Space Perception, Visual Dissonance and the Fate of Standard Representationalism

Presenter: Farid Masrour, Harvard University

Commenter: Adam Pautz, University of Texas Austin

Advertisements

15 Comments

  1. Hello Farid,

    This is a very interesting paper containing much valuable information about “traditional representationalism” which I’ve been struggling to fully understand and document after many years of mostly reading expert descriptions of such, and interviewing/questioning said experts, where possible.

    I consider myself to be a supporter of “Representational Qualia Theory” as is currently described at Canonizer.com as an emerging consensus statement being agreed on, so far, by almost 75% of the survey participators. Obviously, all these experts support the general idea that there are two dramatically different parts of perception. First, there are the initial causal properties of perception, such as the properties of a ripe strawberry when it reflects something like 650 NM light. Also, causally, at the other final result end of the perception process, there is our knowledge of such, which has the phenomenal qualities, like redness, which our consciousness uses to represents such initial causal properties.

    There are obviously many diverse ways people think of “Representationalism”, “Sense Data Theory”, “Russellianism”, “Objectivism” and so on, in the literature. But despite my best efforts to understand many of the traditional descriptions of such, in the literature, I have so far completely failed. You’ve also provided some good concise descriptions of some of these, not unlike what I’ve struggled so long to understand in the literature. So I hope I can use these concise summaries of the literature you’ve provided to illustrate some of what I’m struggling with.

    For example, you describe “Russellianism” as one of the doctrines of “Traditional Reprsentationalism” with: “Contents are structures of object and their represented properties.” When you talk about “Contents”, “structure”, “objects” and “represented properties”, I have no idea whether each of these is referring to initial causal properties or final qualitative properties.

    When you talk about “Representationalism”, as a “thesis” of “Standard Representationalism” as “two perceptual experiences have the same phenomenal character, if they have the same representational content.” By “representational consent” are you talking about the initial causes being represented, or the final conscious qualities of our knowledge, representing them?

    And Finally, the most obvious ambiguity exists in your description of “Objectivism”: as “The spatial properties that experience represents are not mind-dependent properties”. This could mean either of the following very opposite ideas:

    Objectivism_1: The initial causal properties of the perception process, are not mind dependent properties.

    Objectivism_2: The qualities we experience of the final results of the perception process, are not mind independent properties.

    I hope this illustrated the problem I, as someone who considers himself to be a ‘representationalist’ has with so many “traditional” descriptions in the literature. If you are attempting to describe Objectiism_2, what could that possibly have to do with any type of “representationalism” whether “traditional” or not? And if you are talking about Objectivism_1, we can now definitively say that at least multiple experts currently agree that this is what repesentationalism is(since the supporters of such have effectively signed the online statements by supporting the camps) – but if you mean this here, what does the initial causal properties of perception have to do with the any nature of any representations of and judgments about such – whether veridical or not?

    My hope is that if I can better understand what you are saying about these things, I might be able to better understand the rest of your paper, possibly even to the extent that we might be able to get such ‘canonized’, so we might find out how many people may or may not agree with any of these arguments, and how this is trending compared to any competing arguments? And of course, whether I can understand it or not, it’d still be great if we could include any of this in the survey, if others that do understand it, can help with the canonization and support of such.

    Upwards,

    Brent Allsop

  2. Dear Farid, yours is a sophisticated paper while mine is a very sketchy idea by a young student…it may well be there are reasons you mentioned in your arguments to reject the line of thought I put forward! You carefully considered a wide range of strategies the Representationalist may adopt to account for visual dissonance.Yet another possible option comes to my mind, as it has been frequently employed in past debates. A representationalist may argue that shape and spatial properties of the scene relative to the subject of experience’s point of view constitute the non-conceptual content of the experience, while properties such as the non-perspectival size of the angle are conceptually represented. Perceptual phenomenology is determined by non-conceptual content only, so the the experiences E-up and E-tilt are phenomenally different because they represent different perspectival spatial properties, included the shape of the angle formed by the V sign as it presents to the subject in the two different ways – when upwards and when tilted -, modes of presentations that are influenced by subjective perspective. The non-perspectival size of the angle, instead, is represented by applying concepts – e.g. that determinate size.
    Sorry for the very rough way in which I sketched the strategy, but I would like to know what you think about that.
    Thank you!

    Giulia Martina

  3. Dear Giulia, Thanks for your apt comment. My impression is that the view that you’re sketching here is a version of strong perspectivalism. In the paper, I characterize strong perspectivalism as the view that we only experience perspectival properties, which I take to be equivalent to the view that only representations of perspectival properties contribute to perceptual phenomenology. I agree with you that adopting strong perspectivalism seems promising as a strategy for blocking my argument. But I give two replies to this strategy in the paper. One is to directly argue against it strong perspectivalism (pp 9-10). I think the view is phenomenologically under-motivated and has trouble accounting for things such as the fact that accessing perspectival spatial properties is very difficult. I wonder you made of the arguments in that part of the paper. I also think that the challenge the dissonance presents is independent from the issue of perspectivalism. To substantiate the claim I give a second argument (the inter-subjective argument) where the two experiences that are being compared are experiences of the same perspectival property. So the differnence between their phenomenal characters cannot be accounted for in terms of the differences in perspectival properties.

    My conviction that the challenge that visual dissonance poses is independent from the issue of perspectivalism is also based on something that I don’t talk about in the paper. Some of the empirical evidence that supports dissonance shows that our experience of angle size depends on non-perspectival factors. For example, compare E-tilt with E-tilt*, where E-tilt* is the same physical angle as T-tilt which is embedded in the same exact plane as E-tilt but has a different orientation, e.g., rather than pointing away from you is pointing to the left. E-tilt* and E-tilt have the same perspectival size because they are embedded in the same plane. but classic empirical findings show that E-tilt and E-tlit* are experienced as having different sizes.

    I hope this gets the ball rolling and am looking forward to seeing what you think.

    cheers,
    Farid

  4. Hi Brent. Thank you for your comment. If I understand your questions, then my reply is this: in the case of the Russellian view, by objects and their properties I mean the objects of perception and their properties. These are often the causes of our perceptual experiences, but they don’t have to be (for example in cases of hallucination they are not). The situation is, in a way, similar to the contents of thought. On a Russellian view of content, the content of the thought ‘Obama is tall’ is a structured entity made of Obama and the property of tallness. Obama is the intentional object of your thought and tallness the property that is attributed to it. Similarly, a Russellian representationalist about perception will say that one’s perceptual experiences have intentional objects and properties that they attribute to them. If I’m hit in the head and as a result start having a hallucination that presents Obama to me as tall, then Obama is the object that enters content and tallness is the property that enters content. However, Obama and his tallness were not the causes of my perception. Of course, some people deny that hallucinations are perceptions. but for the moment, I’m setting that issue aside. Let me know if this helps.

    cheers,
    Farid

  5. Dear Farid, thank you for anwsering.

    I think I can accept your perspective and comprise the conceptual/non-conceptual content strategy under Strong Perspectivalism. However, this depends on how we read your characterization of Strong Perspectivalism. In the paper you write: “Strong perspectivalism implies that E-up and E-tilt do not have a component that represents the non-perspectival size of the V-sign”, yet according to the conceptual/non-conceptual content strategy the two experiences do have a component – namely, a conceptual component – representing the non-perspectival size of the V sign. Instead, in the above comment you characterize Strong Perspectivalism as the view that “we only experience perspectival properties”, where “experience” may be taken as implying a phenomenal consciousness of the represented properties. Even in this interpretation, however, the conceptual/non-conceptual content strategy may not fall under your characterization. We can think that how we conceive of the non-perspectival size of the angle can have a cognitive influence on how we perceptually and non-conceptually experience the perspectival spatial properties of the scene so resulting in a phenomenal difference; or we can think that experiencing an object as falling under a certain concept (conceptually attributing a property to the object) is phenomenally different from experiencing the object as falling under a different concept or under no concept at all.

    Anyway, I now try to look at your arguments at pages 9-10 against Strong Perspectivalism and see if I understand them properly. I just tell you what I intuitively think.

    1) You write: “First, introspection does not support the idea that we only experience perspectival properties. For example, it does not seem that we only experience an elliptical shape when we look at a tilted coin.” Well, prima facie it does not seem to me we experience the elliptical shape and also the round shape, at least not it the same sense of “experience”. For instance, we don’t see the tilted coin as elliptical in the same time as we also see the tilted coin as round. It would result in a contradictory visual experience.

    2) “This is also reflected in our reports. When we wonder about what we normally experience, the answer often includes non-perspectival properties.” Sure, but the conceptual/non-conceptual content strategy may account for that. What the conceptual content of our experience represents may well be reflected in our reports. Maybe, the point of the idea that we conceptually attribute the property of having such-and-such a non-perspectival size to the angle is that this shows up in our beliefs, though we can sometimes be misguided by how angles appear perspectivally.

    3) “It also makes much sense that we experience non-perspectival properties because those are the properties that we need to detect and respond to for the purposes of survival.” I completely agree. A view that simply denies that we somehow represent and can consciously access non-perspectival properties misses this important point. However, it intuitively makes sense to me that we can experience non-perspectival property “indirectly”, so to speak. After all, how the objective world seems to us as we experience it is dependent on our having a point of view on the world. This doesn’t rule out the possibility of accounting for such perspectivalness in representational terms nor it rules out the possibility of conceptually attributing objective non-perspectival properties somehow on the grounds of our experience of – always objective – perspectival properties.

    4) “Finally, under Perspectivalism, it is much harder to explain the fact that it took us such a long time to invent perspective in paintings. As Figure 5.1 illustrates the invention of perspective in painting involved inventing means to aid us ‘see’ the perspectival properties of objects… As Schwitzgebel (2011) has also argued, the view is grounded in over-analogizing visual experience to painting or photography.” It curiously seems to me that your 4) objection is over-analogizing visual experience with painting, so maybe I don’t understand it correctly. I tend to believe that one constructing an explicit codification of the rules of pictorial perspective can be very difficult even if we normally experienced perspectival properties only – just as it is very difficult to make the rules at the basis of normal vision explicit. Moreover, the way things appear in representational paintings following the rules of, say, Albertian perspective is prima facie very different from the way things appear perspectivally in real life perception…

    However, I understand your idea that the challenge visual dissonance presents is independent from the issue of perspectivalism. I will devote some time in the next days to give a closer look at your inter-subjective argument and tell you what I think.
    Sorry if I’ve been confusing in some of my comments. I should also note I am not fully supporting the Perspectivalism strategy, I’m just trying to explore every open path the representationalist has to defend her general theory of perception in problematic cases. This is because, I think, intuituitions are on the representationalist side as far as perception is concerned.

    Tell me what you think, I’ll be back soon!

    Giulia

  6. Hi Farid, thanks a lot for this really cool paper. Two questions:
    1. Your strategy for arguing in favor of egalitarianism is that you say at the beginning that it is not a crazy view and then in Section 3 you point out about three ways of arguing against egalitarianism that they don’t work. Do you have a positive argument in favor of egalitarianism?
    2. Here is the way I think many of us would describe the example you use (maybe you did address this view, in this case, could you point to the discussion?): content is a matter of attributing properties to an object. If the object does in fact instantiate these properties, then the representation is veridical. So in the case of your example, one of my experiences is indeed veridical, whereas the other one is not – the one that attributes properties that are in fact instantiated is veridical, the one that attributes properties that are not instantiated is not veridical. I can see ways of questioning this if the properties in question were perspectival, but as you argue at length, they are not. So what’s wrong with this story? I don’t think your objection against the ‘isomorfism’ view addresses this, nor do the other two objections.
    3. I’m not sure I see the ‘Objectivism’ premise as a necessary part of Standard Representationalism. This is obviously a sociological claim, but my conjecture would be that the majority of representationalists would allow for mind-dependent properties. But putting the sociological claim aside, is there any reason that should move the representationalist to accept Objectivism?
    Thanks again for the paper,
    Bence

  7. Dear Giulia,
    Thanks for your thought provoking replies. Here are some reactions on my side:

    Your comment 1:
    “I think I can accept your perspective and comprise the conceptual/non-conceptual content strategy under Strong Perspectivalism. However, this depends on how we read your characterization of Strong Perspectivalism. In the paper you write: “Strong perspectivalism implies that E-up and E-tilt do not have a component that represents the non-perspectival size of the V-sign”, yet according to the conceptual/non-conceptual content strategy the two experiences do have a component – namely, a conceptual component – representing the non-perspectival size of the V sign.”

    My reply:
    This is a good point. But let me clarify an assumption that I was relying on implicitly. I distinguish components of experiences (which are themselves experiences) from contents. So even if I accept your proposal that E-up and E-tilt each have a conceptual content that represents the non-perspectival size of the angle, I wouldn’t say that there is a component of E-up and T-tilt that does so, unless the representation of the non-perspectival property has a phenomenology associated with it. But if you accept that there is such a phenomenology then you have given up strong perspectivalism.

    Your comment 2:
    “…, prima facie it does not seem to me we experience the elliptical shape and also the round shape, at least not it the same sense of “experience”. For instance, we don’t see the tilted coin as elliptical in the same time as we also see the tilted coin as round. It would result in a contradictory visual experience.”

    My reply:
    I don’t see why experiencing the coin as perspectivally elliptical and non-perspectivally round at the same time is contradictory. As you yourself admit, these might be different senses of experiencing. It would be a contradiction to experience the coin as non-perspectivally round and non-perspectivally elliptical at the same time. I think perspectival seeing is partly a matter of projectively imagining in perception. We don’t see the elipticality of the tilted coin, but we can easily imagine the shape of its projection on a two-dimensional surface. More on this below.

    your comment 3:
    “What the conceptual content of our experience represents may well be reflected in our reports. Maybe, the point of the idea that we conceptually attribute the property of having such-and-such a non-perspectival size to the angle is that this shows up in our beliefs, though we can sometimes be misguided by how angles appear perspectivally.”

    My reply:
    Nice point. I completely agree with you. The argument from report to the claim that we experience non-perspectival properties has a big gap. I have to either reject the conceptual content view or argue that these conceptual contents make their independent contribution to the phenomenology of perceptual experience. I haven’t done any of these in the paper. I think I can do the former. I’ll say a little more about it below.

    your comment 4:
    However, it intuitively makes sense to me that we can experience non-perspectival property “indirectly”, so to speak. After all, how the objective world seems to us as we experience it is dependent on our having a point of view on the world. This doesn’t rule out the possibility of accounting for such perspectivalness in representational terms nor it rules out the possibility of conceptually attributing objective non-perspectival properties somehow on the grounds of our experience of – always objective – perspectival properties.

    My reply:
    If by experiencing non-perspectival properties indirectly you only mean that the phenomenal facts that comprise experiences of non-perspectival properties are somehow temporally posterior to the experience of perspectival properties, then you’re not a strong perspectivalist and your view would not block my argument. A strong perspectivalist denies that we experience non-perspectival properties. There is another way to read “indirectly” namely, that facts pertaining to the experience of non-perspectival properties are grounded in facts pertaining to the experience of perspectival properties. I take this to mean that strictly speaking we don’t experience non-perspectival properties, which would be compatible with strong perspectivalism. I reject this version of the indirect view.

    your comment 5:
    “However, I understand your idea that the challenge visual dissonance presents is independent from the issue of perspectivalism. I will devote some time in the next days to give a closer look at your inter-subjective argument and tell you what I think.”

    My reply:
    I’m looking forward to seeing what you think about that argument.

    Your comment 6:
    “Sorry if I’ve been confusing in some of my comments. I should also note I am not fully supporting the Perspectivalism strategy, I’m just trying to explore every open path the representationalist has to defend her general theory of perception in problematic cases. This is because, I think, intuituitions are on the representationalist side as far as perception is concerned.”

    First of all, let me tell you that I think your comments are awesome and I’m very glad to have had the chance to see them. Second, let me ask you a question about the view that you’ve sketched here. I’m not seeing why you draw a parallel between the conceptual/ non-conceptual spatial content on the one hand and the perspectival/non-perspectival spatial experiences on the other. Usually, those who support non-conceptual contents do so in three ways: (a) by pointing out that certain aspects of perceptual content are rich and go beyond our conceptual repertoire (b) by arguing that animals are capable of perceptual experiences but do not posses concepts or (c) by relying on the empiricist maxim that since concepts are acquired from perception they cannot be required for it. I don’t see how any of these considerations might entitle us to treat experience of perspectival spatial properties differently from experiences of non-perspectival spatial properties. In other words, if you have a convincing argument that experiences of perspectival properties have non-conceptual content, then the argument, is one of (a), (b), or (c). But these arguments, if plausible, apply equally to experiences of non-perspectival properties. So I don’t see why one of these experiences should be regarded as conceptual and other non-conceptual. If it’s not too much trouble, explain to me why you treat them differently.

    cheers,
    Farid

  8. Hi Bence,
    Thanks so much for reading the paper and the comments. Let me know what you think about the following:

    1. “Your strategy for arguing in favor of egalitarianism is that you say at the beginning that it is not a crazy view and then in Section 3 you point out about three ways of arguing against egalitarianism that they don’t work. Do you have a positive argument in favor of egalitarianism?”

    I take myself to have given a positive argument in section 4 where I argue that naturalists are committed to egalitarianism. This said, I should add that my argument there is not a knock down argument. Basically, I argue that to avoid egalitarianism the naturalist has to make some ad hoc assumptions. Of course the naturalist can make these ad hoc assumptions. But since these assumptions are empirical assumptions their status would be worse than philosophical ad hoc assumptions.

    2. Here is the way I think many of us would describe the example you use (maybe you did address this view, in this case, could you point to the discussion?): content is a matter of attributing properties to an object. If the object does in fact instantiate these properties, then the representation is veridical. So in the case of your example, one of my experiences is indeed veridical, whereas the other one is not – the one that attributes properties that are in fact instantiated is veridical, the one that attributes properties that are not instantiated is not veridical. I can see ways of questioning this if the properties in question were perspectival, but as you argue at length, they are not. So what’s wrong with this story? I don’t think your objection against the ‘isomorfism’ view addresses this, nor do the other two objections.

    I can accept your assumptions that content is a matter of attributing properties to an object and that the veridicality of the content depends on the whether the object instantiated these properties. But in order to conclude from this that one of the experiences is non-veridical, one needs to show that the two experiences attribute different properties to their objects. Section 4 argues that if your criterion for determining the properties that these experiences attribute to their object is a naturalistic criterion, then your committed to the conclusion that they attribute the same property.

    3. I’m not sure I see the ‘Objectivism’ premise as a necessary part of Standard Representationalism. This is obviously a sociological claim, but my conjecture would be that the majority of representationalists would allow for mind-dependent properties. But putting the sociological claim aside, is there any reason that should move the representationalist to accept Objectivism?

    My target is a representationalist view whose ultimate goal is to provide a naturalist solution to the metaphysical problem about qualia. I think that the denial of objectivism would be incompatible with this ultimate goal. For example, an externalist representationalist who thinks that colors are dispositions to cause this or that type of color experience in us cannot achieve this goal. The success of the strategy requires that we do not analyze the properties that enter contents partly in virtue of phenomenal properties. Otherwise, the strategy would suffer from circularity.

    I understand that there are ways to read mind-dependence that do not suffer from the circularity problem. For example, by mind-dependent one might mean that the represented properties are relations between physical properties of objects and non-qualitiative properties of the brain (or the body). But I’d like to distinguish mind-dependence from brain/body-dependence. There are no mind-dependent properties in a zombie world.

    Thanks again for the questions. Let me know what you think.

  9. Dear Farid,
    I see we can now enjoy Adam Pautz’s commentary, so you may be busy with answering. I just explain a bit my perspective on the grounds of your replies.

    You wrote: “I’m not seeing why you draw a parallel between the conceptual/ non-conceptual spatial content on the one hand and the perspectival/non-perspectival spatial experiences on the other”.

    This is certainly a legitimate question. I suppose I was trying to point at two different kinds of content or two aspects of the experiential content each involving the representation of different properties – perspectival as opposed to non-perspectival spatial properties. My intuitive idea, just as you underlined, is that the property of being elliptical and the property of being round are arguably not visually and consciously attributed to the tilted coin in the same sense – there are different senses of experiencing, as you wrote. Here your account of what perspectival seeing is may help understand your reasoning. Maybe we can completely ignore the characterization of such different aspects of content, representing different properties, as being conceptual or non-conceptual. Maybe it is enough for the representationalist to say that both perspectival and non-perspectival properties of objects can be represented and somehow phenomenally experienced.

    However, I imagine I have been guided by an idea that may well turn out to be one we can treat without appealing to the conceptual/non-conceptual content distinction (which is a controversial distinction). I’ve been thinking about the role that the recognition that an object posses a certain non-perspectival spatial property plays in our beliefs. As you aknowledged, we can form beliefs about non-perspectival properties of objects even if we are in a sense misled by our perception of perpsectival properties. A way of dealing with this may be to say that such beliefs are grounded in our perception, so that we do somehow perceive non-perspectival properties – we do not infer or suppose that objects have them – while also phenomenally experiencing perspectival properties – our visual experience of the tilted coin presents us with a coin that looks as if it was elliptical.

    Anyway, I think we can reformulate the discussion in terms of properties such as “appearing elliptical (from here/from this point of view)” and maybe describe our experience as attributing to the coin the property of appearing elliptical from here but also the property of being round, without any need to posit conceptual attribution – even if our conceptual recognition that the coin is round shows up in our verbal reports.

    You wrote: “I distinguish components of experiences (which are themselves experiences) from contents. So even if I accept your proposal that E-up and E-tilt each have a conceptual content that represents the non-perspectival size of the angle, I wouldn’t say that there is a component of E-up and T-tilt that does so, unless the representation of the non-perspectival property has a phenomenology associated with it. But if you accept that there is such a phenomenology then you have given up strong perspectivalism.”

    I think the perspectivalism I was inclined towards is a weaker one – and this is why I was doubtful about comprising “my” approach under your Strong Perspectivalism title. I suspect a representationalist can account for the difference in overall phenomenal character between the experences you cite, and she would deny there are phenomenally experienced components of the experience that give rise to a different appearance of the scene each – such that what the non-perspctival size of the angle is represented in the content of each component and the phenomenal characters of each component differ.

    As for your inter-subjective argument, I confess I have some trouble understanding it! I’m not sure I grasp what the two experiences being compared are like, and exactly what properties you believe their contents represent. I’ll read Adam Pautz’s commentary and see if it helps.

    I leave you with a general – methodological – question. What do you think about constructing supposed counter-examples to representationalism – standard or whatever – as a strategy to show that representationalism is false? This has certainly been a prevailing approach: find two experiences having the same content but different phenomenal character and representationalism – at least, strong pure representationalism – will be defeated. The problem is there’s a common strategy on the opposite side: the representationalist will somehow re-interpret the experiences at issue and point to a different in their contents. I don’t think either side has ever been truly impressed by such type of objections.
    Nevertheless, it can be interesting and philosophically enriching to investigate such cases in depth – a variety of sophisticated answers developed from such debates. (I myself am very interested on that group of alleged counter-examples to representationalism involving the perception of ambiguous figures). What do you think?
    I’m happy to contribute discussing these sometimes very subtle issues, which usually hide all kinds of philosophical assumptions and riddles…I’ll keep following the discussion!

    Giulia

  10. Hi Farid,

    Thanks for a fascinating paper! I am currently trying to make sense of the data you present, and here is what I came up with. Hopefully, it will be of some assistance.

    Let me set aside the issue of Peacocke’s trees and tilted coin cases, as you justifiably do. The case you describe (if genuine) is significantly different, I think.

    It seems that you think (although this is not explicit) that we cannot “read-off” the objective-angle-related content of E-up (and also E-tilt) from its objective-angle-related phenomenology. For example, merely on the basis of that phenomenology, we cannot judge that the angle is, e.g., 20 degrees. The V does not phenomenally seem to be 20 degrees (on your view), right? The reason I attribute this claim to you is that, if, e.g., we could read-off “20 degrees” from E-up phenomenology and “24 degrees” from E-tilt phenomenology, then it would be clear that (at least) one of them is misrepresenting reality, and it would be clear which one it is. Am I right that this issue plays an important role in your view?

    Now, what we realize, introspectively, is that the two experiences “seem to disagree about the size of the angle” (p. 2). Something odd seems to happen. The experiences in question differ in their “objective-angle” phenomenology, so it prima facie seems that one of them must misrepresent, but we cannot phenomenologically tell which one it is (because we cannot read-off the content). Moreover, your considerations from covariational accounts of representation seem to show that the two experiences have the same content. So we have the following three claims:
    1. One cannot read off the objective-angle-related contents of E-up and E-tilt from their respective phenomenology.
    2. The two experiences appear to differ in their objective-angle-related phenomenology. (the difference is related to a difference in the size of the angle).
    3. The two experiences have the same content.

    Your account of this situation is that standard representationalism is false.
    I would like now, first, to mention an example of an illusion that appears to satisfy claims analog to 1-3. I will then attempt to analyze it within standard representationalism, with the aim of suggesting that perhaps such an analysis could also make sense of the E-up, E-tilt case.

    Consider the Muller-lyer illusion. In this case it seems that we feel, phenomenologically, that one of the lines is longer than the other (satisfying the analog of 2), but we cannot say which of the lines is veridically represented with respect to length (which satisfies the analog of 1). It strikes me as plausible (or at least not implausible) to say here that the length of each line is veridically represented (in accordance with 3). What do you think? Is this case interestingly similar to your E-up E-tilt case? (I think the waterfall illusion might also be relevant).

    Now, it seems to me that the representationalist could handle the Muller-Lyer case by saying, first, that the length of each line is veridically represented, but the experience misrepresents the relation “longer than” as obtaining. The next step is to say that the phenomenal property associated with the representation of the length of each line is the same. The phenomenology of “longer than” is accounted for by the visual (mis)representation of the “longer than” relation. This appears to preserve standard representationalism.
    Perhaps, then, a similar analysis is available in the case of E-up and E-tilt. The idea would be that the respective (angle-related) contents are identical, and the phenomenal property associated with each is also identical, but in addition there is a visual (mis)representation of the relation “being bigger than”, which accounts for the phenomenological observation from 2. Could this save standard representationalism?
    (There is at least one important difference between the cases, which is that the E-up and E-tilt experiences are occurring one after the other whereas the lines in the Muller-lyer illusion occur simultaneously. Perhaps this is not an important difference since arguably we can have visual representations of change over time, and so the relevant visual representation that accounts for 2 is a representation of the angle as changing its size across time. Here the waterfall illusion seems relevant: an object is represented as changing its location over time, but each repersentation of it at a time is veridical with respect to its location).

    Thanks again for a challenging paper!

    Assaf

  11. Dear Adam,
    Thanks so much for your terrific comments on the paper. I’m very grateful. They were extremely helpful and clarifying. Also, sorry for the long delay in my reply. I had a lot on my plate for the past few days, but I also had to think hard about the comments. I confess that I haven’t been able to address all the points that you raise. Here I focus on the five following issues that were more central.

    1- The difference between my case and Peacocke’s tree case
    2- The issue of Perspectivalism.
    3- Experiences and their components.
    4- Phenomenal difference
    5- the inter-subjective argument.

    1. The difference between my case and Peacocke’s tree case
    One of Adam’s main comments about the paper is that he is not sure “why this example is supposed to be tougher for standard representationalists than other examples in the literature (Peacocke’s two trees, the tilted penny).” He says later that “Farid hasn’t identified a clear difference between cases like the tree and the coin and his “fingers” case.” But he continues to say that “I agree with Farid that Seeming In- compatibility doesn’t hold in the tree case and the coin case, but I disagree with him when he suggests that it holds in his fingers case.”

    I think there is a very important difference between my case and Peacocke’s tree case. I don’t think that there is even a seeming disagreement (or as Adam puts it seeming incompatibility) about the non-perspectival size of the trees in Peacocke’s case (or if there is, this difference plays no part in the argument). But in the case that I present, I think there is a seeming disagreement between E-up and E-tilt about the non-perspectival size of the angle. Adam disagrees with this claim, but it is puzzling to me that he thinks that I haven’t identified a difference. It is one thing to say that an important difference between my example and the Peacocke’s tree case has not been identified and it is another thing to say that one disagrees with the claim that there is a difference. Suppose that I tell you that Obama is in Oakland to compete in a hip hop dance contest today, but you disagree, holding that Obama is in the white house. Then it is incorrect to say that I have not made a claim about Obama’s whereabouts. I have made such a claim, but you think my claim is wrong.

    Suppose you present subjects with angles of different sizes facing up and then at a later stage with the same angles tilted and ask the subjects to identify how they perceive the non-perspectival size of the angles by either reporting their estimate of the non-perspectival size of the angle or by matching a sample angle. Suppose that it turns out that the verbal report and the matching performance systematically depends on the degree of the tilt of the angle. Then, if we take the subjects’ performance in such a setting as indicative of the way that they are consciously perceiving the non-perspectival size of the angle, we should conclude that there is dissonance: the experiences of the same angle once facing up and once tilted seem to disagree about the non-perspectival size of the angle. Now, I don’t know of any actual empirical study that has done this exact experiment. But the empirical literature that supports Visual Dissonance is of the same kind. You present subjects with the same physical angle or physical line with different orientations or degrees of tilt and study their reports or matching performance pertaining to the perception of the angle size, and the length or curvature of the line. And the results indicate that performance in such cases depends on variations in degrees of tilt and orientation.

    Let’s call the hypothesis that E-up and E-tilt seemingly disagree about the non-perspectival size of the angle, SD. Adam disagrees with SD. His main reason for this is based on introspection. I can’t argue with Adam’s report about the way he experiences the angles. But two points deserve mentioning. First, although my introspection supports SD, I don’t think that my introspection delivers a certain verdict about it. I can’t say that I’m certain about SD. But that’s not to say that my introspection takes a neutral position about the issue. It favors SD. I also don’t think that introspective reports are infallible. Even if my introspection favors SD, SD might be wrong. But I don’t think that we should give up making claims about phenomenal character because of this. There are indirect ways to tap into phenomenal character. For example, since I think that there is good evidence for Visual Dissonance, then I have more reasons to accept SD. In short, my reason for accepting SD is two-fold: direct introspection and indirect empirical support coming from Visual Dissonance. Of course, there is a lot that needs to be said about Visual Dissonance. But I am presupposing it for the purposes of this paper.

    The second point that deserves highlighting is that my main aim in the paper is to argue that Visual Dissonance is an important challenge to standard representationalism. I use the example of E-up and E-tilt to illustrate this challenge. But, as long as Visual Dissonance is correct, the challenge stands even if the specific claim about E-up and E-tilt is wrong. For, if Visual Dissonance is correct, there are pairs of dissonant experiences. In fact, I’m making the task more difficult for myself by choosing E-up and E-tilt as examples. This is because E-up and E-tilt are experiences of the same angle seen from different perspectives. This opens up the possibility of responding to the argument in different ways, for example by adopting strong perspectivalism or access holism. But the threat of strong perspectivalism and/or access holism is much less serious in other cases. For example, there is empirical evidence (Foley 1964, 1972) that our perception of the size of an angle depends on its orientation. Two angles with the same physical size that are embedded in the same plane are perceived differently depending on their orientation. Since the two angles are on the same embedding plane, under the common way of understanding perspectival properties their perspectival sizes will be the same. So strong perspectivalism and/or access holism would be less threatening in this case.

    To summarize, I disagree with Adam that I haven’t identified the difference between my counterexample and the standard challenges to Standard Representationalism such as the tree case. The tree case does not rely on a seeming disagreement about the non-perspectival size of the tree, but my cases do. Second, my justification for believing that there is a seeming disagreement in the case of E-up and E-tilt comes from direct introspection and from the empirical support for Visual Dissonance. Admittedly, this didn’t come out very clearly in the paper. Third, my main aim in the paper is to show that Visual Dissonance is challenging for Standard Representationalism. This general claim does not depend on the specific claim that E-up and E-tilt are example of visual dissonance.

    2. The issue of Perspectivalism
    In the paper, I distinguish between strong and weak versions of perspectivalism. The strong version holds that we only experience perspectival properties. The weak version holds that we experience both perspectival and non-perspectival properties. Adam is not convinced by my argument against perspectivalism and points out that I don’t argue against weak perspectivalism. Here is my reaction. I’ll start with the point about weak perspectivalism.

    Adam is correct that I don’t argue against weak perspectivalism. But, this has a simple reason: I don’t need to reject weak perspectivalism, because weak perspectivalism is compatible with the premise that E-up and E-tilt have an experiential component that represents the non-perspectival size of the angle.

    Strong perspectivalism, on the other hand, denies that we experience non-perspectival properties. This entails that E-up and E-tilt do not have an experiential component that represents the non-perspectival size of the angle. So I need to argue against strong perspectivalism to defend my premise. In the paper, I offer three considerations against strong perspectivalism: first, that our reports don’t support it, second, that detecting non-perspectival properties is more important for survival, and third, that detecting the perspectival properties often is much harder than detecting non-perspectival properties (this is the main point of the painting example depicted in Fig 2.1). Adam is not convinced by the first consideration. He says:

    “First, I don’t think that these are particularly strong objections against Strong Perspectivalism. A natural response is that we do only represent perspectival properties and can attend to them but we typically only care about objective properties, and this is reflected in our thought and talk (Hill 2009, 157-9). So, I think that there is a lacuna in Farid’s actual within-subject argument against standard representationalism: as far as I can see, the standard representationalist might happily just go with Strong Perspectivalism, as for instance Chris Hill seems to do.”

    Adam does not discuss my second and third considerations against strong perspectivalism. Perhaps he thinks that they don’t add much to the first one or that the same response applies to them. I’ll come back to this issue later. But let me start with the first consideration.

    The Pope believes in God, but that doesn’t convince many of us to believe in God. According to Adam, Chris Hill believes that we only experience perspectival properties and our talk and thought only reflects the fact that we typically care about non-perspectival properties. Should that convince us to believe in strong perspectivalism? So far, all we have here is that there is a possible position in logical space that someone occupies and that position disqualifies our reports that we experience non-perspectival properties. But in the absence of a good reason to disqualify such reports, this shouldn’t give us much of a reason to abandon our position.

    Adam’s remark, however, can be taken in a different way. Perhaps, his point is that in the absence of a response to Hill’s claim, we don’t have a particularly strong argument against strong perspectivalism. If having a strong case requires giving a knockdown argument against Hill’s position, then I agree with Adam that we don’t have a strong case. But, there is hardly ever a knockdown argument when it comes to what we experience. All we can do is to evaluate the balance of reasons against and for the various possible positions. And I think the balance of reasons is not in favor of strong perspectivalism. Let me elaborate on this by expanding on the considerations that I offered in the paper. I thank Adam for pushing me to do this.

    Our introspective reports constitute one source of evidence for what we perceptually experience. But there is a problem with relying solely on introspective reports. The reason is that it is not always obvious that what we report is what we perceptually experience as opposed to what we have learned on the basis of what we perceptually experience. So introspective reports in favor of the claim that we experience non-perspectival properties can be explained away. I think this is the strategy that Adam attributes to Chris Hill. But there is a wrinkle here. The problem is that the explaining away strategy can be equally applied to reporting experiences of perspectival properties. Unless more is said, the introspective evidence that we have for the claim that we experience perspectival properties is at best equal in strength to the evidence that we have for the claim that we experience non-perspectival properties. It follows that, unless more is said, we have stronger introspective evidence for weak perspectivalism than strong perspectivalism.

    But introspective reports are not the only sources of evidence about what we experience. Let me consider another source. This will clarify the third consideration that I give against strong perspectivalism in the paper. I’m very good at copying an ellipse if you show it to me on a sheet of paper. But I’m not good at all in copying the alleged perspectival elliptical shape of a tilted coin. This is very puzzling if we assume strong perspectivalism. For, the strong perspectivalist is committed to the thesis that seeing a tilted coin is like seeing an ellipse head-on, at least with respect to the shape properties that we experience. So when I see a tilted coin, setting other things aside, my experience presents to me an elliptical shape in the same way that it presents to me an elliptical shape when I see an ellipse head-on. But then, why is it that I can easily draw the ellipse presented head-on but cannot draw the perspectival elliptical shape of the coin? This is the idea behind the consideration that it took us such a long time to invent perspective in paintings. If we experienced perspectival properties in the way suggested by strong perspectivalism, we wouldn’t have needed to invent perspective. The same reasoning applies when we consider the fact that it takes children so much time to learn to draw 2D images of a 3D scene. Some of us never learn this.

    A similar consideration applies to reports of what we see. Take a triangle and look at it when it is tilted. Try to estimate the perspectival size of its angles. Now, take a picture of the tilted triangle and look at the picture and try to estimate the size of the angles of the projected triangle. I bet you’ll be much worse in the former task than the latter task. But this is very hard to explain under strong perspectivalism. For, under strong perspectivalism, the shapes that you see in the two cases are the same. I’m making an empirical claim here. And I don’t know if anyone has ever tested this. But there are some studies in the ballpark (Gilam 1977, Segwik and Levy 1985) that show that we are often much worse in reporting perspectival properties than non-perspectival properties. (For some discussion see Gilchrist 2012).

    On the view that I favor, the explanation of these facts are easy. Experiencing perspectival properties requires the ability to imaginatively project the scene in front of us onto an imaginary occluding plane. This is a skill that we need to learn through practice. It is easy to apply in certain cases and hard in others. We can make the task easier for ourselves with various methods, one of which is depicted in the Durer painting (Figure 2.1) in the paper.

    Of course, doing justice to this issue requires going over a host of other considerations. Specifically, we need to consider the other reasons in favor of the claim that we only experience perspectival properties. There are a few that suggest themselves. For example, vision starts with receiving a 2D image on the surface of the retina. Relatedly, one might argue that perspectival properties are represented earlier in the visual stream, or perhaps representing non-perspectival properties requires cognition in a sense that representing perspectival properties does not. I don’t think any of these considerations can be turned into convincing arguments and I’d be happy to explain why. But I’ve already said a lot about this. Let’s move to the third issue.

    3. Experiences and their components
    Adam is suspicious about my talk about experiences and their components. He says:

    “… I just have a hard time understanding these claims about the components of my experiences. Suppose you view your V straight on, and then tilted. Now, most philosophers hold that when you view V titled, you undergo a token experience E. But Farid’s argument requires more. It requires that E has a part, namely, TILTnp, which just represents the non-perspectival angle and nothing else. Is he right?”

    Part of the reason that Adam is suspicious about experience talk is that in ordinary English ‘experience’ is not used to pick out an alleged internal mental event. Adam thus avoids talk about experiences all together. Instead he talks about “subjects and their experience properties.”

    I agree with Adam that issues about individuating experiences are very murky. But we need to have a framework at place before formulating and scrutinizing views like Standard Representationalism. The framework that I presuppose without argument in the paper is that we should individuate experiences in terms of instantiations of phenomenal properties by subjects. If you adopt this framework and accept my claim that when you look at the tilted angle you instantiate a phenomenal property pertaining to the representation of the non-perspectival size of the angle and nothing else, then it is trivial that you have an experience that represents the non-perspectival property of the angle and nothing else. This is because according to this framework there are as many experiences within a subject as there are instantiated phenomenal properties.

    Some theorists think that the notion of a phenomenal property is suspicious. But Adam does not seem to belong to this group. He talks about experience properties, and I don’t see much difference between experience properties and phenomenal properties. But if we’re working within a framework that recognizes talk about phenomenal properties, I don’t see any convincing reason yet to discount my framework in favor of Adam’s that replaces talk about experiences with talk about experience properties.

    Admittedly, Adam might convince me that there are good reasons to adopt his framework. But I don’t see how that would change anything substantive about my argument. Note that, I formulate representationalism in terms of experiences. If it turns out that we should abandon this way of talking, then we have to properly modify both the formulation of representationalism and my argument. For example, rather than saying that the subject has an experience that represents a non-perspectival property and nothing else, I should say that the subject instantiates a specific experiential property pertaining to the representation of a non-perspectival property and nothing else. But I don’t see how anything substantive in the argument would change. The argument would still show that two different phenomenal properties (or experience properties) can have the same representational contents, which shows that standard representationalism is in trouble. So I think Adam’s worry about the individuation of experience can be dealt with to his satisfaction.

    But there is another worry that Adam has with my claim that we have an experience that represents the non-perspectival size of the angle. He says:

    “It’s true that when one experiences the V angle at a certain degree of tilt one instantiates more than one phenomenal property pertaining to shape. For instance, one instantiates a rather determinate phenomenal property P that’s common between cases in which one experiences the angle V at the relevant precise degree of tilt. One also instantiates a more determinable experience property D that is common between cases in which one experiences the angle V at various degrees of tilt. (There’s a rough phenomenal similarity between these cases and this determines a determinable experience property, D.) But I don’t see how this entails that when one experiences the V at a certain degree of tilt one undergoes two experience events, one representing the non- perspectival angle and nothing else.”

    Perhaps I’m missing something, but I don’t see the challenge here. Given the above discussion, the issue is not how many experiences there are. The issue is what phenomenal properties are instantiated. More precisely, the issue is: when I have E-tilt (or E-up) do I instantiate a phenomenal property pertaining to the experience of the non-perspectival size of the angle and nothing else? In my view, there is a way that I’m experiencing the non-perspectival size of the angle which is a determinate of the same determinable that is instantiated when I experience the same angle with a different degree of tilt. If there is such a way then I’d say a phenomenal property pertaining to the non-perspectival size of the angle is instantiated. Has Adam given us a reason to be suspicious about this? I’m not sure.

    This ends what I have to say about the issue of the individuation of experiences and its use in my argument. I short, I agree with Adam that the issue of individuating experiences is difficult and perhaps we should abandon talk about experiences in favor of Adam’s framework. But I don’t see that anything substantive about my argument depends on this.

    4. Phenomenal Difference
    Setting all of these issues aside, Adam still worries that I might not have well-motivated one of the crucial premises of my argument. Even if we accept that in having E-up and E-tilt we instantiate phenomenal properties that represent the non-perspectival size of the angle and nothing else, we might deny that these components of E-up and E-tilt differ in phenomenal character. In response to my claim that these components must be phenomenally different because they give us a different impression of the angle, Adam writes:

    “The perspectivalist doesn’t just think the consecutive total experiences represent an objective angle. He also thinks that they represent distinct perspectival angles. Maybe this gives us, or can give us (if we take the painterly attitude), the (mis)impression that they represent different non-perspectival angles simpliciter, thus explaining Seeming Incompatibility. (A kind of introspective mistake.) This explanation does not require that UPnp and TILTnp differ in phenomenal character. It just requires that the total experiences differ overall in phenomenal character.”

    In the paper, I discuss this exact response under the title of access holism (pp.11-12). I describe access holism as the view that in the context of different degrees of tilt, our access to E-upnp and E- tiltnp can be biased.

    Let me just repeat what I said in response in the paper. I do not think that access holism is false. Our judgments about how things are with us phenomenologically can go wrong in many different ways, specifically as a result of contextual factors. But we should note that one can accept access holism and still hold that there is a difference in phenomenal character between E-upnp and E-tiltnp. Therefore, access holism is not sufficient for denying that E-upnp and E-tiltnp differ in phenomenal characters. Accepting access holism might shake our introspective confidence in Phenomenal Difference, but does not imply that it is false.

    We can happily accept access holism and Phenomenal Difference at the same time, and we can do so justifiably if we have independent reasons to accept Phenomenal Difference. I think we do. The reason, in a nutshell, is that Phenomenal Difference is an implication of the empirical finding that visual experience does not have a constant curvature. As discussed earlier, empirical work on the geometry of visual experience suggests that the spatial properties of objects must change as they move from being presented by one system to another system. The seeming non-perspectival size of an angle changes as the result of some rotations. So if we accept Visual Dissonance, we have good reasons to accept Phenomenal Difference.

    5. The inter-subjective argument
    Adam’s main point about the inter-subjective argument is that it does not target representationalism. I agree with him that the argument does not target representationalism. But I intended it to only target Standard Representationalism. I’m explicit about this at the end of section 2:

    “This ends my argument against Standard Representationalism. Assuming the veridicality of E-up and E-tilt, I have given two arguments against Standard Representationalism that I called the Intra-subjective and the Inter-subjective arguments. (p. 15)”

    So why does Adam think that I was targeting representationalism? He correctly points out that the conclusion of the inter-subjective argument is that representationalism is false. So, what’s going on? Am I being inconsistent?

    Here is what’s going on. I characterize Standard Representationalism as a view that is committed to content Russellianism, Naturalism, Objectivism and Representationalism. I characterize Representationalism as the thesis that necessarily two perceptual experiences have the same phenomenal character if they have the same representational content. Throughout the intra-subjective argument, I’m holding my target as committed to Russellianism, Naturalism and Objectivism. As I point out at several places, a representationalist can easily reject Veridicality if they reject naturalism. They can also block the argument by adopting Fregeanism and/or denying Objectivism. So, strictly speaking, the conclusion of the intra-subjective argument is that if we assume Naturalism, Objectivism and Russellianism, then Representationalism is false. In other words, Standard Representationalism is an inconsistent position. Now this carries over to the inter-subjective argument. The conclusion of this argument is also conditional: if we assume Naturalism, Objectivism and Russellianism, then Representationalism is false. So I don’t think there is a disagreement between Adam and I about this point. There was a mis-communication here that I should take the blame for. I should have made it clear that the conclusion is conditional.

    Seeing this also helps us respond to the other worry that Adam has about the inter-subjective argument. He worries that the representationalist can block premise 6 (E-tiltSam and E-tiltPam have the same representational content.) I agree with this point too. But I think the Standard Representationalist has to accept premise 6. It is entailed by the previous premises that the Standard Representationalist is committed to. I do have worries about representationalism. But that’s a topic for a different paper.

    Adam also finds suspicious my claim that a phenomenal difference between Pam ( a subject who doesn’t suffer from Visual Dissonance) and Sam (who suffers from Visual Dissonance) is conceivable. I might be missing something here, but I don’t see the worry. Pam does not suffer from visual dissonance. So, given that Sam and Pam have the same experiences when they look at the non-tilted angle, Pam’s experience when she looks at the tilted angle (whatever it might be like) is different from Sam’s experience when he looks at the tilted angle. Otherwise, Pam would be suffering from Visual Dissonance too. That’s all we need to substantiate phenomenal difference in the inter-subjective argument.

    This ends my replies. I’m very much looking forward to any reactions.

  12. Hello Farid. I have been following this discussion with interest. Of the many points that have occurred to me while reading and rereading this material, there is one in particular that bugs me and that I would like to raise here.

    In your reply to Adam Pautz, you have made a clarificatory remark that makes explicit something that I think was only implicit in your paper. You write:

    “The framework that I presuppose without argument in the paper is that we should individuate experiences in terms of instantiations of phenomenal properties by subjects. If you adopt this framework and accept my claim that when you look at the tilted angle you instantiate a phenomenal property pertaining to the representation of the non-perspectival size of the angle and nothing else, then it is trivial that you have an experience that represents the non-perspectival property of the angle and nothing else. This is because according to this framework there are as many experiences within a subject as there are instantiated phenomenal properties.”

    So basically, you regard an experience as a phenomenal property that the subject is aware of, and you regard a representationalist experience as just such a phenomenal property that is associated in some way with whatever is represented (a content, a distal property such as the nonperspectival size of an angle, an object, etc.).

    Is this correct? If it is, then it presents a serious problem, which is that representationalism can’t be characterized this way. Representationalism holds that the subject of an experience is per se aware only of whatever the experience represents, not of sensations, modes of presentation, or qualia — or phenomenal properties distinct from what is represented.

    Perhaps I have you wrong and you hold that in a representationalist experience, the phenomenal property is identical with what is represented. But there are places in your original paper where that does not seem to be your view. For example, your argument against the idea that visual dissonance involves the misguiding of the subject’s beliefs hinges on the following (p. 20):

    “Because of dissonance, the same spatial quality gives rise to different looks depending on the system that it is presented to. A right angle does not have a single look and cannot be associated with a single phenomenal property. Thus if one associates the concept of a right angle exclusively with the looks that it produces in the Up system, then one will be led to form false beliefs when right angles are presented by the Tilt system. But this only reflects a flaw in conceptual competence. A fully competent user of the concept of a right angle will associate it with different looks in the different systems, and will avoid forming false beliefs.”

    I take it that “look” and “phenomenal property” refer to the same thing here (as well as in the passage above from your reply to Adam). So what is represented, the right angle, is only associated with the look; they are not identical. The subject of an experience has access to the look independent of what objective property the experience represents. Obviously, this is not representationalism. And this matters a great deal, because the argument depends on the dissociation of looks from what is represented. On a true representationalist theory, a veridical experience of a right angle would represent a right angle, and the subject would per se have access to nothing else. The phenomenal character of the experience would be: “right angle.” And this would be true for both the Up system and the Tilt system.

    Thus, this argument is question-begging. It attempts to explain how visual dissonance subjects might form false beliefs from their experiences despite the veridicality of those experiences. But the suggested account takes for granted the falsehood of representationalism — something the representationalist will certainly not grant.

    You may say that, at this stage of your paper, you are not arguing specifically against representationalism but are only seeking to show how certain possible objections to the claim that dissonant visual experiences can all be veridical can be removed in general. Thus, once standard representationalism has been cleared away, we can see how seeming objections to the veridicality of dissonant experiences can be handled. But this is not the situation at all. Defense of the veridicality thesis is only important as an element in the argument against standard representationalism: you need to show that a pair of dissonant experiences can differ in phenomenal character while representing the very same thing. The argument that the experiences represent the same thing is that the same external property (the angle) is being viewed and the experiences are veridical. You need to show that the standard representationalist is committed to the pair of dissonant experiences being veridical. But what we actually have, in the matter of belief formation, is a reason why the representationalist should be committed to these experiences being not veridical.

    Thus, in this passage and in various other places in your paper, it seems to me that you take a nonrepresentationalist view of sense-experience as a starting point. Your tacit view (made less tacit in your reply to Adam) seems to be that an experience consists in awareness of a nonintentional phenomenal property that, when we are fortunate, is associated with an appropriate external item. Now, obviously, you must take whatever view you think is best supported (though I do not share this one — like Adam Pautz, I am an internalist representationalist). But I think it would be better if it were not kept tacit in your paper. This would help to avoid problems like the one — if I have got it right — I have pointed to.

    -David

  13. Dear Assaf,
    Thank you for the really interesting comment and sorry for the delay.

    I’d say the important difference that you mention at the end of your comment is crucial here. You’d get the phenomenal difference between E-up and E-tilt even if you close down your eyes before rotating your finger and your memory of the previous experience is deleted. My claim is that E-up and E-tilt seem to present angles that are different. The point is not that it seems to us that the angles that E-up and E-tilt present are different. In the former, the comparative claim is not in the scope of the verb ‘seem to present’. In the latter, the comparative claim is in the scope of the verb. So I don’t think that your response to the Muller-Lyer carries over to my case (As a side note, I disagree with your interesting take on the ML illusion too. But that’s a different issue.) At least, if you agree with my description of the case E-up and E-tilt do have different angle-size-related phenomenologies and your response is ruled out.

    This said, at the beginning of your comment you say a couple of things that together can be taken as an argument to the effect that I’m not entitled to my claim about the difference in the objective-angle-related phenomenology of E-up and T-tilt. You say:

    “It seems that you think (although this is not explicit) that we cannot “read-off” the objective-angle-related content of E-up (and also E-tilt) from its objective-angle-related phenomenology. For example, merely on the basis of that phenomenology, we cannot judge that the angle is, e.g., 20 degrees. The V does not phenomenally seem to be 20 degrees (on your view), right?”

    Well, yes and no. First, if you define content in ways that naturalistic theories of content understand it (for example, if you think that the content of a state is determined by its causal profile,) then I don’t see any reason to hold that you can read the objective-angle-related content of E-up and E-tilt from their phenomenology. So, ‘No’ if that’s what you have in mind.
    Second, I think that you have introspective access to phenomenal content. So if by content you mean phenomenal content, I’d say you can read off the objective-angle-related content of E-up and E-tilt from their phenomenology. But I think this access is fallible. So ‘yes’ with some qualifications if you have phenomenal content in mind.

    You continue:

    The reason I attribute this claim to you is that, if, e.g., we could read-off “20 degrees” from E-up phenomenology and “24 degrees” from E-tilt phenomenology, then it would be clear that (at least) one of them is misrepresenting reality, and it would be clear which one it is. Am I right that this issue plays an important role in your view?

    Again, the issue is complicated. Even if we focus on phenomenal content, which I think is accessible, it’s not obvious that one of the experiences must be mis-representing if they seem to disagree. For example, if you identify phenomenal contents with modes of presentation, as Fregeans do, then despite the difference in phenomenal content, the two experiences might be attributing the same property to their object. If so, they can be both veridical. Even if we assume Russellianism about phenomenal content and identify phenomenal properties with the properties that experience attributes, it doesn’t follow that one of the experiences will be non-veridical. For, it is possible to hold that the angle instantiates many non-perspectival size properties. Of course, these properties cannot be identical with the physical non-perspectival size of the angle. This view is unpalatable to many, but you can see the intuition behind it if you consider the analogous view in case of color where it is held that each object instantiates many different colors at the same time. If we accept this view (I call it the many-properties view), E-up and E-tilt can be both veridical, although they attribute different size properties to the angle.

    Does this address your worry?

    -Farid

  14. Dear David,
    Thanks for the comment. Like my usage of the term ‘experience’, my usage of ‘phenomenal property’ is neutral about the issue that you’re worried about. As I’m using the term, phenomenal properties can be understood as instantiations of qualia, modes of presentation, relations to an externalistically construed content, or relations to properties of objects, or perhaps something else. I don’t deny the identification of phenomenal properties with representational properties. I’m suspending judgment about the issue. So I’m not taking for granted “the falsehood of representationalism.”

    Does this respond to your worry? Is there a place in the paper were you think I’m assuming a non-neutral position? I’d be very grateful if you could point out such a place.

    Farid

  15. Hi Farid.

    “Like my usage of the term ‘experience’, my usage of ‘phenomenal property’ is neutral about the issue that you’re worried about.”

    Thanks for the clarification.

    “Is there a place in the paper were you think I’m assuming a non-neutral position? I’d be very grateful if you could point out such a place.”

    But I already did! Or so it seems to me. I would be curious to see your response.

    In section 4 of your paper, there are multiple other instances of the sort of problem I had in mind. But on reflection, I don’t think they are very important. They can be reformulated in an acceptable way. To illustrate what I’m talking about, on p. 24 you write:

    “In line with the arguments of the second section of the paper, let us assume that E’s are
    complex with a component, call it Enp, that represents the non-perspectival size of the
    tilted angle. Enp experiences are tilt-independent in the sense that they can be
    components of experiences of non-tilted angles. But due to dissonance, the non-tilted
    angles that normally give rise to Enp are different from the tilted ones. Enp experiences are
    caused by tilted 42º angles but by, say, upward facing 44º angles.

    “This raises a question: what is the content of Enp experiences? We seem to have two
    options. First, we might divide the class of Enp experiences into two, holding that those
    that are caused by tilted angles represent 42º and those that are caused by non-tilted
    angles represent 44. This is the option that in my view the naturalist has to adopt. But
    another option is to privilege the non-tilted angles. On this option, since non-tilted 44 º
    angles give rise to Enp experiences, the content of Enp’s is 44º in both systems. It would
    follow that Enp experiences normally misrepresent tilted 42º angles as 44.”

    In this passage, you introduce an experience type Enp that represents non-perspectival angles and can do so in the context of both Eup and Etilt experiences. You stipulate that Enp is normally caused by different non-perspectival angles in the two contexts. And then you ask what is the content of Enp, meaning, I think, what specific angle does Enp attribute to the V as its non-perspectival angle. And you suggest two possibilities: (a) that Enp might represent one size when the V is up and another size when it is tilted, or (b) that Enp might represent the same size in both cases.

    To see the trouble, consider possibility (a). It violates representationalism to say that one experience type represents two different things. Representationalism customarily individuates experience types by what they represent. So in possibility (a), there is no longer one experience, but two, EnpUp and EnpTilt. EnpUp represents the non-perspectival angle to be 44 degrees, and EnpTilt represents it to be 42. EnpUp and EnpTilt presumably have different phenomenal characters (although they don’t strictly have to, of course). If the experience of the non-perspectival angle is supposed to be veridical in the two orientations, it can’t be experienced by EnpUp and EnpTilt. There must be another experience, say EnpTilt*, that represents a non-perspectival angle of 44 degrees when tilted. This will, according to representationalism, have the same phenomenal character as EnpUp.

    I don’t think this was what you had in mind at all. I think the idea was supposed to be that the same experiences, Enp, with the same phenomenal character, represent two different things in the two contexts (upright and tilted). Since the perception is veridical, a V of 44 degrees would be experienced by Enp when upright and by some other experience, Enp*, when tilted. Both would have the same content, established by causation/covariation, but different phenomenal characters, these being something the experience types have independently. Obviously this is not a bus the representationalist will step onto in the first place.

    You can avoid this difficulty by talking about nonintentional, nonconscious “experiential states” (neural states, perhaps) that become experiences when they get associated with content through some sort of externalistic process (causation, covariation, etc.). Thus, instead of an experience Enp, we would have an experiential state, which covaries with tilted angles of 42 degrees and with upright angles of 44 degrees. As such, it is the vehicle for the experiences EnpTilt and EnpUp in the two contexts. If the standard representationalist wants to avoid this consequence, he needs to come up with a well-motivated reason (which you argue against). If in spite of representing the same non-perspectival property, EnpUp and EnpTilt* have different phenomenal characters, that is the representationalist’s theoretical difficulty. Thus the problem is eliminated.

    But I don’t think this strategy will help out with the problem from section 3.3.

    -David

Leave a Reply

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

WordPress.com Logo

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

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s