Bursting Apart at the Seems

I assume that no interesting controversy exists over whether there is an epistemic sense of “seems”. I question whether there additionally exists a phenomenal sense of “seems”. I question whether there are phenomenal appearances in addition to epistemic appearances.

To get a handle on what the alleged distinction is supposed to be, it helps to consider the following picture. Smith and Jones are two little men in two little opaque boxes. Mediating between the interiors and exteriors of their boxes is a camera that feeds into a computer capable of reliably detecting the presence of dogs near the boxes. Smith’s dog-detecting device has a video readout that displays Smith the printed word “dog”. Jones’ dog-detecting device has a speaker that says “dog” aloud upon dog detection. Both Smith and Jones, via the use of their dog detectors, come to judge that a dog is just beyond the walls of their boxes. Smith and Jones are alike, then, with respect to the ways things epistemically seem to them: insofar as they judge that dogs are present, it epistemically seems to them that dogs are present. However, in spite of these similarities between Smith and Jones, there are notable differences. In particular, there are differences in the evidence they rely on as the bases of their dog-detecting judgments. Smith’s evidence concerns what’s happening on the video display whereas Jones evidence concerns what’s happening with the audio speaker.

Is the above picture a good model for distinct kinds of appearance? Suppose we do something to transform the differences between Smith’s and Jones’ box interiors into a difference between the insides of Smith’s and Jones’ heads. Suppose we kick them out of the boxes and have only their own senses mediate between them and dogs. Smith is deaf but sighted and Jones’ is blind but has his hearing intact. Smith and Jones come to both believe that there is a dog present, but one does so by seeing the dog and the other does so by hearing the distinctive bark. Are the differences that arise in spite of their similarity in judgment worth calling a distinct kind of appearance? Are the ways dogs appear to Smith and Jones epistemically identical but phenomenally distinct? I think not.

If there is indeed a distinction to be made sense of, we need to be able to make sense of two different kinds of cases: one in which phenomenal appearances remain constant while epistemic appearances change and one in which epistemic appearances remain constant while phenomenal appearances change. Neither of the versions of the story about Smith and Jones involve the requisite changes. Whatever changes would be required to change Smith into Jones would change various beliefs Jones had, beliefs like whether he was looking at a monitor versus listening to a speaker

In the first version of the story, Smith and Jones differ with respect to their evidence, e.g. screen vs. speaker. In the second story, there is also a difference with respect to evidence. But it is a big mistake to think that in either story the evidence is something in the heads of Smith and Jones. In the second story, the different evidence is the difference between light reflected and sound emitted. These different kind of events trigger in Smith and Jones certain beliefs which in turn give rise to inferences, the conclusions of which are the common belief that a dog is present.

Fig. 1. Smith and Jones.

Fig. 2. Dog, detected. (Photo by No Fixed Abode)

31 Responses to “Bursting Apart at the Seems”

  1. Richard Brown says:

    Pete, consider the Muller-lyer illusion. When you first see it it epistemically seems that the two lines are of differing length. Once you find out about the illusion you realize that the two lines are the same length and so it no longer epistemically seems that they are the same length, but there is no difference in the phenomenal aspect: it still phenomenally seems that the two lines are different lengths. Is this an example of the first thing that we would need to show in order to get a distinction like this going?

  2. Pete Mandik says:

    Hi Richard,

    I think illusions are very important to consider in connection with this stuff. A person who thinks all appearances are epistemic can attempt to assimilate the explanation of perceptual illusions to the explanation of cognitive illusions.

    Consider, for example, the Monty Hall Problem:

    Suppose you’re on a game show, and you’re given the choice of three doors: Behind one door is a car; behind the others, goats. You pick a door, say No. 1, and the host, who knows what’s behind the doors, opens another door, say No. 3, which has a goat. He then says to you, “Do you want to pick door No. 2?” Is it to your advantage to switch your choice?

    Someone can read about what the right answer is and why it is the right answer and still not shake the urge to give the wrong answer. It might seem to such a person both that one should pick door No. 1 and that one should pick door No. 2. Whatever seemings are involved here are clearly epistemic. Why, with regard to seemings, isn’t the Muller-Lyer illusion exactly the same sort of thing?

  3. Hi Pete,
    I guess I will repeat an example similar to Richard’s.
    Say, Smith sees a dog. Jones says to him, it is not a dog, it is a hologram, and Smith believes him (while still looking at the dog).
    Wouldn’t that be a case where phenomenal appearances remain constant, but epistemic appearances change? (for Smith)

  4. Pete Mandik says:

    Tanasije,

    I guess I’m not seeing what the argument is that what remains constant isn’t just another kind of epistemic appearance.

    Note that the anti-phenomenalist can fully grant that it seems like there’s a distinction between epistemic and phenomenal appearances. What the phenomenalist needs is an argument that there really is such a distinction and an account of in what that distinction consists.

  5. Pete,
    Let me rephrase the example in more general way:
    I take it for granted that two things can look the same (while being of different type).
    One can imagine lot of examples, for example there might be two classes of things (A and B) which have same front side, but different back side.
    Maybe I don’t understand your proposal, but it seems to me if one denies the distinction between phenomenal and epistemic appearances, one wouldn’t be able to account of possibility to say e.g. “A and B look the same (e.g. if looked from side)”. What would the identity consist in, if not in the phenomenal appearance?

  6. Richard Brown says:

    Pete, I am getting confused….you ask”Are the ways dogs appear to Smith and Jones epistemically identical but phenomenally distinct? I think not.” It then looks like you tell a story about information processing and belief formation that completely by-passes conscious experience (redundantly redundant to say that, but these days you have to…sad really….). Is this an attempt to avoid the ‘myth of the given’ (that is sense data)? If not what is going on here? For surely, one might think, I can come to form beliefs based solely on the way things appear to me. Like when I hear the barking dog: I can listen attentively to the barking to determine whether the dog is a big one, small one, vicious one, alarmed one, playeful one, scared one, etc, etc…I can even identify the breed this way. This cannot be explained solely in terms of information comming in and triggering beliefs, can it?

  7. Pete Mandik says:

    Tanasije,

    Hopefully the following helps.

    I am assuming that whatever epistemic seeming consists in, it can be explained in terms of certain kinds of mental events (e.g. states with propositional content). I assume also that if there is a distinct kind of seeming, so-called phenomenal seeming, then it can be explained only in terms of mental events distinct from the ones invoked to explain epistemic seeming (e.g. states with qualia).

    So, a hard-core anti-phenomenalist (about seemings) would explain what it means for something to seem a way for someone in terms of the various judgments that person is disposed to make about the thing in question.

    So, suppose A is a square-based pyramid and B is a cube. Looked at from the bottom, they both seem the same. What, in terms of judgments, can this consist in? One kind of answer the hard-core anti-phenomenalist (about seemings) can give is: George, looking at only the bottoms of A and B, judges them to both be cubes.

  8. Pete Mandik says:

    Richard,

    Let me know if the following helps.

    This is, among other things, an attempt to avoid the so-called given and sense-data. It is also a proposal that lots and lots of stuff can indeed be explained solely in terms of information coming in and triggering beleifs.

  9. Pete,
    I think I once already wrote this as a response on different post, so please excuse me if you find this repetition boring:
    If I show a paper with two circles, one scarab-green and the other spring-green to a person who doesn’t recognize those colors, he/she will still be able to notice they are different. (The case is true for me at least).
    Wouldn’t “different greens”, when one has just one concept for both colors seem like a contradiction if the person has just epistemic appearances?
    And those colors won’t be merely different, but one can build concepts of scarab-green and spring-green based on their phenomenal appearance. What else is there to base the concept on? (I admit I tend to look at the question phenomenologically)

  10. Pete Mandik says:

    Tanasije,

    Why wouldn’t the ability to apply the concepts CIRCLES, COLOR, and DIFFERENT suffice for two colored circles to seem different?

    If the answer is that in the imagined case, there is some particular way in which they seem different, say one seems darker than the other, then my question becomes:

    Why wouldn’t the ability to apply the concepts CIRCLES, COLOR, DIFFERENT, and DARKER suffice for two colored circles to seem different?

    If you answer that there is still some way in which they seem different, then I will add another concept to the list that captures the way in which they seem different. And if we keep up in this excercise, pretty soon either (1) we will hit a point at which it looks like the person had the concepts SCARAB GREEN and SPRING GREEN all along or (2) there is some way in which the colors seem different to the person that outstrips the concepts the person has.

    If there is such a thing as possiblity (2) - a way in which the colors seem different that the person has no concept with which to conceive of the difference - then how could they have any epistemic access to whether there’s a difference in the first place?

  11. Pete,

    I’m not sure how can one defend possibility (1), i.e. that the person had the concepts of SCARAB GREEN and SPRING GREEN all along. What would “already having such concepts” consist in? In some sense the person didn’t have those concepts until he/she learned them (the opposite would make learning impossible). Maybe we can say that the person had the disposition to learn those concepts from the very start, but can we equate the disposition to have (or acquire) a concept, with having the concept?

    I would take (2) as only thing left as possibility, and I would argue that not just that there is possibility of pre-conceptual noticing/seeing of the difference, but that it is a base of forming of concepts (through acquiring power to recognize this phenomenal appearance).
    I’m not sure why you find (2) as problematic. I guess I’m missing something in the wider picture.

  12. Pete Mandik says:

    Tanasije,

    I find 2 problematic because you would need a general argument that there can be no concept acquisiton unless there were non-epistemic seemings. But this is implausible when we look at cases that most people would agree have nothing at all to do with phenomenal seemings like, for example, concept acquisition in highly advanced mathematical areas. The other day I explained to some of my students what Dedekind cuts are. Assuming that I succeeded, they acquired a concept based on their ability to relate it to other concepts that they already had, such as the concepts of the number line and rational numbers. But there was nothing obviously necessarily involving phenomenal appearances there. Why, then, think they are involved anywhere? What’s the argument?

  13. Pete Mandik says:

    Tanasije,

    An additional consideration.

    I assume that things don’t epistemically seem any way at all to a thermometer, and that a thermomter has no concepts. Do things thereby phenomenally seem some way to thermometers when they register the temperature? If so, then I have no further argument beyond saying that’s really weird. If not, then the defender of phenomenal appearances needs to say what phenomenal appearances consist in beyond mere non-epistemic reactions to things.

  14. Pete,
    1.Why would one need to defend that all concepts are acquired from phenomenal appearence, in order to argue that some of them are? SCARAB GREEN might be acquired based on phenomenal appearence, concept of Dedkind cuts might be acquired in different way (through understanding their relation to different concepts)

    2.I would not reduce phenomenal appearences to some kind of sense-data which are caused in the person’s brain by the things in the world. So, I don’t think that things seem any way to thermometers (phenomenally or epistemically). (I’m not sure about that as (I never was a thermometer in my life, but I agree that it would be really weird.)

    So, this comes to the question of what those phenomenal appearances would consist in (then). But doesn’t that put me in a bad position, where I need to reduce to concepts that which I see as a base for the concepts themselves?

  15. Pete Mandik says:

    Tanasije,

    1. Short of a general argument, you then need to supply an argument why SCARAB GREEN can’t just be like DEDEKIND CUT. No argument has yet been supplied. In contrast, I’ve spelled out how it could be just like DEDEKIND CUT.

    2. You are in a tricky position but not an obviously impossible one. You could perhaps spell out what phenomenal appearances would be for a subject who didn’t himself have the associated concepts: what, for example is the phenomenal appearance of scarab green for a subject lacking the concept of scarab green. You could perhaps do that even if you yourself have the associated concepts without necessarily undermining your position.

  16. Pete,

    1.One can teach someone what SCARAB GREEN vs. SPRING GREEN is by pointing, and pronouncing the words. Where does the meanings of the words “scarab green” and “spring green” come from in that case (if we remove phenomenal appearance)? Also, can one teach SCARAB GREEN same way as DEDEKIND CUT, or other way around?

    2. Let me try… I would say that what is seen by that person is a (more or less) specific color. It is not merely a color, it is not merely a green color, and it is not even merely a scarab green color, because probably lot of shades can be scarab green. Having the concepts that person has, he/she can determine it as 1)A COLOR 2)A GREEN. So the person sees color which is A GREEN, but which is not merely A GREEN.

  17. Pete Mandik says:

    Tanasije,

    1. The meanings come from (a) informational linkages between neural states and the colors in the external world, linkages that are neither conscious nor phenomenal and are no different in kind than the way thermomemtor states carry information about temperatures and (b) linkages to previously held concepts like COLOR

    2. Why couldn’t the application of the concepts NOT, MERELY, and A GREEN suffice to account for the appearance to the person that they are seeing a green but not merely a green?

    Here’s another way of thinking about the issues raised in connection with number 2: How would things epistemically seem to you if you had phenomenal states? How would things seem to you, epistemically, if you didn’t? If phenomenal appearances are indeed distinct, would you be incapbable of (epistemicall) noticing the difference from the first-person point of view?

  18. Pete,

    1. If the meanings come from linkages that are not conscious, how come I’m conscious of those meanings? (I’m conscius/aware of the color that is the meaning of taught word - there it is, in front of me)

    2. NOT+MERELY+GREEN wouldn’t do it because it is possible that the person later recognizes the same shade of green (e.g. it was scarab green, so later he/she recognizes it vs. some other shade of green). It won’t be possible if it is just “not merely green” as all shades of green are not merely green.

    I’m inclined to think that things can’t appear epistemically to me, without them appearing phenomenally. (Haven’t thought it through though)

  19. Pete Mandik says:

    Tanasije,

    1. The short answer is: because you apply a bunch of concepts in the right way. In this case, concepts such as COLOR and THERE IN FRONT OF ME. The long answer is here: [link]

    2. So now it’s starting to look like she had the concept of scarab green after all. If some one is able to distinguish a color from all others not only in experience, but also in memory, why isn’t that sufficient for a conceptual grasp of that color? I am assuming that the mere fact she doesn’t know that the English word for the color is “Scarab green” doesn’t preclude her from having a concept of that specific color.

  20. Richard Brown says:

    No, it didn’t help….

    You say you are an anti-realist about phenomenalistic seemings, and yet you seem to think that that does not mean there are no qualia, so what gives?

  21. Pete Mandik says:

    Richard,

    What gives:

    Either qualia aren’t seemings or they are reducible to epistemic seemings (that is, identical to a sub-class of epistemic seemings).

  22. Pete,

    About 2)…
    I agree that having a word is not a requirement for having a concept.
    I’m still not sure what exactly are you proposing when you say that the concept is already there? Do you take position that all concepts are innate?
    Maybe we going in circles:
    I won’t agree that she had concept SCARAB GREEN before she had seen scarab green. (As I said, that would make learning scarab green impossible, or at least senseless.)
    Also I would argue that she didn’t have the concept even while she was seeing it. Because she didn’t have a reason to form a new concept of that color (vs. the GREEN she uses).
    It is only later when somebody puts two green color chips in front of her, that one of them reminds her of the specific color she had seen. She recognizes it, and that presents base for a concept. Some kind of special epistemic role of that particular color vs. the general green.

  23. Pete Mandik says:

    Tanasije,

    I agree with everything you just said. Note, however, that none of it proves that there are phenomenal appearances.

  24. Richard Brown says:

    Dude, that doesn’t give, that takes.

    Look, I personally find the suggestion that qualia are not seemings (or, better, that qualitative states are not in business of generating conscious experiences for the animal in which the occur) to be just crazy and I would need to have strong arguments/evidence to ever consider that this was not the case; I have just had too many acid trips that I consider to be extremely convincing evidence to the contrary (not to mention all of the ORDINARY experiences that I have). Arguing that a notion of epistemic seeming ‘in principle’ can give us an explanation of the animals behavior without appealing to phenomenal apperances is not good enough. There needs to be an argument that it is the best explanation. Now this does not mean that I endorse sense-data, but C’MON, that is not the only way to be a qualia realist!!!

    So, OK, let’s put that aside for the moment and consider your second suggestion: ‘qualia can be reduced to epistemic seeming’. So, I take it that this sense of epistemic seeming has to do with the beliefs that the animal has, right? And the idea is that incomming senory information results in states of the animal which trigger beliefs with the end result being that the animal comes to be in a state of it epistemically seeming that such and such, yes? Example: The dog barks, the soundwaves trigger state S of the auditory system, state S triggers belief B to the effect that there is a dog present and being in state B is what makes it the case that it seems to me that there is a dog here. Now, does ‘epistemically seems’ here equal ‘conscious of’? Is there ’something that it is like for the animal that has the seeming?’ It seems to me that it does, what else could it mean? Then this is to say that belief B makes me conscious of the dog. You suggest that this is an explanation of qualia, which is really our old friend ‘a conscious state is a state that makes me conscious of something’ . But is seems to me that you have shot yourself in the foot. For, just as I can have the belief about the dog or not, and thereby be consicous of the dog or not, so too I can have a belief about my belief that there is a dog or not and so therby be conscious of my belief or not. In other words I can have the belief that there is a dog near or whatever without myself being aware of that state. ‘Sure’ you say, ‘that’s just my argument against the transivity principle. We can have a conscious state without being aware of that state.’ But now what that amounts to is that I can have the belief, caused as it was by dog barking, that there is a dog present without being aware that I had that belief. You have described a zombie. In what sense is there anything that it is like for the animal that has this belief? The question is not ‘what is it like to have the belief?’ The question is ‘what is it like FOR THE ANIMAL to have the belief?’ If the animal is unaware that it has this belief, why think that there is anything that it is like for it to have the belief? If not then your account cannot exlpain qualitative consciousness. In short your account does not work with appeal to some kind of phenomenal seeming, over and above the epistemic seeming.

  25. Richard Brown says:

    ooppss that’s upposed to be ‘with out appeal’….

    We missed you at the NC/DC meeting last night :(

  26. Here is a sketch of a representationalist approach to these issues. For expository purposes, I will follow many functionalists and speak as if for each distinct attitudinal mode, there is a functionallly defined “box” for storing representations. Really, these boxes and even the representations in them may be purely virtual from the point of view of neural realization. But I find it a handy expository metaphor.

    So there will be a “visual experience box” which holds the contents of visual experiences, an “auditory experience box” to hold the contents of auditory experiences, and so on. Of course, there must also be a “belief box” to hold the contents of one’s beliefs, which may be different from how things seem in experience, as in the Muller-Lyer case after one learns it is illusory. The visual experience box contains representations that say that one line is longer than another, but the belief box contains representations that say they are the same length.

    Presumably, perceptual modules move from physical stimulation as input and have as upshot the depositing of representations in the various experience boxes. Acquiring certain concepts involves acquiring both inference rules that govern them in thought, but also (for some concepts), cultivating transducers that can write sentences containing these terms into the various experience boxes (for non-inferential application).

    On this picture, having it look (visually seem) to you as if there is a dog in front of you would be a matter of having a representation that means that there is a dog in front of you in your visual experience box. Having it sound (auditorily seem) to you as if there is a dog in front of you would be a matter of having a representation that means that there is a dog in front of you in your auditory experience box.

    In the usual case, the contents of one’s experiences are simply endorsed and so “copied” into one’s belief box more or less automatically. But one can hold them apart when it becomes salient to do so.

    I don’t know if this is purely epistemic, but it does seem to be enough to give what we need by way of “phenomenal appearances” as the contents of the experience boxes, without having to posit anything other than representations.

  27. Pete Mandik says:

    Richard and Anders,

    I found both of your comments sufficiently thought provoking that I just did a whole new post on this topic.

    Richard, sorry I missed practice but I was in the midst of total swampitude Sunday. But anyway, my recent post, I hope, spells out what I take the reduction to epistemic seemings to be while spelling out the connections to the correct theory of consciousness.

    Anders, in my most recent post I spell out a schematic view of what is going on that I think is consitent with the boxology you describe. However, a lot of that will hinge on how literally to take the boxology.

  28. > Anders, in my most recent post I spell out a schematic view of what is going on that I think is consitent with the boxology you describe. However, a lot of that will hinge on how literally to take the boxology.

    I tried to indicate that I use the boxology only as an expository device to explain a view to people already sympathetic to functionalism. I don’t really place any weight on it.

    All I really want is that at the person level there are such states as experiences construed as concept-exercising intentional states which are distinct from judgements or inclinations to judgement. I don’t really place any weight on the box metaphor as an account of actual sub-personal or design stance facts. It just helps me explain this in terms a naturalist might understand.

    This seems to me enough accomodate the intuition that there is a phenomenlogical difference between sight and even the most super blindsight — yes, sight involves visual experiences, blindsight would not — without positing non-conceptual acquaintance with ineffable private qualia, or anything of that dismal ilk. So it gives us the best of both worlds, and no problems.

  29. Pete Mandik says:

    Hi Anders,

    I appreciate the clarification. I’m glad to see that we agree on the following:
    1. That the target phenomena can be accounted for in terms of concept-excercising states
    2. That the target phenomena can be accounted for without postulating non-conceptual aquantainces with ineffable private qualia

    What we disagree on is whether we need multiply attitutudes beyond belief for the particular task at hand. I worry that your considerations in favor of the multiplication will instead just let in through the back door the nasty things mentioned in 2.

    I’ll take this up further a little later in my response to you over at the thread for this post’s sequel.

  30. [...] also, some of my earlier posts on senses of “seems”: Bursting Apart at the Seems; Bursting at the Seems 2: Electric Boogaloo; Transcending [...]