The Knowing and the Known




Quantum mechanic.

I turn now to the third key way in which K* falls short of P1.

Recall that P1 and K* are as follows:
P1: If I know that I am not a zombie, then phenomenal character is (a certain kind of) conceptualized egocentric content.
K*: Smith knows that Smith has qualia –> (Smith believes that Smith has qualia & Smith has qualia))

When I first introduced this key way, I described it in terms of requiring an identity of the relevant representations and their being true. Now that I’ve made the relevant remarks about the conceptual and egocentric requirements on phenomenal knowledge, I am in a position to formulate the third way in which K* falls short of P1, namely that it leaves out the identification of phenomenal character with certain kinds of conceptualized egocentric contents.

If we modified K* to accommodate only the points made so far concerning the first two of the three ways in which K* falls short of P1, we would have only spelled out as a requirement on phenomenal knowledge that there is an isomorphism between the relevant representations and what it is that they represent. However, the remarks so far do not suffice to show that phenomenal character must be identical to certain conceptualized egocentric contents. Consider an analogous point made about George and the facts he’s able to know about rocks. There must be an isomorphism between facts George is able to know about rocks and conceptual contents that George is able to have. But this alone doesn’t suffice to establish that rocks are made of concepts or conceptualized contents. Concepts and their contents are mental. Rocks are extra-mental.

I want to argue that, unlike rock facts, which are not reducible to any set of conceptual contents, phenomenal character is so reducible. There are four general lines of thought in favor of viewing phenomenal character and rocks as disanalogous with respect to the question of reduction to conceptual content. The first is that we have reason to believe that at least some phenomenal character is constituted by conceptual contents, while we have no such analogous reasons for believing in the conceptual constitution of rocks. The second is that attributing concept-independence to rocks is needed to explain and organize our conceptualizations about rocks in a way that attributing concept-independence to phenomenal character is not. The third is that the very concept of a rock is a concept of a thing that has a reality that outstrips its appearance, whereas the concept of phenomenal character is an appearance concept. The fourth is that with respect to rocks, we have mere knowledge but not certainty. The certainty that we have with respect to qualia could not be achieved if the representations were distinct from what they represent.

Regarding the first point, there are relatively clear cases in which the acquisition and then subsequent application of a concept to experiences contributes to what it’s like to have that experience. One relatively familiar case concerns the way wine may taste quite differently to a novice and to an expert in virtue of the differential between the concepts each can bring to bear on their respective tasting experiences. Such cases lend at least prima facie support to the claim that at least some phenomenal character is conceptually constituted. An analogous case cannot be made for rocks: there’s no prima facie reason for believing that rocks are conceptually constituted.

Of course, one might pursue a plausible case that the pet rocks of the short-lived 1970’s fad were partially conceptually constituted on the grounds that no rock may be anyone’s pet without the rock being conceived of as a pet. So rocks qua pets may admit of partial conceptual constitution. But I hope to be granted the scientific-realistic presumption that rocks, qua geological kind, are in no way conceptually constituted.

petrock
Regarding the second point, the reason we believe that rocks exist independently of our rock judgments is that it helps to explain certain patterns in our judgments. The problem of how we know the external world is genuinely external is a huge problem and I certainly do not pretend to have a solution to it. However, I’m aware of no good reason for being a realist about rocks that doesn’t appeal to the explanatory or other theoretical utility of positing them as existing independent of our conceptualizations (see Mandik and Weisberg 2008 esp pp. 225-227). In contrast, there is no explanatory burden that phenomenal character bears that cannot be borne by a reduction of phenomenal character to a certain kind of conceptual content. This sort of point will be developed further in future posts.

Regarding the third point, consider the following. If phenomenal character outstrips conceptual content, then phenomenal consciousness would be noumenal for the person who has them. But if anything should be phenomenal as opposed to noumenal, it should be phenomenal consciousness. Phenomenal properties are appearance properties. Perhaps, one might object, there are two senses of appearance, an epistemic sense and a phenomenal sense. I address this suggestion later in a future post wherein I discuss the question in terms of whether the way e.g. colors appear to us in experience outstrips our ability to conceptualize them.

Regarding the fourth point, there is a degree of certainty that attaches to our knowledge of our own phenomenal states that does not attach to our knowledge of rocks. If, however, our representations of phenomenal character were non-identical to what they are representations of, then it would be possible to have the representations without their being true. But as long as that is a possibility, then I cannot be certain that I have accurate phenomenal representations.

Previous Posts:
1. Introducing Transcending Zombies
2. Anti-Skeptical Maneuvers
3. I Know I’m Not a Zombie
4. Some Remarks on Phenomenal Knowledge
5. The Egocentricity of Phenomenal Knowledge

10 Responses to “The Knowing and the Known”

  1. Eric Thomson says:

    This seems to be the lynchpin upon which your entire argument rides. If you could establish that phenomenal content is conceptual content, then that would be seriously bad ass. I’m not yet convinced though let me talk about each reason.

    Incidentally this one topic deserves an entire book.

    The first is that we have reason to believe that at least some phenomenal character is constituted by conceptual contents, while we have no such analogous reasons for believing in the conceptual constitution of rocks.

    [T]here are relatively clear cases in which the acquisition and then subsequent application of a concept to experiences contributes to what it’s like to have that experience. One relatively familiar case concerns the way wine may taste quite differently to a novice and to an expert in virtue of the differential between the concepts each can bring to bear on their respective tasting experiences. Such cases lend at least prima facie support to the claim that at least some phenomenal character is conceptually constituted. An analogous case cannot be made for rocks: there’s no prima facie reason for believing that rocks are conceptually constituted.

    Wine taste discrimination seems just as likely to be a case of perceptual learning, so our concepts track the changes in perceptual acuity, not vice-versa.

    The second is that attributing concept-independence to rocks is needed to explain and organize our conceptualizations about rocks in a way that attributing concept-independence to phenomenal character is not.

    Regarding the second point, the reason we believe that rocks exist independently of our rock judgments is that it helps to explain certain patterns in our judgments. The problem of how we know the external world is genuinely external is a huge problem and I certainly do not pretend to have a solution to it. However, I’m aware of no good reason for being a realist about rocks that doesn’t appeal to the explanatory or other theoretical utility of positing them as existing independent of our conceptualizations (see Mandik and Weisberg 2008 esp pp. 225-227). In contrast, there is no explanatory burden that phenomenal character bears that cannot be borne by a reduction of phenomenal character to a certain kind of conceptual content. This sort of point will be developed further in future posts.

    This will be interesting to see. Intuitively, I think I can make finer perceptual discriminations than I can conceptual distinctions, which would suggest we need a different type of representational format to account for perception.

    The third is that the very concept of a rock is a concept of a thing that has a reality that outstrips its appearance, whereas the concept of phenomenal character is an appearance concept.

    Regarding the third point, consider the following. If phenomenal character outstrips conceptual content, then phenomenal consciousness would be noumenal for the person who has them. But if anything should be phenomenal as opposed to noumenal, it should be phenomenal consciousness. Phenomenal properties are appearance properties. Perhaps, one might object, there are two senses of appearance, an epistemic sense and a phenomenal sense. I address this suggestion later in a future post wherein I discuss the question in terms of whether the way e.g. colors appear to us in experience outstrips our ability to conceptualize them.

    This seems key. I guess if you define concepts the way Churchland does, so even our color space is a kind of conceptual space, then you might be right.

    I worry that this is going to end up with you saying something is conceptual, and others saying “Well I wouldn’t count that as conceptual.” And we’d have a kind of uninteresting semantic dispute. You might want to beware of defining ‘concept’ so promiscuously that any representational content is conceptual content. (Unless you don’t believe in nonconceptual contents…)

    The fourth is that with respect to rocks, we have mere knowledge but not certainty. The certainty that we have with respect to qualia could not be achieved if the representations were distinct from what they represent.

    Regarding the fourth point, there is a degree of certainty that attaches to our knowledge of our own phenomenal states that does not attach to our knowledge of rocks. If, however, our representations of phenomenal character were non-identical to what they are representations of, then it would be possible to have the representations without their being true. But as long as that is a possibility, then I cannot be certain that I have accurate phenomenal representations.

    A couple of points. Fist, your argument would suggest that conscious contents are representational contents, but not necessarily conceptual contents. It seems any representational content will get us the epistemic buffering you talk about.

    Also, I am not sure we have such certainty because of things like Blindness Denial (which, incidentally is actually controversial, as they may be hallucinating). I think we do have higher credence, so it is a matter of degree, but not quit at certainty.

  2. Pete Mandik says:

    Great comments, Eric. Thanks.
    A few further thoughts:

    1. Even if you’re right that the other story about wine tasting is equally probable, this still doesn’t defeat my point about there being a disanalogy between the perception case and rocks. There aren’t two equally plausible theories about rocks, one which says our concepts track rock changes and another that says rocks track conceptual changes.

    2. I definitely want to avoid a trivializing argument whereby any old representation is a concept. In my latest post, I say some pretty specific things about what I take concepts to be. By the way, I’ve talked to Paul Churchland about this sort of point and, in conversation at least, he concedes that it is perhaps best, in dealing with zombie and Mary arguments, to have more stringent criteria for concepts than can be satisfied by a trained up feedforward network.

    3. If you think blindness denial and other weird stuff undermines first personal phenomenal certainty, then you’re not really in the audience for Transcending Zombies argument (and should be referred instead to the Josh Weisberg’s and my Type-Q Materialism argument which shows how to get to materialism via a lack of first-personal certainty).

  3. Josh Weisberg says:

    Pete–

    You write:

    >If, however, our representations of phenomenal character were non-identical to what they are representations of, then it would be possible to have the representations without their being true.

    So, if our representations of phenomenal character are identical to what they represent, then it is not possible to have the representations without their being true (by contraposition).

    Is this right? Viz: This sentence has 14 words. It is identical to what it represents, but it is false. So, not all self-representing sentences (mental states, etc.) are true. What ensures that connection for phenomenal knowledge?

    Also, might there be other ways of achieving certainty about our phenomenal states? One is a Shoemaker/Lewis trick: phenomenal states are defined as states accompanied by distinct belief states that accurately represent the phenomenal state. If a state is not so accompanied, it is not a phenomenal state. Now, one can ask about “pseudo-phenomenal” states–ones not accompanied by accurate beliefs (is there anything it’s like to have them? etc.), but it seems that certainty for actual phenomenal states is guaranteed.

    Or, there is a special relation of acquaintance holding between phenomenal states and beliefs about phenomenal states. Or “phenomenal concepts” are in part constituted by phenomenal character, though that character is distinct and can exist independently of the beliefs. So there is no space for error in phenomenal knowledge.

    Anywho, it seems that there might be other ways to achieve certainty besides your conceptual-identity route. Maybe you’re going to kill these alternatives, or maybe they violate something the qualophiles already accept? (Dave Beisecker brings in “Uncle Wilfrid” here, but that doesn’t bother folks as much as it should…)

  4. Pete Mandik says:

    Heya Josh,

    When we’re talking about something with sentential content, there’s an ambiguity in “what does S represent?” that I need to be more explicit about. On one reading, an acceptable answer is simply the grammatical subject of S, on another reading, the answer must be a nominalization of S. Compare, then, answering “What does ‘Smith is running’ represent?” with “Smith” vs. “Smith’s running”. The latter reading is what I have in mind when I make my identity claim (though I should spell this out more in the future). So “This sentence has 14 words” is not identical to what it represents, since what it represents is that sentence’s having 14 words and it is not a 14-word sentence.

    Regarding the other ways of achieving certainty, I should maybe say more about what I don’t like about them, but it boils down to there seeming to be no reason for positing a separate existence for the states of affairs that phenomenal beliefs depict.

    BTW, what are the citations on the relevant Shoemaker and Lewis? I’d like to check ‘em out.

  5. Josh Weisberg says:

    Yo.

    I’m not sure I get you’re answer to the self-rep thing. I was suggesting that some kinds of self-rep are open to misrepresentation–what does the properly nominalized reading of an egocentric phenomenal representation look like, such that it avoids that possibility? (This really is a clarificatory question!) I take it a model for your claim is something like the cogito?

    I think you probably need to point out the problems with the other ways. If qualophiles can defend one of them, can’t they avoid your argument? Here, again, Uncle Wilfrid lurks, though oddly qualophiles seem to forget what our dear uncle taught us!

  6. Pete Mandik says:

    Hi Josh,

    Here’s one way to model what I’m after.

    Consider sentence S:

    (S): This sentence is, in part, about there being a ten-toed chicken on Mars.

    Now, it may be false that there are any Martian chickens, but that doesn’t affect whether S is true.

    An analogous belief would be B:

    (B): This belief is, in part, about there being a ten-toed chicken on Mars.

    What makes B true isn’t some state of affairs that can obtain independently of B. What makes B true is B itself. B is identical to what it represents.

    —-

    Another way to get at what I’m after goes like this.

    I identify qualia with certain doxastic seemings, so phenomenal character, what it’s like to be me, just is what I believe things to be (restricted to a certain class of beliefs). (The restriction is to single out just the beliefs that are state-conscious).

    What phenomenal character is is just what these beliefs are about (in the nominalized-sentence sense of ‘about’). Believing P suffices to make it true that what it’s like for me is that P is true. It doesn’t matter whether P is true. Phenomenal character just is what my belief represents.

  7. The operating characteristics of a competent semantic network might help resolve the problem of self representation in sentential propositions. For example, in a computer simulation test of my neuronal model of a semantic network, the network learned to define a tiger in terms of several predicates; e.g, “a tiger is a cat”, “a tiger has stripes”, etc. When the network was asked “What is a tiger?”, I was surprised that its first response was the identity relationship “A tiger is a tiger” (see *The Cognitive Brain*, Ch. 6 “Building a Semantic Network”, pp. 108-109). So it seems, if my theoretical model of the cognitive brain is valid, that the “truth” of self-identity is an innate property of the semantic mechanisms in the human brain.

  8. Josh Weisberg says:

    Pete–

    Hmmm…. very interesting.

    Questions:

    I thought the thing we’re supposed to be certain about is quite specific phenomenal content–that is, that things seem green, or seem red, or what have you.

    But what assures knowledge of that difference? It seems that all these ‘about’ sentences are true in the same way–they can’t be false because they refer back to themselves. But that doesnt get you to knowledge of specific character. So, my certain knowledge of “This sentence is about blorbs” is the same as my certain knowledge of “This sentence is about shlorbs”. But given that I don’t know what blorbs and shlorbs are, I can’t be said to know the difference between blorbs and shlorbs. Now, I can (I suppose) tell the syntactic difference here. But that seems quite different from what I know when I know I’m having a red, as opposed to a green phenomenal state–I know things about the similarities and differences in my phenomenal experience–that it’s more like purple than blue, say. Don’t you need that level of specificity? Is the syntactic deal enough? (I’ll look for the Lewis and Shoemaker reference…)

  9. It seems clearly the case that occurrent phenomenal content is beyond any challenge of veracity, in contrast to beliefs about phenomenal content, which can be true or false. Example: Your phenomenal experience of the moon at the horizon is that it is larger than the moon seen at its zenith. This is beyond challenge. Your belief (?) that the moon’s projected image has actually changed in size from horizon to zenith can be challenged and is, in fact, false. But your occurrent phenomenal experience of this belief is beyond challenge.

  10. Pete Mandik says:

    Hi Josh,

    One problem with the “blorbs” and “shlorbs” case, is, not knowing anything at all about blorbs, I have no concept of blorbs and can’t have any belief about blorbs as such. Not just anything that can happen with a sentence has an analog in belief. Nonetheless, I think you raise a good point. Part of your question, I take it, is something like, if certain conceptual contents are what determines certainty, then what determines specificity?

    On my view what sets the upper and lower bounds of specificity goes something like this. Experience is going to be no more specific than the concepts brought to bear, so conceptual repertoire determines the limit of maximum specificity. But experience is not going to be as general as just any old conceptual mental state because not any old conceptual mental state is going to be able to get involved in the Allocentric-Egocentric Interface. For example, a content like “Historical events happen in the past” is not going to be driven by and drive egocentric representations in such a way that makes it a content of conscious experience.