double earth cases’ and his arguments based on them. I do not deny this. But, whatever the correct view, I claim that there will be some sense in which R necessarily represents roundness at place p. (There are purely ‘Fregean’ views which reject this roundness claim; but I think they should be rejected, not only because they violate the intuitive shape claim, but for other reasons that Chalmers gives in his ‘Eden’ paper – for instance, they do not accommodate certain ‘internal connections’ among experiences in different modalities.) The roundness claim, however it is to be explained, is enough for my argument against biological theories. Indeed, as Shoemaker explains, even his determinable claim would be enough. Even if one thinks (as Shoemaker does but many do not) that the multiple realizability argument provides a some reason to reject the biological theory, I hope that it is worthwhile to point out that it is also subject to this quite different style of argument.

    Shoemaker’s reply is extensive and contains much else of interest, which I discuss in a more complete version of my reply here:

    https://webspace.utexas.edu/arp424/www/shoemaker.pdf

    In this more complete version, I raise some questions which I was hoping Shoemaker might address about why (given his intentionalism) he merely leans towards something in the vicinity of my externally-directed intuitions, and about how his functionalism is consistent with what he says about BIVs. I also answer a question Shoemaker asked: why do I think that the multiple realization argument is not enough to refute the biological theory? And I offer some related remarks about the possibility of indeterminacy in our concepts of experience and about the ‘primitivist’ view I favor. I would be grateful for any feedback on any of these matters, and I again thank Shoemaker for this extremely helpful comments.

  2. I am indebted to Douglas Keaton and Thomas Polger (K&P) for their extensive comments, which will help me to improve the paper. I will address the main issue. My argument has two premises. First, R has certain externally-directed properties necessarily. Second, N doesn’t. So R isn’t N. K&P’s response is to accept the second premise (indeed they say *nothing* has the externally-directed properties essentially – in (iii) below I argue that this goes too far), but they reject the first premise across the board. Interestingly, other biological theorists I’ve heard from favor the very opposite response! They accept some instantiations of the first premise but reject the second premise. As I discuss in the paper, Ned Block is an example – although it seems to me that the theory he ends up with is not a biological theory. In his “Type Materialism” paper at p. 442, Brian McLaughlin at least says that this sort of view is a possibility.

    I have four points in response to K&P. I will focus on the externality intuition and the matching intuition, since even K&P find them “plausible”. (By the way, by an “intuition”, I just mean something we have some pretheoretical justification for accepting – I should say this in the paper. This does not commit me to a faculty of intuition, nor to special intuition episodes. Nor does this commit to the view that the source of the justification is introspective.)

    (i) Most philosophers working on perception accept the externality and matching intuitions, and regard them as having pretheoretical support. I have in mind all intentionalists (even weak intentionalists), all disjunctivists (some may object to ‘matching’ but I think they would misunderstand), inseparatists like Horgan-Tienson, Chalmers, Siewert, to name just a few. In fact, as just noted, even biological theorists like Ned Block accept the externality and matching intuitions and regard them as having strong pretheoretical support, in spite of K&P’s claim that no biological theorist will accept the intuitions. And, in his comments, Shoemaker says he leans towards something in the vicinity (and in fact his intentionalism seems to so commit him), and he says it rules out biological theories. Many also likewise accept the grounding intuition (e. g. those who think phenomenality is tied to accessibility) and the justification intuition (all ‘internalists’ and even many ‘externalists’). This is relevant because it strongly suggests what I say in the paper, namely that the intuitions (especially externality and matching) do enjoy extremely strong (not, as K&P say, ‘superficial’) pretheoretical support.

    (ii) K&P at one point say that they will provide a “theory-based deflationary explanation of each intuition”. This suggested to me (maybe I was wrong) that they were going to try to “explain away” the seeming plausibility of the externality and matching intuitions – in which case it would not be a serious cost of the biological theory if it forsakes them. In my opinion, K&P have not done that – at least I do not see how they have. They imply that the sentence ‘necessarily, if one has R, then one has an experience as of a round thing at a certain place’ is ambiguous. (In a longer version of this response linked to below, I express doubt that it is ambiguous in the way they suggest, but let me just grant their claim now for the sake of argument.) They say it is false on both readings. This does not explain away the fact many people find it pretheoretically very plausible that it has a true reading.

    (iii) If there were an argument against the externality and matching intuitions independent of the biological theory, then the fact (as I take it to be – and K&P agree) that the biological theory must forsake them would not count so strongly against it. But there are no strong arguments against the externality and matching intuitions. One argument (not one that K&P make) is based on ‘strange scenarios’, such as the brain in the vat. This is the fourth objection in my paper. I argue that it is unsuccessful in the paper, and in my response to Shoemaker.

    (iv) At one point, K&P offer a different argument. They reject the externality and matching intuitions on the basis of the very strong claim that “nothing at all essentially has the kinds of externally-directed properties that concern Pautz”. Would they let me know their argument for this strong claim? Obviously, it is an argument I need to address! In any case, I think there are strong reasons to doubt the claim. (i) As I say in the paper, the property *believing that a round thing is present* has such properties necessarily, contradicting K&P’s sweeping claim. For instance, this property is such that, necessarily, if an individual has it, then he is in a state that matches the world only if a round object is present. (ii) In section 6, I argue at length that the kind of ‘world-involving’ physical properties which with Tye and Dretske would identify R might have all four externally-directed properties necessarily. (So, contrary to K&P’s footnote 2, I explicitly deny that physical properties cannot have these properties necessarily. Therefore my argument differs from the one they attribute to Immanuel Kant and John McDowell – it is not that heavy! Thanks to K&P’s comments, I will in future versions make this clear at the very start of the paper. I hope it is also clear that my argument is nothing like Max Black’s.) Also, as Shoemaker explains in his comments, certain other physical (or physically realized) properties might have such properties necessarily. They are just not biological properties: they involve the environment in certain ways. This is why Shoemaker agrees that the externally-directed intuitions rule out the biological theory (although he thinks multiple realizability is enough to do that).

    (v) Another argument against the externally-directed intuitions – one that K&P seem to be making in some places – is based on the biological theory itself. If R is N, and if (in accordance with my premise 2) N does not have the externally-directed properties necessarily, then, of course, neither does R. But, as I just said, there are other views (defended by Tye, Dretske, Shoemaker, Chalmers, etc.) on which R is identical with a physical (or maybe non-physical) property that does have the externally-directed properties necessarily. Such views accommodate the externally-directed intuitions. I myself find the externally-directed intuitions ‘sacrosanct’. I cannot deny them. So I cannot accept the biological theory and have to go with an alternative. To others, perhaps, the intuitions are not ‘sacrosanct’, as I fully acknowledge in the paper. But, in any case, in view of my point (i), I think it must at least be admitted that the externally-directed intuitions are pretheoretically extremely plausible. And K&P themselves admit that the externality and matching intuitions, at least, are “plausible”. So, I hope that at least some folks will agree that the paper does something new and worthy of consideration: it demonstrates a hitherto unidentified cost of the biological theory (beyond the alleged cost of denying multiple realizability, etc.), namely that it must forsake all such intuitions. This provides a new reason for rejecting the a biological theory and accepting one of the aforementioned alternative theories. Many would say a strong reason for rejecting the biological theory, for many find the intuitions extremely compelling. Since decisive refutations are not to be found in philosophy, that’s about the best one can do!

    I have addressed what I think is the main issue. K&P also raise issues about total realizers, whether anyone holds my target, and ambiguity. It may be worth mentioning that Block, McLaughlin, Hill (who was once a biological theorist but who has recently rejected the theory) and others have looked at the argument and discussed it with me. None accused me of mischaracterizing the view. And I did not follow K&P’s criticisms about total realizers, because I never use ‘total realizers’ in the section they criticize (maybe I’m missing something). I address these issues in more detail in a longer version of my comments here:

    https://webspace.utexas.edu/arp424/www/kp.pdf

  3. Thanks to Adam for his detailed response to our commentary. My impression is that we have a bit of cross-talk going on, which I’m sure is my fault more than Adam’s or Tom’s. Perhaps all I am after with this comment is a bit of clarification.

    There’s a joke that says that any one time there are exactly five type-identity theorists in the world. Whenever someone adopts the identity theory, one of its former adherents decides to drop it. Point being: type-identity theorists are rather rare. But I wonder if Adam’s target is actually type-identity theorists. Adam says his target is theories that claim that experience types are identical to neural types. But, this rather underspecifies things. Block, in his early paper “What is Functionalism?” points out that even reductive functionalists who identify mental states with realizers, are not really identity theorists, since the allegedly purely neural states that they favor are infected with an ineliminable functional characterizations. Type-identity theorists need not be reductive functionalists — and part of my concern is that Adam is assuming that they are.

    Now, Adam says that this is not so, and as evidence he points out that he doesn’t use the term “realizer” very often in his paper. But that is not really what matters. What matters is that he uses “functional role” a lot in his paper. By definition, a property that can be separated from a functional role is a functionally characterized property — in other words, a core or total realizer.

    Thus, here is his description of the central argument:

    “The separation argument is meant to provide a general argument, independent of consideration of any particular theory, for the conclusion that N, by contrast to R, has its externally-directed properties only contingently, in accordance with premise 2. It begins by pointing out that there are possible separation cases in which N is stripped of its relevant (wide and narrow) functional role, and more generally is stripped of all relevant relations [to] roundness. Then it says that in these cases N is also stripped of the four externally-directed properties, thus establishing premise 2. This follows immediately from the standard functionalist account, but (to repeat) is plausible on any reasonable theory of the externally-directed properties.”

    The idea of wide and narrow functional roles is a functionalist idea. They occupants are realizers, by definition. So, whether or not Adam uses the word “realizer,” he is appealing to realizers. He writes that his argument is independent of consideration of any particular theory, but I cannot see how he can say that and also use talk of functional roles.

    Now, the quoted passage says “N is stripped of its relevant (wide and narrow) functional role, and more generally is stripped of all relevant relations to roundness.” Maybe Adam meant, “N is stripped of its relevant (wide and narrow) functional role, OR AT ANY RATE is stripped of all relevant relations to roundness.”

    But an identity theorist need not agree that the neural property that is identical to R can be stripped of all relevant relations to roundness. There are two different kinds of relation to roundness, one of which innocent, one of which is semantically loaded. Quoting from our commentary:

    “(a) being a property such that, if anyone has that property R, then he has an experience property like that which is had by me when I see a round thing at a certain viewer-relative distance d and position p.

    “(b) being a property such that, if anyone has that property R, then he has an experience that presents the world as though there were a round thing at a certain viewer-relative distance d and position p.”

    How could a property that has (a)-relations to roundness be stripped of its relevant (a)-relations to roundness? A property N that has the (a) property need not be a property that is functionally characterized in any sense — it need not “play a functional role” in the strong sense that is needed to make sense of the separation cases that drive Adam’s paper.

    If Tom and I say that experience-types are identical to non-role-playing properties we are adopting a view of experience types that is (I suppose) at odds with 99% of the philosophical community, which may explain why Adam has not heard our stripe of objection before. But we took it that his paper was directed at the minority view for which we are here attempting a defense.

  4. Hi Doug. Thanks very much for the follow-up. I am not sure, though, how the points you bring up bear on the evaluation of the argument (maybe I’m missing something). In your main comments, you accept premise 2 (indeed say *nothing* could have the externally-directed properties necessarily) but reject premise 1. You did not explicitly provide any arguments for rejecting premise 1 or for the sweeping claim (which seems open to clear counterexamples, as I explained in my comments). But I do not think they could have to do with the points you raise here, about the details of formulating the biological theory and about how my use of ‘functional role’ is somehow problematic. But let me address these points anyhow.

    I am aware that “experience types are identical to neural types” under-specifies things because it is compatible with Lewis-style realizer functionalism, and indeed already address the point in the paper. That is why there is always a ‘necessarily’ in my formulation (often in italics) – which you seem to have missed here. That rules out Lewis-style realizer functionalism. For more here, see footnote 5 of my paper. It seems, though, that I need to highlight this even more in future versions – so thanks for bringing it up.

    Doug also objects that “[Adam] uses “functional role” a lot in his paper”. I did not see how this is a problem – again, maybe I’m missing something. It just means that biological properties are apt to be caused by certain things and to cause certain things. In that sense, they have functional roles. Even biological theorists like Ned Block, Brian McLaughlin, etc. say so; they just do not think functional role constitutively enters into the account of experience. Doug says “Tom and I say that experience-types are identical to non-role-playing properties”. Do Doug and Tom deny biological properties are apt to be caused by certain things and to cause certain things?

  5. Hi everyone, I have been enjoying the discussion here so far.

    It seems to me that there is no problem for what we might call the ‘post-Lewisian’ biological theorist. To put this in a 2-D framework we can say that, roughly speaking, the Lewisian-style functional story gives us (part of) the primary intensions for R. Neuroscientific theories give us the secondary intension. A standard assumption in this debate is that the primary and secondary intensions of R are necessarily identical (i.e. that the PI of R picks out the same SI in all possible worlds) but the biological theorist can deny this (I in fact do).

    If this is the case then the biological theorist can agree with you that anything that is an experience will have the properties you are worried about but they will disagree that the neural states in question must have those properties and this is because the the primary intension may pick out different properties in different possible worlds (since the PI=/=SI). So the identity of R with the neural property N is SI necessary but not PI necessary just like all of the other scientific identities. So your externality intuitions are produced by the truth that anything that satisfies the PI of R will have externality, matching, ect. And your separation intuitions are produced by Slug being in one of the possible worlds where the PI of R does not pick out N.

    Notice that if we neglect this distinction we could give an exactly parallel argument for the (false) conclusion that [](water=H20) is false.

    Now one can, and most do, deny that the PI and SI of phenomenal terms can come apart in this way. But there is some evidence that this happens (pain asymbolia for instance) and higher-order theories of consciousness predict that this would be the case so this is a viable strategy.

  6. Hi Richard, thank you very much for chiming in. Also, thanks very much for all of your work in organizing the conference, which I am enjoying and learning much from.

    Roughly, my argument is as follows:

    1. R necessarily represents roundness
    2. N does not
    3. So R is not N

    In two-dimensional semantics, the primary intension of an expression is only relevant to evaluating certain *epistemic* claims in which the expression figures (for instance, the claim that a certain sentence-token a priori, or the claim that someone believes so and so). But 1 and 2 are evidently not claims of this kind; they are (de re) modal claims. So the primary intension of ‘R’ can have nothing to do with the evaluation of the premises of the argument – or its validity. In particular, contrary to what you suggest, the argument in no way requires that the primary intension and the secondary intension of ‘R’ coincide. (A side-note: ‘R’ is not an ordinary English term of commonsense psychology, but a technical term. Given my stipulation, the primary intension of ‘R’ might naturally be glossed with a description along the lines of ‘the salient experience property common to so-and-so cases’. Alternatively, I might stipulate that ‘R’ refers to a phenomenal concept in Dave’s sense (in his ‘The Content of Epist. of Ph Belief’), in which case it will be epistemically rigid. In any case, contrary to your suggestion, it cannot be glossed with a description based on commonsense platitudes about causal roles in the Lewis way.) Similarly, the primary intension someone associates with ‘water’ is not relevant to the evaluation of a de re modal argument he puts forward about water. I should add a brief note about this issue to my paper – thanks very much for bringing it up.

  7. I should have said that I might stipulate that ‘R’ *expresses* (rather than *refers to*) a phenomenal concept (since of course it refers to an experience-type) – in which case it is epistemically as well as subjunctively rigid. But – to repeat – nothing in the argument turns on the primary intension of ‘R’, for the reason I gave.

  8. I’ll have to think about this a bit more as it really does seem to me that the force of the externality and matching intuitions comes from taking the primary intension of R (I am happy to let you gloss it your way…though I don’t see what’s wrong with my way). But if you really just mean to be talking about secondary intensions then my intuitions go back to (v) in your response to Tom and Doug…I guess I should go back and re-read your response to the BIV cases…

  9. I went back to look at the ‘strange cases’ section but ended up re-reading the entire paper. In setting up the externality case you say

    Since ‘R’ is a rigid designator, the de dicto externality intuition entails the
    de re claim that R has the following (second-order) property necessarily:

    The externality property: being a property such that, if anyone has that
    property, then he has an experience as of a round thing at a certain viewer-relative distance d and position p

    Isn’t taking the de dicto to entail the de re just taking the PI and SI to coincide? And that is what takes you from the harmless platitude (the de dicto intuition) to the thing which causes the problem for the biological theorist (the de re claim). And in so far as the intuition is de dicto it involves primary intensions, right?

  10. Hi Richard,

    Thanks for your thoughts. Four points in reply.

    First, the move from the de dicto modal claim to the de re one does not require that the PI and SI coincide. For instance, ‘Necessarily, water is H20’ entails ‘Water is such that it is necessarily H20’, even on 2Dism. The validity of the move only requires that the relevant term be a rigid designator.

    Second, the de dicto versions are enough to undermine the biological theory. I just move to the de re because I like presenting the arguments in the form of applications of Leibniz’s Law. As I say in a foot-note on counterpart theory in the objections-reply section, I could just as well state the argument without using Leibniz’s Law, just relying on the de dicto externally-directed intuitions. So I don’t really even need the de dicto to de re move.

    Third, I could also have just started out by stating the intuitions de re: R is such that necessarily one who has it has an experience as of a round thing, etc. In that case, again, I wouldn’t really need the de dicto – de re move.

    Fourth, again ‘R’ is a technical term and so its semantic properties are up for stipulation. In the paper I only say what its secondary intension is – when I say it rigidly picks out a certain experience-type. I could stipulate that it expresses a phenomenal concept in Dave C’s sense, in which case the PI and SI coincide. There’s really no problem because of my first three points. But if there were, I could just fall back on this move, and the argument would not be affected.

    Hope that helps!

    Btw, I gather you like HOT theory. Of course, the proponent of HOT theory needn’t accept the biological theory. He could accept some kind of functionalism or intentionalism – in which case he’d have no problem with my argument and could accept any externally-directed intuitions he wants (see sect 6 of my paper for relevant discussion).

    Thanks again for the thoughts.

  11. OK, so let’s start simple: Because I have no idea what “as of” means, I can’t regard any claim about “as of”-ness as obvious or sacrosanct. If the “as of” claim is topic neutral, then I suppose it is a true comparative claim: “What is happening now is like what would be happening if….” But if the “as of” claim is “externally directed” such that it is essentially a source of belief or justification, then I don’t accept it. Why should I? Only, as far as I can see, if I have certain commitments about the semantic-cum-epistemological duties of sensations. Kant and McDowell give transcendental reasons along those lines. But Pautz suggests that those are not his reasons. So I don’t get it.

    Now, I know that people use the expression “as of” all the time. But that is no recommendation. I honestly have no idea what it means if it doesn’t either mean “similar to” (topic neutral) or “of” [while withholding judgment on veridicality] (sem/epist). The first is no problem for a type identity theorist; and the second we deny is an essential feature of sensations. If “as of” means anything else, I don’t know what it is.

  12. Hi Tom, thanks very much for your original comments and for this follow-up – they’re a huge help to me.

    I agree that ‘as of’ is a quasi-technical term – even if it is often used. That’s a good point, I think. In the long (pdf) version of my response to your comments, I tried to explain it. But, in a way, the whole issue can be side-stepped. For, as I also pointed out, ‘as of’ can be entirely removed from the statement of the externality intuition, so that it reads:

    (#) Necessarily, R is an experience of a round thing at a certain place.

    I and many others say that (#) has a true reading. It is commonly pointed out that there is a success-neutral, or “bracketed” reading of ‘has experience of an F’ on which it does not require that there be an F present. For instance, I can say that I had an illusory experience of a red thing, or a hallucination of a pink elephant, etc. It is relative to such a ‘success-neutral’ reading that (#) is true.

    Now, as I said in my response, I was just using ‘as of’ to indicate that I had in mind the success-neutral reading. But if ‘as of’ causes problems, then we can simply work with (#). Now, one cannot object against (#) that “I have no idea what it means”. It is just ordinary English (aside from ‘R’, which has been explained).

    Btw, the success-neutral issue is not very important. For one can formulate claims in the same ball-park using obviously success-neutral language, for instance

    (@) Necessarily, if an individual has R, then something looks round to the individual.

    But let me just stick with (#), because ‘has an experience of an F’ clearly has a success-neutral reading.

    Another point is this. My argument just requires that (#) has a true reading. This does not put me under an obligation to provide a detailed theory of what that reading is. In particular, I am not under an obligation to say that it is Tom’s semantic/epistemic reading (which, btw, I am not sure I followed). As G. E. Moore said, you can know something without knowing its analysis. For instance, you can know something framed in terms of ‘knows’, without being in the position to provide an analysis of ‘knows’. Similarly, we can find it very intuitive that (#) has a true reading, without having a theory of (#).

    So far, so good: many find it extremely plausible that (#) has a true reading. Now, Tom says ‘experience of a round thing’ has a topic-neutral/comparative reading. Now the claim that it has this reading is a very controversial semantic/philosophical claim (it is denied by Jackson in *Perception* on the basis of counterexamples). But let me just play along. (Btw, I just use ‘externally-directed property’ as a tag for whatever properties are expressed by the predicates used to formulate the externally-directed intuitions (e. g. ‘is an experience of a round thing’), leaving it completely open what they are. So if they include comparative properties, then those count as ‘externally-directed’ – contrary to Tom’s claim that these comparative properties cannot possibly count as externally-directed.)

    Now Tom says that “[the comparative reading] is no problem for [the biological theorist]”. I didn’t follow this. We can still construct an argument against the biological theory. (To make things simple, I will not attempt to state the argument in the form of an application of Leibniz’s Law.) On the comparative reading, together with Tom’s biological theory according to which R is necessarily identical with N, (#) is true iff

    (#a) Necessarily, neural state N is the kind of state typically caused by round objects.

    And (#a) is evidently false. So, if (#) has Tom’s topic-neutral/comparative reading, then, on Tom’s biological theory, it is false relative to that reading. (Notice that all this is all already in my paper! As many have pointed out, Smart’s topic-neutral/comparative account is basically a functionalist account. And in the paper, I discuss functionalist accounts of the externally-directed properties at length. I argue that, on such accounts, the externally-directed intuitions are inconsistent with the biological theory.)

    Now, Tom identifies another reading he thinks (#) has – the ‘semantic’ reading (which again I am not sure I followed – but that is not important). He also admits that relative to that reading (#) is false.

    In general, on Tom’s theory, (#) has no true reading. Indeed, the title of one of the sections in his and Douglas’s comments is “The externally-directed intuitions are ambiguous and false”. (Right? Maybe I’m missing something.) *This is the crucial point.* For this is a serious cost of the theory (in my view, a decisive cost – but at least a serious cost). Because most folks would say that it is *extremely* plausible, pretheoretically, that (#) has a true reading. And they are (I think!) right.

    Similarly, as Tom admits, the biological theory is inconsistent with the matching intuition. This is another cost of the theory (in my view, a decisive cost – but at least a serious cost). Because the matching intuition is almost universally accepted and regarded as having strong pretheoretical support. Similarly for the grounding intuition and the justification intuition. That the biological theory violates these intuitions across the board – at least some of which enjoy nearly universal acceptance – is a cost of the theory that goes beyond its violation of the much more controversial ‘multiple realizability intuition’.

    Now Tom and Doug might disagree. They might not regard this as a serious cost. They say in their comments ,“We’re not sure what the big deal is. Plausibly, nothing at all essentially has the kinds of externally-directed properties that concern Pautz”. If they were right here – if there were some argument for accepting this sweeping claim – then of course the fact that the biological theory violates the externally-directed intuitions would not be a great cost of the theory: every theory would violate them. But, as I say in the paper and in my response to Tom and Doug, I think there are problems with this sweeping claim, and they did not give an argument for it. For instance, the property *believing that a round thing is present* is a clear counterexample. It has such properties necessarily, contradicting their sweeping claim. For instance, this property is such that, necessarily, if an individual has it, then he is in a state that matches the world only if a round object is present. And, in section 6 of my paper, I argue at length that the kind of intentional properties which with Tye and Dretske and Chalmers would identify R might have all four externally-directed properties necessarily – thereby supporting their theories over the biological theory. Also, as Shoemaker explains in his comments, certain other physical (or physically realized) properties might have such properties necessarily. They are just not biological properties: they involve the environment in certain ways. This is why Shoemaker agrees that the externally-directed intuitions rule out the biological theory. For these reasons, Tom and Doug’s sweeping claim seems false. Would they Tom and Doug let me know their argument for this strong claim? Obviously, it is an argument I need to address!

  13. About this: “As many have pointed out, Smart’s topic-neutral/comparative account is basically a functionalist account.” It is true that Lewis gave it that reading, and looking backward that can seem like a natural reading of Smart’s intent. But I am not in the least sure that a nascent functionalism is really there in Smart’s work. That is part of my concern. I would be much mollified if one could show that the Smart analysis really generates a full-blown functional analysis.

    This is not precisely on topic, but I think it would be helpful to know whether taking Smart’s road to type-identity really lands one in Lewis.

  14. Hi Doug, thanks for the comment. My remark “As many have pointed out, Smart’s topic-neutral/comparative account is basically a functionalist account” was just made in passing. I think it is correct, but I nothing of substance hangs on it. The crucial point, it seems to me, is that on the biological theory all the externally-directed intuitions come out false. Since many find at least some of these extremely plausible (and in fact you and Tom say some are ‘plausible’), that seems to be a serious cost – and one worth pointing out. Now maybe you and Tom do not think it is a serious cost, because “Plausibly, nothing at all essentially has the kinds of externally-directed properties that concern Pautz”. But, again, that sweeping claim seems open to counterexamples involving belief properties. And there are non-biological theories of experience that accommodate the externally-directed intuitions. (It might help me to know your argument for the sweeping claim, so I could address it.)

    (Btw, you seem to think that I was saying that Smart’s account yields Lewis-style functionalism about experience. This is not so. I was talking about a Smart-style account of the ‘experience of’ locution. I was just saying that the Smart-style account of the ‘experience of’ locution might be described as a functionalist account of the ‘experience of’ locution. I did not mean that this account of the ‘experience of’ locution yields a (Lewis-style or other) functionalist account of experience itself. In fact, what was under discussion was the conjunction of this account of the ‘experience of’ locution and a non-functionalist, biological account of experience. Hope that makes sense…)

  15. Mostly I am sympathetic with Adam’s argument here. But a few small comments.

    (1) On the dialectic with Richard: I think Richard could reasonably point out that one time-honored strategy (e.g. Kripke’s) for dismissing apparent modal intuitions is by pointing to nearby epistemic/contextual intuitions (in effect
    based on primary intensions) instead. Combining this Kripke/2D model for explaining away modal intuitions with the Lewis-style view of phenomenal concepts (where they have functional primary intensions, biological secondary intensions) is a move to be taken seriously. Of course Adam can reasonably reply that with the epistemic/modal and 2D distinctions in full view, he still clearly has the modal intuition (not just the epistemic/contextual one). The opponent could come back by saying that the apparent support for the modal intuition is really support for the epistemic/contextual one (and by saying it’s illegitimate to stipulate a rigid primary intension for phenomenal concepts when it’s controversial whether there are any phenomenal concepts that behave like this). I think that Adam is ultimately right here, but there’s room for an opponent to make at least a little bit of hay.

    (2) Re the dialectic with Tom: I think there’s a bit more to be said about giving comparative readings to (@) and (#). I think that insofar as there are comparative readings at all, one natural comparative reading of ‘looks X’ is something like ‘looks the way X things actually look to me’. If we give that reading to (@) and (#), that explains a sense in which they are true without giving any support to Adam’s argument. Again, Adam can argue that they are true in some other sense — but this takes a bit more work than the above.

    (3) Re the dialectic with Sydney: I think there’s a bit more to be said for the BIV argument here. I’ve argued elsewhere that there’s a strong case that Matrix-style BIV’s have true beliefs and in particular have beliefs about roundness*. Now as Adam observes, this doesn’t entail that BIVs don’t have experiences representing roundness (or some other property that we represent such as Edenic roundness), and indeed my own account has it both ways by saying a BIV’s experience represents both roundness* and Edenic roundness. Still, it’s at least a little costly to appeal to two layers of experiential content. And there’s certain a prima facie case that the contents of belief should mirror the contents of experience. So there’s at least a prima facie three-premise argument here for the conclusion that BIVs “shape” experiences represent shape* properties and not properties represented in our shape experiences. And I think that the resulting view is a pretty respectable one (it’s basically Brad Thompson’s view). Of course the view has to deny a certain phenomenological intuition about commonness of properties represented — but this intuition is now getting fairly abstruse and is perhaps deniable (especially if one can accomodate something close by pointing to common Fregean content). At least, the costs of denying the intuition need to be weighed against the costs of denying one of the three premises, and I don’t think it’s completely cut and dried which wins.

    (4) My own view is that any biological theorist should be a “big-state” theorist. Even without Adam’s argument, small-state theorists are subject to the objection that it’s very implausible that a small state in a Petri dish could support a visual experience like ours. Now Adam’s argument adds extra force and rigor to this sort of objection to a small-state theorist. But in any case, big-state theorists clearly aren’t subject to either this sort of argument or Adam’s. And I think that quite a few biological theorists (e.g. Searle, and Ned Block in some moods) are best understood as big-state theorists.

  16. Hi Dave. Thanks very much for all of these points (and the expression of sympathy for my argument). Let me try out some reactions.

    (1) Re the Kripkean proposition confusion strategy: I have three points. (i) I introduce ‘R’ as a rigid designator of the salient experience property Adam has in a certain case C. For the sake of argument, let me just grant that “it’s illegitimate to stipulate a rigid primary intension for phenomenal concepts when it’s controversial whether there are any phenomenal concepts that behave like this” (even though we are both believers in phenomenal concepts that behave like this). Even so, as I said in response to Richard, it look as if the primary intension will not be a rich functional intension; it will simply be a meager, case-based primary intension, glossed by the description ‘the experience property Adam has in C’. (Why should it be that the only way we have of thinking of such a property is via a rich functional primary intension?) Now, if the biological theorist adopts the Kripkean proposition confusion strategy, then he says that, when I seem to be intuiting that an externally-directed claim is necessary, I am really only recognizing that the primary proposition is necessary. But if the primary intension is a meager, case-based primary intension, then this attempt at explaining away the intuition of necessity does not work. Because in that case the primary proposition is something like: that the experience property that Adam has in C has externally-directed property P. And that’s certainly not necessary. (ii) To avoid the Kripkean proposition-confusion ploy, I can use a trick that Lewis uses in *Plurality* (at p. 250). I could introduce ‘R’ but from then on just use the pronoun ‘it’. Then the intuition is: *it* is necessarily an experience of a round thing, etc. It is difficult to see how the Kripkean proposition confusion strategy could get off the ground here. For here I am not referring to R under a rich functional primary intension. (iii) In general, the Kripkean proposition confusion strategy seem implausible to me. In fact, it is in tension with famous remarks Kripke made elsewhere in *Naming and Necessity*: “Don’t ask: how can I identify this table in another possible world, except by its properties? I have the table in my hands, I can point to it, and when I ask whether it might have been in another room, I am talking, by definition, about *it*.” (52-3). Similarly, I think can ask of R: must *it* be an experience of a round thing, etc.? I think it must be, and I just find it implausible that I am confused about what I am considering.

    (2) Re the actuality-based move: I call this actuality-based move the ‘rigidification ploy’ in my paper and already criticize it at length in note 29. The externally-directed claims just cannot be analyzed in this way, for several reasons. I won’t repeat those points here but I’d be grateful to hear what people think of them. (I am not sure, btw, that Tom had this response in mind. On this response, the externally-directed intuitions come out true in a sense, but quite compatible with the biological account. By contrast, in their comments Tom and Doug repeatedly say that they are false.)

    (3) Re BIVs: (i) I gather that the argument is: the BIV’s belief accurately represents roundness*, not roundness; the content of the belief mirrors the content of the experience; so the experience R represents roundness*, not roundness. As we’ve talked about before, I myself do not accept your argument that the BIV has true beliefs and in particular has beliefs about roundness*. Clearly, the BIV has a false belief, when it has a tomato-like experience and says ‘there is a round before me’! The BIV is the paradigm of a deceived individual! That intuition counts for something, and in my (humble) opinion the argument you give is not enough to undermine it (though I fully realize that I incur the responsibility of saying where it goes wrong). So, in effect, I deny the first premise of this argument. (Whereas you retain my shape claim in the face of the argument by rejecting the second premise, the mirror principle.) So I retain the mirror principle (which I find very plausible) by saying that that both the BIV’s experience R and its belief represent roundness. (ii) You say rejecting the first premise is a cost. But I don’t think it is great cost, because I find that premise very counterintuitive, as I just said. (Of course, I have to reply to your argument in your Matrix paper, but I have a handy excuse for not doing that: my comments have already been too long.) Further, I think the costs of the pure Fregean view (or any other view) which forsakes my roundness claim are much more serious. First, it forsakes the roundness claim. (Btw, at SPAWN a couple years back, even Brad said that was on firm ground! – making me confused about his stance on it.) Second, it does not accommodate the internal connections between the contents of experiences in different sense modalities, as you discuss in your ‘Eden’ paper. Third, it faces all of the problems you raise early on in your ‘Eden’ paper, which lead you to reject it (and accept my shape claim).

    (4) Re Big States: We have to distinguish between three theories. First, there is the local biological theory that R is necessarily identical with the local neural property L that comes and goes with R. Probably no one holds this. Second, there is the global biological theory that R is necessarily identical with a more global neural property G: undergoing L in neural context C. Since the laws of nature and the even wider neural context are contingent matters, property G is not necessarily apt to be caused by the presence of a round thing and apt to cause behavior appropriate to such a thing. Third, there is the big state theory that R is identical with *having neural property N and being so constituted, and living under such laws of nature, that N is apt to be caused by a round thing at distance d and place p and is apt to cause behavior appropriate to such a thing*.

    I agree that my arguments do not apply against the big state theory. My arguments apply to the local biological theory and the global biological theory; since the local biological theory is probably not occupied, in my paper I focus on the global version throughout. But, as I explain in my paper, the big state theory is not a version of the biological theory that is my target. And it is not the view taken by Block, (erstwhile) Hill, McLaughlin or Searle. It is way to close to functionalism.

    Hill (in *Sensations*), McLaughlin (in ‘A Naturalist-Phenomenal Response to Block’s Harder Problem’) and Searle (in *The Rediscovery of the Mind*) all explicitly accept the possibility of ‘absent role’ cases in which one has R (or the like) but one has no state that is apt to be caused by a round thing at distance d and place p and is apt to cause behavior appropriate to such a thing. Indeed, this is part of their case against functionalism and for the biological theory. This means that they do not accept the big state theory, for on that theory absent role cases are impossible. Block agrees (in ‘two neural correlates of consciousness’) with you that the pitre dish case suggests that experience-types cannot be identified with local neural properties; but he says (in ‘comparing the major theories of consciousness’ and elsewhere) that they can be identified with somewhat more global neural properties involving re-entrant processing. This suggests he accepts the global theory, not the big state theory. Further, he certainly thinks one could have R without its being accessible, and so I would imagine without its being poised to affect behavior (indeed in his most recent ‘Mesh’ paper he says the individual G. K. might have experiences that he/she is constitutionally unable to act on). Again, this suggests he doesn’t accept the big state theory, for on this theory having R necessitates having a state that is so poised to influence behavior.

    So all these authors probably accept something like the global biological theory. This theory is undermined by my arguments. Now you say “Even without Adam’s argument, small-state theorists are subject to the objection that it’s very implausible that a small state in a Petri dish could support a visual experience like ours.” Does the same point apply against the global biological theory that was my target? If it does, then my argument would have less interest, because then the global theory could be refuted without relying on my argument. But I don’t think it does. It is not crazy to suppose that, if you excised all my neuroanatomy but the quite *global* neural property with which these philosophers identity R, then I would still have R. I raise this and other problems with the pitre dish argument against the global biological theory in section 4 of my paper. I also argue at length that none of these problems afflicts my externally-directed arguments against the global biological theory. The pitre dish argument relies on a dubious intuition that does not withstand scrutiny; my externally-directed arguments rely on independent intuitions which I think are most more robust and less susceptible to argument. So (naturally!) I do not agree that my arguments just “add extra force and rigor to [the pitre dish argument]”. They succeed where the pitre dish argument fails.

  17. Evidently the title of one of the sections of our response is causing some confusion, but that is easily remedied. The title we gave the section was, “The ‘intuitions’ are ambiguous and false,” but to be more exact it should have been, ““The ‘intuitions’ are ambiguous, and false under the intended reading.” This explains why Doug and I can consistently say that the first two intuitions are plausible: we find them plausible under a reading other than the intended reading. On the alternative reading, the intuitions are not a problem for the identity theory, which we took to be the target of Pautz’s attack on biological theories.

    So, contrary to what Adam sometimes says in his response, we do not claim that the first two intuitions are false under both disambiguations. I take it this is worth noting because Pautz regards the costs of flatly denying the intuitions as too high, because too many people find them prima facie plausible. So we’re at least not crazy, according to him, if we give a reading of the intuitions on which they are plausible. Similarly, we do not say that the matching intuition is inconsistent with biological theories; what we say is that its epistemic disambiguation is (“corresponds to the way things are”), but that its non-epistemic disambiguation (isomorphism) is compatible with biological theories but not up to the work that Pautz requires of it.

    Now, what are the disambiguations? In our comments, Doug and I gloss them as follows:

    (a) being a property such that, if anyone has that property R, then he has an experience property like that which is had by me when I see a round thing at a certain viewer-relative distance d and position p.

    (b) being a property such that, if anyone has that property R, then he has an experience that presents the world as though there were a round thing at a certain viewer-relative distance d and position p.

    We’re not wedded to this being the final word on how to make the distinction. The important thing is that a distinction like this one can be made, and that Pautz argument depends on something like reading (b) where experiences are essentially such that they are poised for some epistemic and semantic duties, like being accurate (matching) and grounding or justifying beliefs. According to me and Doug, (a) is plausible, but it is just what I called above a comparative claim, a topic-neutral reference fixer. (a) does not imply that R has any externally directed properties essentially.

    Adam says that his argument works with something like (a), and he offers (#), (@), and (#a):

    (#) Necessarily, R is an experience of a round thing at a certain place.

    (@) Necessarily, if an individual has R, then something looks round to the individual.

    (#a) Necessarily, neural state N is the kind of state typically caused by round objects.

    Since “of” and “looks” have the same kinds of ambiguities that Doug and I point out in “as of”, I will focus on (#a) for the moment, which appears to be clearer. The problem with (#a), says Pautz, is that it is false. I assume he has in mind that causation is contingent and relational, so it’s not true that any state is necessarily such that it is typically caused in some way or another. But this is just why (#a) is not a good substitute for our (a). (a) does not say that R is necessarily caused in a certain way; rather, it says that R is like what happens when I have states that are typically caused in a certain way. The comparison is important because that is what takes the brunt of the essentiality claim, thus insulating (a) from the fact that nothing is essentially such that it is caused in a typical way. And, as Dave notes, the comparative is neutral on “location” of the property to which R is compared, and so does not carry any essential external-directedness. (It’s not the actuality that does the work; that’s a red herring.)

    Now Shoemaker, Doug, and I all think that for Pautz’s argument to go through his externally directed properties have to be something like epistemic, semantic, representational, or intentional properties, or at least proto versions thereof. Since I (like most philosophers) don’t think anything has intentional or representational properties essentially, I say that nothing has externally directed properties essentially. Pautz finds this silly, pointing out that beliefs and representations have externally directed properties essentially. Here a number of things might be said. I suppose there is no trouble with the de dicto necessity of “representations have representational properties essentially.” Whether I should also accept that there is a de re necessity of the representationalness of beliefs, I am not so sure. But more importantly for present purposes than whether Doug and I engaged in some rhetorical hyperbole is what we learn about Adam’s argument from his reply. First, as everyone knows, type-identity theorists rarely if ever offer that theory as a theory of belief or other intentional or representational states. So beliefs per se are irrelevant. Now, if representationalism is the correct theory of sensations, than the identity theory is probably false. But, second, Pautz purports to be arguing that biological theories of sensations are false on pretheoretical grounds, not by way of a defense of representationalism. Moreover, he tries to argue that his externally-directed properties are not representational/intentional properties. If the argument were just that sensations are essentially intentional, and that brain states are not essentially intentional, then I still might object but the argument not be new.

    So, if he Adam to slap our wrists for over-playing our claim that denying his intuitions would be no problem for a biological theorist, we’ll take the correction. But the cost appears to be that he is conceding that the kinds of externally directed properties he has in mind are representational. And this is tantamount to conceding that the intuitions are theoretical claims, not pretheoretical Moorean truths. Then he needs an argument.

    This is why our comments focus almost entirely on the “set up” of Pautz argument, and not his detailed refutations and accommodations of various theories and objections. He needs to make the biological theorist accept his starting points for the whole operation to get off the ground. In a footnote Doug and I compare Adam’s argument with that of Kant or McDowell; in his reply, he indicated that he wishes to distance himself from those kinds of arguments. Frankly we think he would be better served by actually arguing, with McDowell, that a proper theory of experiences must explain how they can essentially be such that they (and not any pre-mental animal states) are poised for epistemic and semantic duties. This, as Doug and I have said, would probably involve distinguishing “experience” from mere “sensation.” At that point I would say that I am only aiming to give an account of sensation. But Adam would be in a position to dismiss the significance of any account of that, for that would be uncontroversially biological—it would just be the claim that the in-the-head part of experiences (“sensations”) are identical to things in the head.

    Here’s another way of talking about the same kinds of ideas: I think it’s actually quite useful and important that Pautz mentions, if only in a footnote, that “a thread running through much recent work on perception is an emphasis on role of experience in making possible, and justifying, external thought” fn. 4 (p.2). I claim that biological theorists are not in that group; for one thing, biological theories of mind (i.e., identity theories) are not first and foremost theories of perception at all. Of course, ultimately a biological theorist (or, rather, the group of us) will need a compatible theory of perception. What I deny is that a criteria for a theory of the ontology of sensations (e.g., an identity theory) is that it be such that the states it discusses are essentially such that they are able to do a job in a theory of perception. And that doesn’t bother me in the least.

    Of course “small-state” theorists, as Dave called us, do have to explain or explain away some odd consequences; but every theory has some odd consequences, even and especially the true ones. (Parallel lines can cross?!?) I don’t think Adam has shown that there is a problem we identity theorists hadn’t considered and that an identity theory is entirely unable to handle.

    (NB: I will be traveling until Sunday, so my responses between now and then will be slow and/or non-existent.)

  18. 1. What are the True Readings?

    Hi Tom, thanks very much for the thoughts. In your initial comments on my paper it seemed that you were agreeing with me that on the biological theory Externality and Matching are false, but said “this is no big deal”. Now you say that the biological theorist can supply them with true readings. That’s progress! But would you please specify the true readings you think they have? Let me say why it seems to me that you have not sufficiently explained the readings relative to which Externality and Matching are true, on the biological theory. Again, Externality and Matching might be formulated as follows:

    Externality: Necessarily, R is an experience of a round thing at a certain place.

    Matching: Necessarily, anyone who has R is in a state that I-matches the world only if a round object is present at a certain place.

    Here I use ‘I-matching’ to indicate the intuitive sense of matching I introduce in the paper with examples. In that sense, when one has R, one is not in a state that matches the world only if an elephant is present (even if there is some function from the state onto elephants); rather one is in a state that matches the world only if a round thing is present.

    Now I can say why I think you have not sufficiently explained the readings relative to which Externality and Matching are true, on the biological theory. Let us start with Externality. You say that the true reading is not my (#a). That cannot be the true reading, for the simple reason that (#a) is not true – you agree here. Then you say “this is just why (#a) is not a good substitute for our (a)”, where (a) is as follows:

    (a) being a property such that, if anyone has that property R, then he has an experience property like that which is had by me when I see a round thing at a certain viewer-relative distance d and position p.

    But (a) is definitely not a possible reading of Externality. It is not even a sentence! And, unlike Externality, it does not contain a modal operator. Dave offered you a true reading of Externality: a kind of actuality-based reading. But you don’t accept that reading either. You wrote “It’s not the actuality that does the work; that’s a red herring”. Further, in my paper (at foot-note 29) I argue at length that the actuality-based analysis of Externality cannot be correct, and that this strategy does not in any case provide a general solution to my arguments.

    So: what, in your view, is the true reading of Externality that the biological theorist might supply?

    Now let us turn to Matching. In your initial comments you wrote

    “Pautz glosses “matches” as “corresponds to the way things are” (p.6), which is a plainly epistemic or semantic notion of matching [AP – small point: I did not why this locution is plainly epistemic or semantic – for instance, as I show in the paper, a disjunctivist could provide an analysis of this notion and Matching without appealing to content]. But **according to the biological theorist, conscious mental states do not have any such properties essentially**. The raw content of experience may or may not have a structure that “matches” the world in that it is isomorphic to and/or caused by the world in various ways. But **plainly those would be contingent, and it would be false that those matchings would occur “only if a round object is present”** —for everything is isomorphic to indefinitely other things, and thereby “matches” them in indefinitely many ways that are not epistemic/semantic.”

    Now you say:

    “We do not say that the matching intuition is inconsistent with biological theories; what we say is that its epistemic disambiguation is (“corresponds to the way things are”), but that its non-epistemic disambiguation (isomorphism) is **compatible with** biological theories.”

    I’m sorry, but I’m a bit confused. In the first passage, you are saying that Matching has no true reading, on the biological theory. Now you are saying it has a true reading. In view of the first passage, it is difficult to see what might be the true reading it has, according to you. Could you please say what the true reading is, according to you? (Remember, Matching is formulated in terms of I-matching, which I explain in the paper. Foot-note 32 of my paper might be relevant here.)

    2. Are Matching and Externality Theoretical Claims?

    You wrote:

    “[Adam] is conceding that the kinds of externally directed properties he has in mind are representational. And this is tantamount to conceding that the intuitions are theoretical claims, not pretheoretical.”

    Well, I said in my note 7, and a few times in my comments, that Externality and Matching are neutral on such theoretical issues and entirely pretheoretical. Externality is just framed in ordinary English: it is not a theory of any kind but a pretheoretically plausible claim. Matching is formulated in terms of I-matching, a pretheoretically intuitive notion that I explain in the paper.

    Again, my claim is: Externality and Matching are true, that is, have true readings. That is pretheoretically extremely plausible. That is all I’m claiming, and all my argument requires. Before you seemed to be saying that, on the biological theory, they don’t have true readings – which would be a serious cost. Now you seem to be saying that even on the biological theory they might have true readings. But, what are those true readings?

    (Btw, you mention a foot-note where I say that my arguments concern the grounding and justification of thought – an important theme in the philosophy of perception. I only had in mind my grounding and justification arguments. I do not think the externality and matching argument concern the grounding and justification of thought.)

    3. Is My Argument Already Made Somewhere?

    You wrote “we don’t think Adam has shown that there is a problem we identity theorists hadn’t considered”. In your initial comments you wrote “Pautz . . . reinvented Black’s objection to Smart’s identity theory.” You also say that my argument is similar to an argument in McDowell. But, for reasons I have explained, my argument is not Black’s, nor is it in McDowell. And it’s also quite different from, and independent of, the pitre dish argument. (In fact, as I say in section 4 of paper, I agree with you that this argument fails. We agree on something! Sweet!! ☺.) Are arguments of the kind I give really made anywhere else or considered by biological theorists? Where did you have in mind? If so, I should address what has been said about this in my paper.

  19. OK, I shouldn’t have posted without rereading the paper! I missed footnote 29.

    Re BIVs: I know you reject the Matrix arguments and you don’t think they have beliefs about roundness*. I was addressing the part of your reply to Sydney where you allow for the sake of argument that he is right about roundness* and resist the move from there to the claim that they don’t represent roundness (or whatever) — just observing that there are costs to that resistance.

    Re big states: OK, I wasn’t really clear on how you were understanding the big state vs global state distinction. The big states I had in mind are all internal states. Roughly you could think of them as involving at least the kind of internal goings-on that would constitute a supervenience base for an internalist functionalist, plus further specifications, e.g. involving biological realization. I guess you’d count that as a “global” state rather than a “big” state. Anyway, I don’t really see the argument that a global state G like this couldn’t represent R necessarily. I guess part of the argument is that if G is internal, there are possible worlds where it is dissociated from being caused by R and causing R-appropriate behavior. I take it that the global state theorist will allow that there are such worlds — worlds with weird laws of nature or weird things going on between brain and behavior — but will hold that there are worlds in which G is R-directed (in the original broadly representational sense) all the same. I take it that then the objection is that they can’t adequately explain this, i.e. explain why G is R-directed in those worlds. I don’t really see this. Obviously a causal theory or a crude behavioral theory won’t explain this, but this theorist can appeal to all sorts of internal things, especially rich internal functional roles (certain sorts of structurally spatial processing go with representing space, and so on). Maybe you think that no appeal to internal processing can explain things adequately here — but then you really have an argument against any sort of reductive internalism (including narrow functionalism), not just against biological theories, and presumably that argument needs a bit more explaining.

  20. Thanks very much, Dave.

    Re BIVs: I see, thanks.

    Re G: Yes, I think that biological theorists can’t adequately explain this, i.e. explain why G is R-directed in those strange worlds. In particular, I think that they would need to accept the kind of “primitivism” about the sensory representation relation (the “directedness relation”) that I have devoted my life to arguing for. As you know, I myself accept this primitivist view. But I understand ‘the biological theory’ to be committed to a fully reductive theory of consciousness. So this primitivist view is not open to the biological theorist. (If he accepted it, I’d no longer call him a ‘biological theorist’, and his view would become very similar to my own.)

    I argue that the claim that G might be R-directed in those strange worlds requires primitivism in a couple of places. I argue for it in the paper we’re now discussing in my response to the “second objection” about magical theories. I say a bit more in note 32. I also argue for this claim in my paper ‘Phenomenal Intentionality as Irreducible’ (https://webspace.utexas.edu/arp424/www/simple.pdf), esp. section 7.

  21. Hi Dave, I just wanted to make a clarification that is present in the paper but that I should have also made in response to your comment. On the view you have in mind, G somehow necessarily represents roundness without qualification – that is, *even in separation cases in which it is not apt to be caused by a round thing at d and p nor apt to cause behavior appropriate to such a thing.* In my paper, I call this *unqualified internalism about sensory intentionality*. Then my Leibniz’s Law would fail, even if R necessarily represents roundness.

    My reply in the paper was that from the biological theorists point of view unqualified internalism about sensory intentionality is unworkable because there is no reductive theory of the two-place sensory representation relation compatible with unqualified internalism. Call this the *unworkable response*.

    Now Dave wrote: “Maybe you think that no appeal to internal processing can explain things adequately here — but then you really have an argument against any sort of reductive internalism (including narrow functionalism)”.

    That might suggest that unqualified internalism about sensory intentionality is a widely held view. So it might suggest that it would be very hard to establish my unworkable response. Here is where the clarification comes in. Unqualified internalism is in fact not a widely held view.

    In general, as I explain in the paper, what ‘internalists’ about intentionality (or some forms of intentionality) accept is *qualified*. Thus, Lewis (in e. g. in *Reduction of Mind*) and Jackson (e. g. in *Phil of Mind and Cognition*) imply that G has the relevant content, *only provided that it is in ensconced in a certain type of system, and certain laws of nature prevail, so that it plays a certain functional role with respect to other internal states, behavioral outputs (narrowly characterized), and external states of affairs*. What makes Lewis and Jackson internalists or narrow functionalists about (some aspects of) intentionality was that they say that the relevant functional role does not bring in facts about the subject’s actual environment or history. Further, they say that the relevant functional role concerns *potential* interactions with the environment, so that a state of a brain in a vat might count as instantiating it. In consequence, Lewis and Jackson imply that the relevant, content-determining input-output functional role of G in humans is determined by the characteristics of total internal system in which is ensconced together with the laws of nature (as well as facts about ‘the appropriate population’) – in that sense it is narrow even if world-involving.

    In fact, in general, internalists about (some aspects of) intentionality almost always accept this kind of qualification. Thus, in his internalists days, Fodor only said that internal duplicates support the same (narrow) intentional states, only provided that they agree in their behavioral dispositions (narrowly construed). As I discuss in the paper, such qualified internalists about intentionality (by contrast to unqualified internalists) have no commitment to the counterintuitive claim that, even in strange cases (like Fischer’s pulsars case or my separation cases) in which our neural states play radically different roles, there is sameness in attitude and content.

    Unqualified internalism about sensory intentionality is quite different. Again, it says that G somehow necessarily represents roundness without qualification – that is, *even in separation cases in which it is not apt to be caused by a round thing at d and p nor apt to cause behavior appropriate to such a thing.* This view is more or less unoccupied, I think. Maybe John Searle holds it, but he has been roundly criticized. I think it is very implausible. I think Dave is right that there might be isomorphisms of some kind between neural states like G and the represented shapes and other properties, but, as I explain in my paper and in more detail in sect. 7 of the ‘Ph Intentionality as Irreducible’ paper linked to above, they will not get the unqualified internalist a reductive theory of the two-place sensory representation relation (for reasons concerning the ‘scrutability of reference’). In sum, I just wanted to make it clear that my unworkable claim only concerns the kind of extreme (nearly unoccupied) unqualified internalism about sensory intentionality that would be needed to save the biological theory, and not (somewhat more plausible) typical, qualified forms internalism. I think the considerations I give in my papers for this unworkable claim are pretty strong, but I would be happy to hear any thoughts you might have.

  22. Adam, I thought I agreed with Dave’s last two comments. But then I was confused by your response to them. I’m also confused about why you say that Doug and I claimed that the identity theory rejects all readings of the Externality Intuition; we claimed that it is ambiguous and then argued that one reading is plausible but not a problem for the identity theory. Plainly you find our approach equally confusing.

    So I am not sure I can diminish our mutual perplexity, but here it goes. I’ll first try working it out with the biological view (aka, identity theory), as Doug and I did in our response. But then I’ll try it with a view that neither you nor we hold, to see if that helps.

    I. The Identity Theory

    I gathered that the Externality Intuition is: Sensation R necessarily (essentially) has the Externality Property. You say the Externality Property is the following second-order property of R: “being a property such that, if anyone has that property [R], then he has an experience as of a round thing at a certain viewer-relative distance d and position p” (5).

    Doug and I say that there is a reading of the Externality Intuition that is plausible, not a problem for the identity theorist, and does not support your Leibniz’ Law argument. That reading understands the Externality Property in terms of something like (a) rather than something like (b). So the Identity Theory Friendly Reading (ITFR) of the Externality Intuition is, to a first approximation:

    (ITFR) If one has an experience with R then one has a property [R] such that, necessarily (essentially), if anyone has that property, then one has an experience property like that which is had when one sees a round thing at a certain viewer-relative distance d and position p.

    Nothing about (ITFR) is a problem for a biological (inside-the-head) theorist. The biological theorist (Identity Theorist) holds that R is identical to some inside-the-head/body property N, where that identity is a posteriori and necessary. And the Identity Theorist is happy to say that N is necessarily such that having it is similar to (“like”) having a property that is had by me when I see a round thing at a certain viewer-relative distance d and position p. Of course (ITFR) mentions facts that are not in the head; but here the use of things “outside the head” is merely to fix the reference of the “inside the head” property to which R is compared. If you like, we can eliminate the mystery by resolving the reference fixer and just putting N in that position.

    Now presumably it is a contingent fact that N is typically/normally/optimally caused by a round thing at a certain viewer-relative distance d and position p. But that is not problem. For R, also, is merely typically/normally/optimally caused by a round thing at a certain viewer-relative distance d and position p.

    What matters is that N is necessarily identical to R. That is enough to preserve the claim that R is necessarily like the property that (contingently) is had by me when I see a round thing at a certain viewer-relative distance d and position p.

    II. Humean Bundle Theory

    But forget this identity theory business. The question is about the Externality Intuition in general. Consider the following theory: Suppose we hold a substantival dualist bundle theory, of the sort that is often attributed to Hume: the mind is just a bundle of sensations, where those are independently existing entities. Having R is just having a certain one of those basic entities in your bundle. Suppose, perhaps contra Hume, we allow that there could be a primitive similarity relation among the sensation entities. In that case, R can still have (ITFR), but there is no reason to think that this kind of R is essentially externally-directed in the sense required for your argument to go through. Right?

  23. Thanks Tom. The problem is that ITFR is not true unless you use an actuality-based reading. Just go to a world in which individuals have R but no one sees red things. The only way to fix this is to change the relevant part of ITFR to “when one *actually* sees a round thing at a certain viewer-relative distance d and position p”. In effect, that seems to be what you are doing (even though you previously said that the actuality is not doing the work). As I explain in note 29 of my paper, I quite agree that if this actuality-based version of ITF is the only reading of Externality, then the argument from Externality does not work (even without the help of Hume). But I raise objections to the claim that the actuality-based version of ITF provides the correct analysis of Externality in that note. Any thoughts on these objections would be welcome. There I also argue that in any case the move does not defuse the Matching and other arguments, and so does not save the biological theory. It is not a sufficiently general strategy.

    Also, what is the *true* reading you now think Matching has? I didn’t quite follow because in your initial comments you wrote “plainly those would be contingent, and [on the biological theory] it would be false that those matchings would occur “only if a round object is present”.

    Finally, any thoughts on 2 and 3 in my comment 22?

  24. Hi Adam — some random thoughts trying to get the dialectic straight here. I take it that the main selling point of your argument compared to the Petri dish argument was that it refutes global-state theories of consiousness, not just local-state theories (with neither refuting not-fully-internal big state theories). One initial worry is that you convince me that no-one holds global-state theories of mentality, then you’ll convince me that the main selling point of your argument is that it refutes a view that no-one holds. But I take it that the thought is that lots of people hold global-state theories of consciousness while very few hold global-state theories of intentionality. So your argument could still interestingly refute the former.

    Now, I’m sure there are some people who hold global-state theories of consciousness without holding global-state theories of intentionality. I suspect that there are fewer who hold global-state theories of consciousness while denying global-state theories of intentionality, though. And I’m inclined to think this combination of views is already undermined by H&T-style arguments for the link between phenomenology and intentionality.

    Also, I think we probably need to disinguish the sociological claim that there are few developed purely internalist theories of intentionality from the sociologial claim that there are few pure internalists about intentionality. My guess is that there are quite a lot of the latter. Maybe Searle, H&T, Strawson, Farkas, and others. (I’d be at least somewhat sympathetic myself, at least about the first dimension of content, apart from the Fisher cases — but these push equally against pure internalism about consciousness.) It’s just that they haven’t developed theories of intentionality to match their views. But of course developing theories is hard.

    On the substantive point: I looked at the section on “magical” theories in your paper. The main argument there (apart from the title!) seems to be come to the observation that we don’t have a good theory to date and you can’t see how one would go. Now, I think these observations carry some weight, but then it’s early days yet, and no-one understands intentionality well. So it strikes me that these observations are challenges to the pure internalist rather than powerful arguments. Maybe I’ve missed the substantive argument in the vicinity — if so I’d be interested to see it.

  25. Hi Dave, thanks for the helpful comments.

    One response to my argument to is accept unqualified internalism: it’s *metaphysically necessary* that, if one has the physical property G, then (no matter what the laws or context) one stands in the sensory representation relation to roundness. (Btw, this is not important, but I’m not sure whether the people you mentioned would accept unqualified internalism in this sense. I thought at least some of them were neutral on the mind-body problem, whereas unqualified internalism is not – e. g. it entails Zombies with G are metaphysically impossible. I think only Searle clearly accepts this unqualified internalism.) This would save the biological theory, because now G as well as R would have the relevant “directedness” property with *metaphysical necessity*.

    I think maybe I was misleading – really sorry about that. I didn’t mean to say that unqualified internalism about sensory intentionality is false. I also didn’t mean to say that there is no interesting theory of intentionality compatible with it – sometimes you seem to think this is what I was after. Maybe H&T, Farkas, etc. accept unqualified internalism about sensory intentionality and there is a true theory of intentionality compatible with that – even if we cannot think of it. (Maybe, in your terms, representing uninstantiated perfect roundness could somehow be metaphysically necessitated by the physical property G.) As I tried to get across in my comment 19, I just say that there is no *reductive theory of the sensory representation relation in physical-functional terms* compatible with it. In this sense, I think that it requires a broadly non-reductive, “primitivist” account of the sensory representation relation (which btw is the kind of account I like – I’m into magic – and I thought you, Farkas etc might like, given your skepticism about reductive physicalism). If this is right, it means that unqualified internalism not compatible with the ‘biological theory’ as I understand it. Because it is part ‘biological theory’ as I understand it that there is a purely reductive theory of consciousness in *physical-functional terms* – no unreduced intentional relations, etc. (That’s all I meant by saying the response is unworkable – unqualified internalism is not something the biological theorist can accept – not that it is unworkable simpliciter.) If the biological theorist accepted this kind of primitivism, I’d no longer call him a ‘biological theorist’, and his view would become very similar to my own.

    It is true – as you point out – that I do not argue in detail in the paper that there is no reductive theory of the sensory representation relation in physical-functional terms that is compatible with unqualified internalism (making it unavailable to the biological theorist). But that is because it is a huge task and one I carry out a paper I direct the reader to in a footnote – the ‘Phenomenal Intentionality as Irreducible’ paper linked to in my comment 19 above, esp. section 7. Here my arguments are systematic – not simply that no one has devised such a theory yet.

    A couple of other small points. First, you say that “H&T-style arguments for the link between phenomenology and intentionality” refute the combination of the biological view (R=G) and the claim that G only contingently represents roundness. This is roughly my point. I am using H&T style inseparatist intuitions to argue against the biological theory. I do this by arguing that on biological theories it is very likely that G only contingently represents roundness (whereas on inseparatism R necessarily does). I do this with separation cases involving G, and by arguing that unqualified internalism (G necessarily represents roundness even in separation cases) is unavailable to biological theorists (again understood as an attempt at a purely reductive theory) because there is good reason to think it can only be developed in a non-reductive way. I was hoping that this is interesting because H&T style inseparatist intuitions have not (so far as I know) already been used to argue against the biological theory. And it’s different from the argument about Pitri dishes (or pulsar worlds) – and I think maybe a bit better since those arguments seem to me to rely on (in my view pretty shaky) functionalist intuitions, as I explain in my paper.

    Second, since decisive arguments are not to be found in philosophy, I’d be happy with a strong argument against the biological theory based on externally-directed (aka inseparatist) intuitions. So if the success of the biological theory (again construed as an attempt at a purely reductive physicalist theory) hinges on the claim that there is some *purely reductive physicalist* account of the sensory representation relation consistent with unqualified internalism (because only that would secure the inseparatist intuitions in a way compatible with the biological theory), and there is strong reason to doubt that claim, I think that’d go some way to showing that the biological theory is mistaken – it’d be a little bit of progress. And I do think there is strong, systematic reason to doubt the claim that there is some *purely reductive physicalist* account of the sensory representation relation consistent with unqualified internalism. Again they’re developed in most detail in the ‘Phenomenal Intentionality as Irreducible’ (esp sect. 7). Take a look if you’re interested – I hope those arguments have at least some force!

    Whew! Does any of this clarify my thinking on these issues (and seem reasonable)?

  26. Hi Dave and Adam. Thanks for the thoughtful comments! I have really been enjoying working through theses issues.

    Dave has definitely got the dialectic down right; though I would dispute how much hay could be made in the end 🙂

    Adam, you take the externality, matching, etc, intuitions to be a priori don’t you? If so there must be a necessary primary intension in play and that will allow the proposition confusion response to work.

  27. Adam, OK, I thought you were doing something else when you were talking about “actuality.” I thought you had to be, because if the topic-neutral reference fixer is not given an “actual” (de re) reading, then the theory is functionalism rather than an identity theory. So, if I now understand you, the “technical objection” that you call “the rigidification ploy” just is the a posteriori identity theory properly stated! (Compare the “rigidification ploy” in the claim that water = H2O.)

    I would say you are trying the “relationalist ploy” of taking the relational stuff in the topic-neutral reference fixer to be essential rather than merely reference fixing… and then complaining that a non-relationalist theory doesn’t explain the essentially relational facts. (Of course it doesn’t.)

    But suppose I am right about what theory I hold, never minding whether it is a ploy; you think it is still a losing tactic. Yet as far as I can tell, your various complaints all involve switching back to the other readings (what I was calling “semantic/epistemic” readings) of “of” or “as of.” On the a posteriori identity theory, sensations do not have any essential role in the truth conditions for beliefs, etc. It’s just not a theory of that stuff.

    Even if everything I say is wrong, you’ve got a long road ahead of you. As Dave points out, the strategy enumerative refutation means that you will always have to refute the next theory.

    On (2), I would say that the true but non-externality interpretation of the matching property is just the same as above: matching comes out to be a similarity claim about other states of the critter, rather than a claim about whether R “corresponds to the way things are” (p.6)

    On (3), the similarity of your arguments to others in the literature, it’s not very important. But if you want to convince someone who doesn’t already agree with you, I think you should actually be friendly to some neo-Kantian or McDowell-style arguments. Those arguments attempt to show that sensations have to have properties that are, roughly, externally-directed in the way that matters to you. If you don’t engage some such argument, then you’ve got nothing to say to someone who doesn’t already find your intuitions “sacrosanct” except to continue to attempt to refute all alternatives.

  28. Hi Tom, thanks very much. So it looks as if you are going for an actuality-based account of Externality after all – that clears things up. Now it also seems like you have decided to also give the same account of Matching, whereas in your initial response you and Doug said that biological theorists ought to say it’s false on all readings. Progress!

    The idea is that for any world W, N (hence, on your view, R) counts as an experience of a round thing at p (etc.) in W, because it’s similar to the state that people have in a *distinct* world W* (viz our actual world) when they see a round thing at p.

    So, on your account, if, in a world W, N is normally apt to be caused in Charlie (who is a normal member of a population that naturally evolved different wiring than us) by a square thing, and to cause square appropriate behavior, then even in W N counts as an experience of a round thing, and matches the world only if a round thing is present, because this state N is similar to the state that people have in a *distinct* world W* (viz our actual world) when they see a round thing at p. So, on your account of Externality and Matching, Charlie always has non-veridical experiences when he has N, because of the facts about N in our world! (Seems a bit unfair! :))

    If this were plausible, the biological theorist might accommodate the modal claims, Matching and Externality.

    But I think it is not plausible. In my note 29, I raised three problems for this kind of account of Externality and Matching: (i) a worry about apriority, (ii) a worry about arbitrariness (similar to one raised by Shoemaker in his comments regarding a similar strategy for handling BIVs), and (iii) a worry about belief. I didn’t see your responses to these worries, so let me reproduce them.

    (i) On the biological account, combined with an actuality-based account of Externality and Matching, it might actually be the case that we separated systems (as it might be, brains in vats), and hence R might actually fail to be an experience of a round thing, and an experience that matches the world only if a round thing is present. As against this, intuitively, however the actual world might turn out to be, R is such an experience.

    (ii) It arbitrary to suppose that ‘experience of’ facts and the matching-facts, concerning individuals in other worlds (like Charlie above) depend on facts about our world. Should they not depend on facts about their worlds? And, if they do depend on facts about some other world, why our world as opposed to any other?

    (iii) Presumably, if Charlie has an experience of a round thing, he can believe this. (Tom, you say that you are not giving a theory of belief – I agree – but this is still plausible I take it.) But, on the actuality-based account, it is hard to see how he might, since this would require that he has a belief whose truth-conditions involve our unique and highly specific world, which is quite remote from their worlds. How might he manage to have such a belief? (This is similar to a worry Soames raises about rigidified descriptivism about the semantics of names.)

    Any thoughts on these? To me, they seem pretty strong. You say ‘they involve switching back to the other readings’, but I do not see how this might be so. Could you please explain how there’s switching back in each of i, ii and iii?

    Now you conceded that your actuality-based account might have to be rejected because of the kinds of problems I raise. But then you say I would have to give an argument from ‘enumerative refutation’. I explained above that I have a systematic argument. Our most sophisticated and our only developed naturalistic theories of things like ‘experience of’ and matching appeal to input or output factors – roughly causation/indication and behavior. In separation cases (e. g. Charlie above, the simple system, etc.), N is entirely stripped from these: it is not apt to be caused by a round thing at p or to be caused by a round thing at p. So (setting aside actuality-based accounts bc of the above problems) to accommodate Matching and Externality the biological theorist would have to develop a radically unorthodox, reductive physicalist unqualified internalit account. No one (but me!) has discussed how to do this but that doesn’t mean it cannot be done. However, in my sect. ‘phenomenal intentionality as irreducible’ (linked to above) I offer a systematic argument that this cannot be done in a reductive physicalist way. I now realize I should put that into my paper – this is one of the many helpful things I have learned from the exchange here. But it is not quite right to suggest that I don’t have a systematic argument and that I am proceeding by enumerative refutation. I welcome thoughts on this systematic argument because I am very interested in this issue and like I said would like to put the argument into the biological theory paper.

  29. Just wanted to correct a typo: I was referring to sect. 7 of my ‘phenomenal intentionality as irreducible’ (linked to above). it’s in this short section that I provide a systematic argument against the kind of reductive unqualified internalism needed to save the biological theory (the rest of the paper is concerned with other views that wouldn’t help the biological theorist).

  30. Adam, this has been fun, but I fear that we are well down the path of diminishing returns. I’ll say something quick about (i)-(iii) and try to summarize where we stand.

    Your (i) involves matching the world in a sense that I deny sensations must essentially match the world, viz., being an experience “of” a round thing in the second (epistemic/semantic) sense that Doug and I disambiguated. The arbitrariness in (ii) only arises if we think that the matching is related to something like truth or application conditions, something epistemic/semantic. And as to (iii), as you acknowledge, I have denied that experiences are essentially such that they validate, justify, or fix beliefs. (I probably think they are not even necessarily/essentially such that they cause beliefs. Why should they be?)

    This should all be obvious, because it is clearly part of your view that experience is essentially epistemic/semantic. In your last comment you say, “Our most sophisticated and our only developed naturalistic theories of things like ‘experience of’ and matching appeal to input or output factors – roughly causation/indication and behavior.” To me this is evidence that you think experience is necessarily representational, epistemic, or semantic—such that it requires explanation in terms of causation/indication and behavior. So, as you would say, Progress! But the identity theory, as I have repeatedly said, denies that experience is necessarily representational, epistemic, or semantic. (At best, experiences have those features contingently.)

    But this is beside the point: Onlookers, if any, will get the idea. And of course we’re not about to convince one another. You challenged us to explain away your claims. Doug and I did so by (1) disambiguating similarity claims from epistemic/semantic claims; and (2) asserting that the identity theory can handle similarity claims, and we claims that is all that is required. In my view, you’re continually conflating the mere similarity view back into something epistemic/semantic. In you’re view, I’m continually refusing to see that similarity must be made sense of in epistemic/semantic terms (or proto epistemic/semantic terms.) Gridlock!

    As I understand it, you now want to break the impasse by appealing to arguments for irreducible phenomenal intentionality. Great, but that then argument will be bearing *all* the weight. If you can show that experience is essentially intentional, then of course every identity theorist will concede—for none of us think that our theories are theories of essentially intentional states. But your argument for the essential intentionality of experience [phenomenal] states (I presume) does not involve asserting that some “intuitions” are “sacrosanct.” I assume, rather, that you are arguing that we should accept a certain theory—for the theory of phenomenal intentionality, whatever its merits, is a theory (not an intuition, not self-evident) and highly controversial (not even close to sacrosanct.) But then you won’t need any of these appeals to intuitions or separation cases that we disagree about. Progress!

  31. Tom, thanks for your latest thoughts. All this has been very helpful to me.

    You say “it is clearly part of your view that experience is essentially epistemic/semantic” and “you are arguing that we should accept a certain theory [the phenomenal intentionality theory]”. It is in fact part of my end theory that, on the right readings (accounts) of Externality and Matching, they concern intentionality. But this is not part of my argument against the biological theory, as I said in the paper and in many of my posts. My argument is simply that Externality and Matching (which are not “theories” but pre-theoretically very plausible) are true – that is, have true readings. But, I argue, the biological theory is incompatible with Externality and Matching having any true readings. In the paper and in the posts, I discussed two (exhaustive) strategies available to the biological theorist for supplying Externality and Matching with true readings: actuality-based or not. But I argue that neither is successful.

    (1) The Actuality-Based Account
    On the actuality-based account you recently proposed, the interpretation of Externality and Matching that makes them true is not ‘semantic’. In a world W in which N fails to play the round-role, N (and hence on your view R) is ‘of’ a round thing, and matches the world only if a round thing is present, because in our world W* N is like the state that we have when we see a round thing. So on your account Externality and Matching are true, relative to a certain reading. You then agree with these widely-shared intuitions – the ones (as you often point out) I say are ‘sacrosanct’ (since the intuitions are neutral on the matter of interpretation and you think they have true interpretations). You just don’t think they are true on the usual ‘semantic’ theory of them.

    Now I will go through the problems I raise in the paper for the actuality-based account and the replies you gave in your last post, saying why in my opinion the replies fall short.

    (i) Suppose we all turn out to be or – to take a less extreme case – just somehow never see round things. We have N, and hence, in your view, R. (On your view, but not on others, this is a possible scenario.) Clearly, if this were the actual scenario R would still be an experience of a round thing, and match the world only if a round thing is present. (On the view Sydney Shoemaker leans towards, it would be an experience of a d-round thing – but as he explained this is enough to undermine the biological theory.) But on your account this is not so, because if this scenario is actual then N (that is, on your view, R) is not like the state we have when we see round things, for the simple reason that in that case we never see round things.

    Your reply was that this “involves matching the world in a sense that I deny” – the “(epistemic/semantic) sense”. But the objection is neutral on this. The objection is that in *some* sense, if we all turn out to be BIVs or somehow never see round things and have R, then R is still an experience of a round thing, and matches the world only if a round thing is present; but on your account this is not true in *any* sense.

    Moving from worlds considered as counterfactual to worlds considered as actual is a common way of showing that ‘actuality-based’ strategies are unsuccessful.

    (ii) My second objection was this. Suppose Charlie in world W has N, but he is a normal member of a population in which N has the function of indicating square things rather than round things. On the actuality-based account, there is a sense in which whenever he has N he is in a state that matches the world only if a round thing is present, because in some other world, W*, (viz. our actual world), N is like the state people have when the view round things. So on this account in a sense his visual states are always ‘non-veridical’ even when they occur in the present of what they have the biological function of indicating, etc. This looks very arbitrary – after all, there is another world W** in which N is like the state people have when they view trapezoids. Why then does not N in Charlie in a sense match the world when a trapezoid is present? The objection was just there is no reading of the intuitive ‘matching’ notion on which it arbitrary in this way. (Shoemaker raised a similar point in connection with the ‘counterfactual’ strategy for handling BIVs.) Again, you say that this objection requires reading ‘matching’ in a ‘semantic’ way, but it does not. It just says there is no reading of the intuitive ‘matching’ notion on which it arbitrary in this way, whereas on the actuality-based account there is.

    (iii) My third objection was that, if you are right and Charlie in W has an experience of a round thing whenever he confronts a square thing, it should be *possible* that he should believe he has such an experience; but it is hard to see how he might, since on your actuality-based account that would require that he have a belief whose truth-conditions involve the unique and highly specific distinct world W* (viz our world). You say that this objection requires that that “experiences are essentially such that they validate, justify, or fix beliefs” – which you deny. I am not sure why you say this – it just requires that if a believer has an experience of a round thing, it is *possible* that he should believe this. What could prevent him?

    So, pending other responses to (i)-(iii), your actuality-based strategy for finding true readings of Externality and Matching seems unsuccessful.

    (2) Reductive Unqualified Internalism
    On this account, by contrast to your actuality-based account, Externality and Matching have true ‘semantic’ readings. Return to Charlie. Charlie in world W has N, but he is a normal member of a population in which N has the function of indicating square things rather than round things, and is apt to cause behavior appropriate to square things. Once the actuality-based strategy is ruled out, the only option for the biological theorist who wants to say that Matching and Externality are in a sense true is to find facts about *Charlie’s W world* that could make it the case that N (on his view R) is ‘of’ a round thing and matches the world only if a round thing is present – even though in his world N is not apt to be caused by a round thing or apt to cause behavior appropriate to such a thing. That’s an interesting – even if very heterodox – theory of sensory intentionality. It is not an issue biological theorists have addressed. I give a systematic argument that the idea cannot be developed in a reductive way and so is not available to the biological theory.

    Now your reply here is that I’m “appealing to arguments for irreducible phenomenal intentionality”. But we must distinguish. In rejecting reductive unqualified internalism I am not appealing to an argument for any positive theory. All I need is an argument against a positive theory – a reductive physicalist form of unqualified internalism, which is a theory of intentionality no one has discussed and I argue cannot be developed.

    To sum up: the biological theorist, it seems to, cannot supply Matching and Externality with any true readings – whether ‘non-semantic’ (the actuality-based account) or ‘semantic’ (reductive unqualified internalism). Since it is extremely pretheoretically plausible that Matching and Externality have some true readings (as witness the fact that they are almost universally accepted), this provides a very strong reason to reject the biological theory (I think) – one distinct from the usual considerations about multiple realizability. I think, though, that there’s been progress in the sense that I now better understand your position. I have found this exchange extremely helpful – it will lead to lots of changes in the paper.

  32. Just wanted to correct a typo. Next to (i) I should have said: Suppose we all turn out to be *BIVs* or – to take a less extreme case – just somehow never see round things.

  33. Hi Adam, the kind of biological theory I hold isn’t supposed to be (2) is it? On my view the primary intension of phenomenal concepts is given roughly by the truckload of folk-psychological platitudes…so pain is the thing which is typically painful, is caused by this and that, in turn causes, etc, etc. No one disputes that here this picks out some kind of brain state. As a biological theorist I think that this is secondarily necessary but primarily contingent. This gives us a reading of externality and matching that makes them true.

    You object,

    (i) I introduce ‘R’ as a rigid designator of the salient experience property Adam has in a certain case C. For the sake of argument, let me just grant that “it’s illegitimate to stipulate a rigid primary intension for phenomenal concepts when it’s controversial whether there are any phenomenal concepts that behave like this” (even though we are both believers in phenomenal concepts that behave like this). Even so, as I said in response to Richard, it look as if the primary intension will not be a rich functional intension; it will simply be a meager, case-based primary intension, glossed by the description ‘the experience property Adam has in C’. (Why should it be that the only way we have of thinking of such a property is via a rich functional primary intension?) Now, if the biological theorist adopts the Kripkean proposition confusion strategy, then he says that, when I seem to be intuiting that an externally-directed claim is necessary, I am really only recognizing that the primary proposition is necessary. But if the primary intension is a meager, case-based primary intension, then this attempt at explaining away the intuition of necessity does not work. Because in that case the primary proposition is something like: that the experience property that Adam has in C has externally-directed property P. And that’s certainly not necessary.

    But this doesn’t work. The externality and matching intuitions are supposed to be a priori and so depend on a necessary primary intension. For instance here the necessity comes from interpreting ‘experience-type’ as including only those that have externality and matching conditions. Matching no longer seems necessary when one has the above non-necessary intension in mind. N is the secondary intension and so at worlds evaluated as counter-factual that have N we say that there is R there as well. so in possible worlds where H2O is not watery considered as counter-factual are worlds that have water that isn’t watery. So to in separation cases you have a seeing of a round thing that doesn’t have matching and externality but so what? That’s what happens when we evaluate these weird worlds counter-factually (Kripke talks about these kinds of cases in N&N when he discusses worlds where there are creatures who see like we do but do so by sening heat (rather than photons) this is a world where there is light even though they do not pick it out in the way that we do (and light is not heat in this world)).

    (iii) In general, the Kripkean proposition confusion strategy seem implausible to me. In fact, it is in tension with famous remarks Kripke made elsewhere in *Naming and Necessity*: “Don’t ask: how can I identify this table in another possible world, except by its properties? I have the table in my hands, I can point to it, and when I ask whether it might have been in another room, I am talking, by definition, about *it*.” (52-3). Similarly, I think can ask of R: must *it* be an experience of a round thing, etc.? I think it must be, and I just find it implausible that I am confused about what I am considering.

    I agree with the sentiment here expressed by Kripke (that is just the thing that gives us the weird results when considering worlds as counter-factual so there is no tension). When you ask of *it* if it must be an experience of a round thing you ask ‘must anything which looks to someone as this does to me be an experience of a round thing?’ and the answer is obviously yes. Are you confused about what you are considering? No; you are considering the way things appear to you. But none of that is a problem for the identity theoriest.

  34. Here’s my last comment!

    I’d like to thank Sydney, Tom, Doug – and Dave and Richard too – for all of the extremely helpful comments. They have helped me a lot. I also thank Richard for organizing this great conference. I have enjoyed it and learned a lot from it.

    Let me also address Richard’s most recent points – sorry for taking awhile. Richard wondered: how can I say that the primary intensions are contingent when I want to say the relevant claims are apriori as well as necessary? That is because I was assuming for the sake of argument that Dave was right that “it’s illegitimate to stipulate a rigid primary intension for phenomenal concepts when it’s controversial whether there are any phenomenal concepts”. Of course, if I accepted 2dism (I find the issue too hard to have a view), I’d say that the primary as well as secondary intensions are necessary.

    My main reply to the prop confusion objection is just an elaboration of what Dave suggested – that with the epistemic/modal and 2D distinctions in full view, we still clearly have the modal intuition (not just the epistemic/contextual one).

    At the end of your comment, you seem to agree. You write “when you ask of *it* if it must be an experience of a round thing you ask ‘must anything which looks to someone as this does to me be an experience of a round thing?’ and the answer is obviously yes.” I guess I am not sure about the talk of an experience looking a way to one (objects look various ways but not experiences), but it seems like you just have in mind my R (which is what is comment between a certain class of phenomenally identical cases), and you are agreeing it *must be* ‘of’ a round thing. I am not sure, though, why “none of this would be a problem” for the identity theory that R=N, for I argue (with separation cases etc.) that the biological theorist must deny that N *must be* of a round thing. (They’re summarized in my comment 32.) So we get that R is not N.

    I also have a question for you about how HOT relates to some of this (I would have thought that the most natural way to develop HOT wouldn’t be a version of the biological theory) but we’ll have to continue this by email! – Thanks again

Comments are closed.