Qualitative Inaccuracy and Unconceived Alternatives* Daniel Stoljar, ANU Pereboom (2011) is an extremely rich investigation of some of the central questions in contemporary philosophy of mind.1 In the foreground are the usual suspects from the current scene, but Kant and Russell loom in the background, and footnotes elaborate connections to people as apparently remote from the normal run of things as Dilthey and Derrida. With so much covered one is inevitably forced to focus on some things, setting aside others. Here I will concentrate on two ideas contained in the first part of the book: Pereboom’s ‘qualitative inaccuracy hypothesis’ and his criticisms of ‘the epistemic strategy’, a strategy I have defended in a number of places (e.g. Stoljar 2006, 2009). 1. The Qualitative Inaccuracy Hypothesis There is an analogy between anti-materialist arguments that concentrate on conscious states like itches or pains, and anti-materialist arguments that concentrate on secondary qualities, such as colour or heat. Suppose for the purposes of illustration we concentrate on this very simple argument: Argument A. P1. It is conceivable that there is a world that is physically exactly as the actual world is but in which this lemon is not yellow. P2. If this is conceivable then it is possible. C. It is possible that there is a world physically exactly as the actual world is but in which this lemon is not yellow. On the assumption that this lemon is yellow, C is inconsistent with materialism on most understandings of that doctrine. Is Argument A persuasive? Here is one reason to think it is not. There are various legitimate things one might have in mind by ‘this lemon is yellow’. On one interpretation—which we may call (in deference to Chalmers 2006) the postlapsarian interpretation—what is intended is something like: ‘this lemon has the property which makes certain physical objects reliably cause a distinctive category of sensations in us.’ Then P1 asserts that it is conceivable that there is a world that is physically exactly like the actual world but in which this lemon does not have that (postlapsarian) property. But that is something materialists will deny: in the world at issue, they will say, this lemon does have that property. On a different interpretation—which we may call the prelapsarian interpretation—what is intended is something like: ‘this lemon has a qualitative property with a distinctive sensuous nature, a property apparently distinct from any physical or functional property the lemon may have’. Then C follows, but materialists will now deny its relevance to materialism: in the actual world, they will say, this lemon does not have that (prelapsarian) property, so it doesn’t matter if one can imagine a physically identical world in which it does not. Of course, the materialist will add, in perception lemons routinely seem to have the prelapsarian property (just as they may also routinely seem to have the postlapsarian property—there is no problem with them seeming to have both properties). But, in at least this respect, perception is qualitatively inaccurate. If so, and if these interpretations exhaust the possibilities, Argument A poses no threat to materialism. Pereboom’s qualitative inaccuracy hypothesis, as I understand it, adapts this approach to the case of consciousness. Consider this equally simple conceivability argument:                                                                                                                 *

[Acknowledgements]

1

All references to Pereboom 2011 unless otherwise indicated.

 

1  

Argument B: P1. It is conceivable that there is a world that is physically exactly as the actual world is but in which I do not feel as if I have an itch in my toe P2. If this is conceivable it is possible. C. It is possible that there is a world physically exactly as the actual world is but in which I do not feel as if I have an itch in my toe. On the assumption that I do feel as if I have an itch in my toe, C is inconsistent with materialism. Is Argument B persuasive? Pereboom (if I understand him2) says not. There are various legitimate things one might have in mind by ‘I feel as if I have an itch in my toe’. On the postlapsarian interpretation what is intended is something like: ‘I am in a state which disposes me to introspectively represent that I feel a certain way, and disposes me to act as if I feel that way’. But materialists will deny P1 so understood: in the imagined world, they will say, I am in this (postlapsarian) state. On the prelapsarian interpretation, by contrast, what is intended is something like: ‘I am in a simple qualitative state with a distinctive introspective nature, a state apparently distinct from any physical or functional state I may be in’. But materialists will deny that C threatens materialism so understood: in the actual world, they will say, I lack this (prelapsarian) state, and so it doesn’t matter that I can imagine a world in which I am not in it. It may of course be that in introspection I routinely represent myself as being in this state (just as it may also be that in introspection I routinely represent myself as being in the other state—there is no problem with its seeming to me that I am in both states). But in this respect introspection is qualitatively inaccurate. If so, and if these interpretations exhaust the options, Argument B is unpersuasive. 2. Assessment of the Qualitative Inaccuracy Hypothesis How plausible is this response to argument B (and by extension the other anti-materialist arguments)? I will raise three objections. First, it is not clear that the interpretations on offer exhaust the options. Above we said: ‘I am in a simple qualitative state with a distinctive introspective nature, a state apparently distinct from any physical or functional state I may be in’. But this is to put it in code.3 What does the code mean? I see two possibilities. It might mean that I am in a state of feeling as if I have an itch in my foot—after all the state of feeling this way does fit the description given. But then Pereboom’s proposal is just eliminativism of the most straightforward kind (people not itchy, etc) except that it is put in code. But eliminativism in code is no more plausible than eliminativism not in code (in fact it is worse, since it is in code). Alternatively, it might mean something richer, e.g., that I am in state with no parts at all (that is what ‘simple’ means) or perhaps (even more strongly) that I am in a state (a) which has no parts and (b) whose essence is completely revealed to me in introspection. Then the eliminativist element in Pereboom’s account is both less straightforward more defensible; to deny that I am in a state that has these remarkable properties is not to deny that I am itchy. But the problem is that now his response to Argument B won’t work. It is crucial to that response that the pre- and postlapsarian interpretations exhaust the field. But on the suggestion we are considering they do not. In particular, friends of argument B will insist that what they have in mind is neither the pre- nor the postlapsarian interpretation, but simply the                                                                                                                 2

Actually, I take Pereboom to be expressing sympathy with this view rather than endorsing it outright, but I will ignore this complication in what follows. 3

To be fair, the code is mine not Pereboom’s. He himself tends to use the language of ‘modes of presentation’— though of course that too is to put matters in code.

 

2  

claim I feel as if I have an itch in my foot, i.e. the claim (they will say) on its usual interpretation.4 My second objection concerns the picture of introspection that is being presupposed in this account. Pereboom clearly holds an inner sense view of some sort; which sort? The standard modern taxonomy here is Shoemaker’s who distinguishes the object perception model from the broad perception model (see Shoemaker 1996). At one point (p. 19, fn. 24) Pereboom expresses allegiance to the broad perceptual model. But it is central to that account that introspective representations are just beliefs, and it is unclear that beliefs will do for Pereboom’s purposes. For what it is like to be believe falsely that one is itchy is not what it is like to be itchy5, so it is hard to say that it is an open possibility—something that might be true “given what we humans now rationally believe” (p.14)—that we are not itchy but merely believe that we are. At other points Pereboom seems to be committed to the object perceptual model, for example, when he talks of ‘introspective experiences’ (p. 23). But such accounts are in serious trouble, as Shoemaker points out. I may bear a perception-like experiential relation to itches where these are understood as something that seems to be in my toe, but the suggestion that I bear such a relation to the fact that I feel as if I have an itch in my toe is very counterintuitive. Hence it is unclear to me that there is any account of introspection that suits Pereboom’s purposes. He needs an account of what ‘introspective representations’ are which leaves it an open possibility that we have false introspective representations of phenomenal states. But if introspective representations are beliefs this is implausible, and if they are a kind of perception, they do not exist. Third, Pereboom has not made the hypothesis of qualitative inaccuracy plausible. To be successfully applied to Argument B, this hypothesis must not be merely that introspective error happens occasionally but that it routinely or systematically happens. Do we have any reason to believe this? Pereboom (p. 22) certainly makes a plausible case that error can happen, but making errors sometimes is quite different from doing it systematically. Indeed, on the matter of systematic error the colour case is quite different from the consciousness case. In the colour case, it is plausible that we make routine errors because of the interaction of two elements: (a) the rich way that objects such as lemons intuitively seem to be when we see them; and (b) the relatively poor way that we know such objects need to be in order to provoke that sort of state in us. In the case of consciousness, however, we have no such counterpart argument, and in particular we are missing a counterpart to (b). How exactly does my mind or my brain need to be in order that I representing myself truly as feeling itchy? I don’t think anybody knows the answer to this question. Hence I don’t think we know that the way that I am (or would be) when I truly represent myself as being itchy is poorer than the way that I represent myself to be. 3. Pereboom’s criticism of the Epistemic Strategy. Turning now to the epistemic strategy, this approach has two parts. The first part makes a hypothesis about our current epistemic position, viz., that we are (currently) ignorant of a certain kind of relevant fact. This part of the account is a contingent hypothesis; we obviously can’t prove that it is true from the armchair, nor can step out of ourselves and say what fact is                                                                                                                 4  A precisely symmetrical point may be made about the postlapsarian interpretation. The coded statement ‘I am in a state which disposes…’ might mean that I am in a state of feeling as if I have an itch in my foot or it might mean that I am in state that meets a certain very straightforward causal description. If the first, Pereboom’s response involves asserting that materialism about itches is right; if the second, his response is open to the charge that he has not exhausted the options.

 

5

For one thing, to get someone to falsely believe that they are itchy would require considerable set up, akin to the hazing example discussed by Pereboom on p. 22, but to get someone to be itchy is quite easy.

 

3  

it we are ignorant of. The case for this part of the epistemic view—the ignorance hypothesis, I call it—depends largely on reflection on our general situation and the consideration of historical cases. The second part of the epistemic strategy suggests that if the ignorance hypothesis is true, Argument B (to continue with this simple example) is unpersuasive. A good way to bring this out is as follows. There is an implicit quantifier in both P1 and P2 of that argument—‘exactly as’ means ‘in every respect’. If the ignorance hypothesis is true, this quantifier may be interpreted so as the facts of which we are ignorant are included in its domain or not. Let them be included; then P1 is implausible as a claim about conceivability, at least in the relevant sense. Let them not be included; then P1 is plausible but the argument will not threaten materialism because all that will now be established is that the fact that I have a certain feeling comes apart from some of the physical facts, not from all. Pereboom makes two objections against the epistemic strategy. The first is that it is no more plausible than a certain objectionable form of scepticism. “It is generally agreed”, he says, “that quantum mechanics (QM) is well supported by the relevant evidence (EV)” (p. 73). But now imagine a skeptic who says that EV provides no reason at all to believe QM. Why so? Pereboom imagines the skeptic pointing to the fallibility of human beings, and their history of wrong turns and mistakes in the past—in other words, just as proponents of the epistemic strategy do. As Pereboom notes, this sceptical position is in some ways analogous to the sceptical realist position on the problem of evil.6 A perhaps closer analogy is the position taken by Kyle Stanford in the course of discussing the problem of unconceived alternatives in philosophy of science (see Stanford 2006). Stanford would not argue, like Pereboom’s skeptic, that there is no reason to believe QM. He would rather argue that it is probable that QM is false because it probably will be superseded by a currently unconceived alternative, just as previous physical theories were superseded in that same way. Obviously Stanford’s position is controversial. My own view is that Stanford’s claim about our contemporary epistemic situation is plausible given the historical data he provides, but the further conclusion he draws concerning scientific realism is much less so.7 Whatever is true with Stanford, however, there is no doubt that the position of Pereboom’s skeptic is quite implausible. Maybe it is true that QM is likely to be false because there is an unconceived (or as Pereboom says ‘unspecified’) alternative to it, but it does not follow that EV provides no reason at all to believe QM. A body of evidence may give us reason to believe something, even if we also have more reason to believe something else, and even if that very body of evidence gives us more reason to believe something else. So much then for Pereboom’s skeptic; what has this to do with the epistemic strategy? Pereboom says that the proponent of the epistemic strategy holds a position analogous to his imagined skeptic; hence, since the skeptic’s position is implausible, so too is the epistemic strategy. That is his first objection. His second objection concerns a way to respond to the first objection. To defend himself, Pereboom says, “the QM-skeptic would require a partially filled-out hypothesis with significant probability—one, let’s say, that could really be true. By analogy …[the epistemic strategy]…would benefit from a partially specified hypothesis about the non-experiential that could really be true” (p. 74). Pereboom goes on to say that best way of providing this partially filled-out hypothesis is to adopt his qualitative inaccuracy view. In                                                                                                                 6

As Pereboom notes, Bennett 2009 develops this aspect of the issue. I discuss why I do not find the analogy to the problem of evil convincing in Stoljar 2009. 7

 

For discussion of these matters see Devitt 2010 (ch.4) and Godfrey-Smith 2008.

4  

sum, when it is adjusted so as it is plausible, the epistemic strategy is no alternative to his own. 4. Assessment of Pereboom’s criticism of the Epistemic Strategy How plausible are these criticisms of the epistemic strategy? In response to the first objection, while there are similarities between Pereboom’s skeptic and the proponent of the epistemic strategy, there is also a major disanalogy. In effect, the proponent of the epistemic strategy is objecting that what is conceivable according to argument B is not in fact conceivable, or anyway is not conceivable in the relevant sense; the underlying idea being that what is conceivable or not is a function of our epistemic position. If we think of conceivability as evidence for modal claims, the epistemic strategy is that we do not have the evidence that we think we do. That is quite different from the skeptic who agrees we have a particular bit of evidence (EV) but insists that it provides no reason at all to believe QM. Saying we do not have some evidence is one thing, saying the evidence we have does not support some hypothesis is quite another. Could one adjust matters to bring these positions more closely into alignment? The only option I see here is to imagine a version of Pereboom’s skeptic who says, not that EV provides no support for QM but that we do not have EV at all, but instead have a distinct body of evidence EV* which does not provide support for QM. However, the problem with this position is that the existence of an unspecified alternative (no matter how it is made plausible) lends it no support. If EV is observational evidence, for example, we have it whether there is an unspecified alternative or not. This is quite different from the conceivability case. In short, there is no relevant analogy between the kind of sceptic Pereboom describes and the proponent of the epistemic view. In response to the second objection, the insistence that there must be a partially filledout hypothesis is a very common reaction to the epistemic view but it misunderstands that view. The epistemic response makes the hypothesis that we are ignorant of relevant facts, and then goes on to say that this hypothesis bears directly on the plausibility of the Argument B. To object to that, you either need to deny the ignorance hypothesis or dispute that it bears on the argument (or both). It is not an objection to demand a partially filled-out hypothesis about the facts of which we are by assumption ignorant. Obviously the epistemic strategy cannot provide that; that is the whole point. Of course, it remains the case that the ignorance hypothesis needs to be defended, and as we have noted one obvious way to do this is to appeal to historical analogies. Pereboom notes that in my own discussion analogies to Broad’s emergentism about chemistry and Descartes’s discussion of language in the Discourse loom large, but he rejects both on the ground they do not involve phenomenal consciousness. I think this is mistaken. In the first place, Pereboom himself appeals to an analogy that does not involve phenomenal consciousness, the secondary quality analogy. I think there are problems with this analogy as I have said, but it cannot be rejected simply because colours (etc.) are not states of consciousness. In the second place, even if they do not concern phenomenal consciousness, the Broad and the Descartes cases do present predicaments that are deeply analogous to our own. In particular, in both cases we find two elements: a conviction on the part of the relevant theorists they have a complete grasp of the relevant facts (if not in detail then in outline), and a conviction that modal arguments of various kinds are persuasive. In both cases it was the conviction of completeness that turned out to be false. The epistemic strategy says that something similar is true in our own case—it doesn’t undermine the relevance of these cases that they do not concern phenomenal consciousness. References

 

5  

Bennett, 2009 ‘What you don’t know can hurt you’, Philosophy and Phenomenological Research 79 (2009) pp. 766-74 Chalmers 2006 ‘Perception and the Fall From Eden’ in Perceptual Experience ed. Tamar Szabo Gendler and John Hawthorne (Oxford: Oxford University Press), pp. 49-125 Devitt, M 2010 Putting Metaphysics First: Essays on Metaphysics and Epistemology (Oxford: Oxford University Press) Godfrey-Smith, P. 2008 "Recurrent Transient Underdetermination and the Glass Half Full," Philosophical Studies 137 (2008): 141-148. Pereboom, D 2011 Consciousness and the Prospects of Physicalism (New York: Oxford University Press) Shoemaker, S. 1996 The First Person Perspective and Other Essays (Cambridge: Cambridge University Press) Stanford, K 2006 Exceeding Our Grasp: Science, History and the Problem of Unconceived Alternatives (New York: Oxford University Press). Stoljar, D. 2006 Ignorance and Imagination (New York: Oxford University Press) Stoljar 2009 ‘Response to Alter and Bennett’ Philosophy and Phenomenological Research 79 (2009) pp. 775-784

 

6