Perceptual Content and the Content of Mental Imagery

Presenter: Bence Naney, University of Antwerp & University of Cambridge

Commenter 1: Nick Wiltsher, University of Leeds

Commenter 2: James Genone, Franklin & Marshall College


  1. James, thanks a lot for these comments. I think there are a couple of things I need to clarify. My aim here was not to argue against the Dependency Thesis and for the Similar Content View. My aim was to argue for a specific version of the Similar Content View and show that it may explain some odd features of mental imagery better than either the Dependency Thesis or other versions of the Similar Content View. So my goal here is to do something positive – to give a specific enough version of the Similar Content View that could say something interesting about the similarities and differences between perception and mental imagery.

    I understand that you’re coming from the direction of relationalism and, as a result, you are drawn to the Determinacy Thesis. I also understand that many relationalists are very much moved by the ‘introspectively obvious’ feature of perception that it somehow puts us in contact with the world. As you say, when we see an apple, we experience it as physically present – and this is a very different phenomenology from mental imagery, where there is no such experience.

    I am extremely suspicious of this ‘feeling of presence’ as an essential (or even necessary) feature of perception per se. I do think that some of our perceptual states are accompanied by this ‘Gosh, this thing is right in front of me’ phenomenology, but we have no reason to think that all of them do. So I would not want to generalize from some special cases of perception to something about perception per se.

    I do have things to say elsewhere about how one could try to explain this ‘Gosh, this thing is right in front of me’ phenomenology in a representationalist framework (short story: the attributed properties are tropes). But as this is not a general feature of perception and the paper is about similarities and differences between perception per se and mental imagery per se, I’m not sure I need to address this here.

    I’d love to hear more about what you think of this strategy and of the general question about why relationalists take this phenomenology to be universal. Thanks again, Bence

  2. Hi Nick, thanks for these careful and excellent comments. I think the first point you raise is really important: the explanandum of the Dependency Thesis (visualizing) may not be the same as the explanandum I am interested in (mental imagery in general). I was always assuming that the Dependency Thesis can and should be extended to other, non-visualizing instances of mental imagery, but I think you’re right that this may not be so. So maybe the Dependency Thesis is not my direct competitor, which is, in some ways, a good thing.

    Your second point is that what I take to be an uncontroversial account of perceptual content is not so uncontroversial after all. I may have to concede this – but with a proviso. I do think it’s uncontroversial that perceptual representations attribute properties to objects. Now, how these properties are attributed is of course an open question. I say that these attributed properties are spatially (as opposed to syntactically or linguistically) organized. I also used the term non-propositional, but I am prepared to take this back, for the sake of those who have a very weak conception of what propositional content looks like.

    I do have to disagree with the third point though – although the point you raise is an important one and I hope to turn it to my own advantage. The problem you raise is that in my account it would be difficult to talk about the exercise of attention while having mental imagery (as in, shifting your attention from some visualized property to another) without somehow falling back to a version of the ‘special picture theory’, whereby I would need to posit some mental entity that is the object of attention. Now, I think this is a problem for all accounts of mental imagery. All accounts of mental imagery need to be able to explain how we can shift our attention from the color of the visualized apple to its shape and on the face of it, this sounds like it entails the existence of some mental entity that would serve as the object of our attention. But my account is that attending to a visualized property is just representing it (or attempting to represent it) with more determinacy. So there is no ‘special picture’ that our attention would sweep across. All we have is attributed properties, some more determinate, some less determinate. All attention does is to change this determinacy. So I would say that my account avoids the ‘special picture’ implications of the problem of attention in mental imagery that you drew attention to better than other views.

    Thanks again,

  3. Hi Bence, and thanks for your reply. I realize that you weren’t trying to criticize the Dependency Thesis directly, but you did offer evidence that it was less explanatorily powerful than the Similar Content View. As I mentioned at the end of my comments, I wasn’t convinced that those considerations were decisive, or even something a defender of the DT should be worried about given that the main consideration I raised seems to favor the DT with respect to explanatory power. Even if it turns out the SCV has an edge with respect to the points you mention, it seems worth considering other data points in this context.

    As far the phenomenological consideration I raised, I can understand why you’re suspicious of the ‘feeling of presence’ idea, since it’s somewhat vague and underspecified, though it is meant to be intuitive. And while I don’t have any new arguments for why phenomenal presence should be considered a feature of all perceptual experiences, I’m not sure why you think there’s no reason to think that it’s true given that others have tried to establish something like this point, e.g. Strawson in ‘Perception and its Objects’. It would be useful to have examples of perceptual experiences that don’t have this characteristic. Typical candidates such as afterimages and phosphenes have been nicely dealt with in a recent paper by Ian Phillips.

    In any case, it sounds like you’re willing to admit that there’s an important phenomenological difference between many cases of seeing and visualizing with respect to the way objects are experienced, and I gather that’s what you want to account for by claiming that we attribute properties as tropes in the relevant perceptual cases. I’d like to know more about what you’d propose for the visualization cases. One strategy, I’m guessing, would be to claim that we attribute universals in that case, but I’m not entirely sure what this would mean. Are universals perceivable properties? So if you’re willing to grant that explaining some kind of phenomenological difference with respect to objects in typical cases of seeing and visualizing is important, it would be helpful to learn more about how this account would go.

    Thanks again for an interesting paper!

  4. Bence, how would you characterize the experience of the subjects in the following experiment? Are they perceiving the triangle or just visualizing/imagining it?

    Seeing-More-Than-is-There (SMTT)

    If a narrow vertically oriented aperture in an otherwise occluding screen is fixated while a visual pattern is moved back and forth behind it, the entire pattern may be seen even though at any instant only a small fragment of the pattern is exposed within the aperture. This phenomenon of anorthoscopic perception was reported as long ago as 1862 by Zöllner. More recently, Parks (1965), McCloskey and Watkins (1978), and Shimojo and Richards (1986) have published work on this striking visual effect. McCloskey and Watkins introduced the term seeing-more-thanis-there to describe the phenomenon and I have adopted it in abbreviated form as SMTT. The following experiment was based on the SMTT paradigm (Trehub 1991).


    1. Subjects sit in front of an opaque screen having a long vertical slit with a very narrow width, as an aperture in the middle of the screen. Directly behind the slit is a computer screen, on which any kind of figure can be displayed and set in motion. A triangular-shaped figure in a contour with a width much longer than its height is displayed on the computer. Subjects fixate the center of the aperture and report that they see two tiny line segements, one above the other on the vertical meridian. This perception corresponds to the actual stimulus falling on the retinas (the veridical optical projection of the state of the world as it appears to the observer).

    2. The subject is given a control device which can set the triangle on the computer screen behind the aperture in horizontal reciprocating motion (horizontal oscillation) so that the triangle passes beyond the slit in a sequence of alternating directions. A clockwise turn of the controller increases the frequency of the horizontal oscillation. A counter-clockwise turn of the controller decreases the frequency of the oscillation. The subject starts the hidden triangle in motion and gradually increases its frequency of horizontal oscillation.


    As soon as the figure is in motion, subjects report that they see, near the bottom of the slit, a tiny line segment which remains stable, and another line segment in vertical oscillation above it. As subjects continue to increase the frequency of horizontal oscillation of the almost completely occluded figure there is a profound change in their experience of the visual stimulus.

    At an oscillation of ~ 2 cycles/sec (~ 250 ms/sweep), subjects report that they suddenly see a complete triangle moving horizontally back and forth instead of the vertically oscillating line segment they had previously seen. This perception of a complete triangle in horizontal motion is strikingly different from the tiny line segment oscillating up and down above a fixed line segment which is the real visual stimulus on the retinas.

    As subjects increase the frequency of oscillation of the hidden figure, they observe that the length of the base of the perceived triangle decreases while its height remains constant. Using the rate controller, the subject reports that he can enlarge or reduce the base of the triangle he sees, by turning the knob counterclockwise (slower) or clockwise (faster).

    3. The experimenter asks the subject to adjust the base of the perceived triangle so that the length of its base appears equal to its height.

    Results: As the experimenter varies the actual height of the hidden triangle, subjects successfully vary its oscillation rate to maintain approximate base-height equality, i. e. lowering its rate as its height increases, and increasing its rate as its height decreases.

    This experiment demonstrates that the human brain has internal mechanisms that can construct accurate analog representations of the external world. Notice that when the hidden figure oscillated at less than 2 cycles/sec, the observer experienced an event (the vertically oscillating line segment) that corresponded to the visible event on the plane of the opaque screen. But when the hidden figure oscillated at a rate greater than 2 cycles/sec., the observer experienced an internally constructed event (the horizontally oscillating triangle) that corresponded to the almost totally occluded event behind the screen. The experiment also demonstrates that the human brain has internal mechanisms that can accurately track relational properties of the external world in an analog fashion. Notice that the observer was able to maintain an approximately fixed one-to-one ratio of height to width of the perceived triangle as the height of the hidden triangle was independently varied by the experimenter.

    These and other empirical findings obtained by this experimental paradigm were predicted by the neuronal structure and dynamics of a putative brain system (the retinoid system) that was originally proposed to explain our basic phenomenal experience and adaptive behavior in 3D egocentric space (Trehub, 1991). It seems to me that these experimental findings provide conclusive evidence that the human brain does indeed construct phenomenal representations of the external world and that the detailed neuronal properties of the retinoid system can account for our conscious content.

  5. Hi Arnold,

    This is from Trehab 1991, right? It is a very nice illustration of cases where perception and mental imagery mix seemlessly – something that I take to be a very important feature of the relation between perception and mental imagery (see my stuff on amodal perception). I should use this experiment (at the moment I have the modified Perky experiment) for showing how my account can handle these cases nicely because of the similarity of content between mental imagery and perception. I have no idea how the defender of the Determinacy Thesis could explain these cases – James? Thanks again – it’s a really cool experiment.

  6. Hi James, thanks a lot for the follow up. I would really like to know more about what relationalists think about this experience of ‘wow, this cup is bang in front of me’, when we see a cup and why they think it’s the sine qua non of perceptual experience. As I said, I can totally see how sometimes this is true of our perception. Huxley wrote a whole book on this kind of experience (Doors of Perception), but he thought of these as special, very special indeed. Most of our perceptual experiences are rather less extraordinary (say, when I flip over the pancake, I experience none of this feeling of presence stuff…). Another thing I’m sometimes worried about is that this may be a visuocentric phenomenology – is it part of the relationalist view that this phenomenology is there in all sense modalities? I realize that we moved away from the paper’s topic, but as you brought this up, I would love to hear more. Thanks again,

  7. Dear Bence Nanay, thank you for the interesting paper.

    I have some questions of clarification which should be regarded just as efforts to better understand by a young student.
    Your proposal, as I take it, is a way of accounting for the phenomenology of perceptual or “quasi-perceptual” experience in terms of the experience content – applied in the specific case of a discussion about mental imagery. So, the similarity of content between perception and mental imagery should explain – or at least covary with – their similar phenomenology.
    If the nature of the content involved (a content consituted of perceivable properties) is the same in perception and imagination, this should account for the similar phenomenology. But I’m not sure whether you think that the phenomenology of perceptual experience and mental imagery are identical – such that, say, there is no phenomenal difference between perceiving a red apple and visualizing/imagining a red apple.
    If we stick to the idea of accounting for phenomenology in terms of content, then perceiving and imagining have the very same phenomenal character – they are phenomenally indistinguishable. So, the different source of the increased determinacy of the represented properties attention may succeed to realize makes a difference only at the level of our theorical explanation of the nature of the two kinds of experiences.
    Instead, if you aknowledge there ideed is a phenomenal difference – despite the undeniable phenomenal similarity –, may it be that the way perceptual content and the content of mental imagery change does play some role in determining the phenomenology of the two kinds of experience? Is it possible that the difference between “the dynamics of how the represented properties, and, importantly, the determinacy of the represented properties change in response to the allocation of attention” is somehow reflected in the phenomenal character of the two kinds of experience? Take the example of the house where I lived as a child: in the case of perception, as you write, “ I can just look: the exact shade of the curtain’s color is there in front of me to be seen”; by contrast, in the case of mental imagery, I attentively search for a more determinate representation of the curtain’s color that may or may not be present in my memory. Maybe the fact that the “visual search” attention performs is such a different process does make a phenomenal difference in the end.
    Thank you, and sorry if I just missed your central points.


  8. Hi Giulia, thanks a lot for this excellent question. So, I’m definitely not claiming that the phenomenology of perceiving x is the same as the phenomenology of having mental imagery of x. The aim is to give an account of the relation between perception and mental imagery that would explain both the similarities and the differences of their phenomenal character. I was focussing on the similarities here, but, you’re right about this, the differences are equally important. And some of the differences between the phenomenology of perception and mental imagery may well be along the lines you suggest. Even if perception and mental imagery has the exact same content, they may still have different phenomenology (even if you accept, as I am tempted to, intentionalism, ie, the view that phenomenology of perception in one sense modality supervenes on perceptual content in that sense modality). But as I understand your last paragraph, that’s exactly the route you’re following. Thanks again, Bence

  9. Hi Bence, your last comment made me realize that we’re not talking about the same phenomenon. I agree that the Huxley-esq “wow, there’s a cup!” is a special kind of experience that is far from the aims of your paper to need to explain. But that’s not what I was describing. What I have in mind is the fact that in the vast majority of cases, when I see an apple, it looks as though I could reach out and grab it, but when I visualize an apple, that experience is phenomenologically different in virtue of the fact that the visualized apple doesn’t look such that I could touch it. This seems to me to be a completely ordinary everyday phenomena, and it struck me that if your view is going to have an explanatory advantage over the Dependency Thesis, it’s one of the central features that you’ll need to account for. Your idea about attributing tropes to locations in the visual field might help here, but I’m not sure how the rest of the story is supposed to go. If what makes objects seem physically present is that tropes are attributed, then as I mentioned before, I’m imagining you’d say we attribute universals in the case of visualizing. But it’s not clear it what sense universals as such are perceivable properties, particularly if you accept that there are tropes. I hope this clarifies the point I was trying to make.

  10. Hi Bence,

    I’m glad you found my comments helpful. As I said, they’re not so much points of disagreement as requests for explanation, which you’ve given. I have a couple of follow-ups.

    On the point about propositional content and so on: I think saying that content *attributes* properties is uncontroversial, yes. Saying that it’s *constituted by* properties is perhaps less so, but again, maybe I have a more literal sense of constitution in mind than the one you intended.

    I suppose my main difficulty with talk in terms of propositional and non-propositional content is grasping quite what people intend by those terms. As you say, there are some very weak views out there; for example, you might think that at minimum, people mean that content is structured, but then there are views on which propositions are unstructured. So it’s not always clear to me what is being denied when content is described as `non-propositional’. You’ve clarified that nicely now: it’s a certain sort of organisation that’s being denied.

    On the point about attention: I absolutely agree that this issue of how to explain imagery without reference to dubious objects of experience is one faced by all views of mental imagery, and I’m sure that your view is no worse off, at least, than many competitors. No doubt you can characterise the experience in terms of attention without referring to such objects; I was just not sure how to do so. Again, your explanation helps. Here’s something to consider: suppose I visualise an apple, and then attempt to do so with a greater degree of determinacy, and succeed in doing so — its colour is now more determinate. Is there any interesting sense in which I am visualising the same apple as the one I started with? I’m tempted to say no (if we assume that I didn’t from the beginning have some particular apple in mind, like the one currently in my fruitbowl).

    Thanks again for the paper — plenty to think about.

  11. James wrote: “What I have in mind is the fact that in the vast majority of cases, when I see an apple, it looks as though I could reach out and grab it, but when I visualize an apple, that experience is phenomenologically different in virtue of the fact that the visualized apple doesn’t look such that I could touch it.”

    This is an important phenomenological difference between perception and visualization/imagery and, as James stresses, it has to be accounted for. Don’t we have to acknowledge that (a) for all instances of visual perception there is a retinal image that resembles the shape of the perception, and (b) the visual percept is located some distance in front of us in our egocentric phenomenal space? In my own experience, if my eyes are closed all mental images seem to be located somewhere behind my eyes, but if my eyes are open I am able to project an image onto a distant perceived surface or some other object in my visual world. So, for example, I’m able to imagine a cloud in the shape of a whale or imagine the shape of a whale on the wall in front of me. Yet it doesn’t feel as if the whale in either case is real. On the other hand, in the SMTT experiment that I described earlier, the moving triangle does feel like the *real thing out there in front of me* (reported by all other subjects that I tested) even though there is no retinal image that resembles a triangle. So the SMTT experience seems to fall somewhere between perception and mental imagery. Similar content doesn’t seem to work here. What are your thoughts about this , Bence?

  12. Theoreticians,

    Very Interesting Paper, responses, and discussions, Bence, Nick, James, Arnold…

    It seems most of you seem to be arguing or agreeing that visualizing can be very similar to, even indistinguishable from seeing. But, within my mind, when I imagine the redness quality of an apple, no matter how hard I try, the memory of such just doesn’t compare – even though there seems to be some less vivid similarities.

    Also, many of you seem to be saying that when you imagine or visualize an apple, it is not relational to you, or that you can’t reach out and grab it. Again, this seems very different to the way I imagine such. When I imagine an apple, it most definitely does seem physically present, right in front of my eyes, where I am imagining it, sometime even already holding it.

    I wonder how many people would agree that this is evidence that there is great amount of diversity in the way all of our minds work. Could it be that many of the divers theories are correct, according to everyone’s own mind, it’s just that we all have very different minds?


    Brent Allsop

  13. Hi Brent,

    Thanks for this. My own take on your questions, which may differ from that of others:

    (1) Though we’ve conducted this discussion mostly in terms of similarities, I wouldn’t agree that visualising can be `indistinguishable’ from seeing, and I’m not too happy with `very similiar’, even. In this, I line up with various philosophers in apparent disagreement with Bence and various other philosophers on his side.

    (2) Your second point about how imagining seems to you highlights one difficulty of investigating the subject: how to get a fix on the phenomena involved, when the experience is pretty much entirely private? One way is perhaps with creative experimental design; another way is with precise and detailed description of the phenomena as they appear to us. Bence, I think, is more inclined towards the first; I’m more inclined towards the second; perhaps in the end you can’t do without both.

    (3) I don’t think your diversity idea follows, though. In the first place, if all the theories in the game turned out to be true of some mental phenomenon or other, you might reckon that actually, they were each describing different phenomena, rather than being competing theories of the same thing. Second, it doesn’t follow from variation across people in capability or even phenomenology that their particular mental states demand different explanations from others. Take perception. Suppose I’m short-sighted and colour-blind, and you’re neither. We have different perceptual capabilities, and different phenomenology in similar circumstances, but we still share a basic sort of mental state which can be treated as an explanandum.



  14. Thanks Nick, and special thanks for making the thread more discussion-like with your last post.
    The point you raise in the earlier comment concerning the individuation of the object we visualize is really excellent and I’m not sure how to answer it. If you visualize your grandmother’s armchair, you can shift your attention from its color to its shape and this will not change the object I visualize. But if I visualize an apple (AN apple) and shift my attention from its shape to its color, I’m not sure it’s still the same object. What do you think? I agree with everything you said in response to Brent. Thanks again, Bence

  15. Brent, thanks a lot (and I’m flattered to be called a theoretician…), all of your three points raise very important issues.
    1. You are right: the question about the relation between mental imagery and perception is twofold: in what respects are they similar and in what respects are they different (there are, of course, many other questions about the relation between the two, eg, whether one presupposes the other, but I’ll leave these on the side). I have been focusing on the question about the similarities, but the one about the differences is as important.
    2. I very much agree and this will be the gist of my response to Arnold and James. We can visualize something in our own egocentric space and, even more so, we can have mental imagery of something in our egocentric space. But we can also have mental imagery of something in an ‘abstract space’. I’ll write more on this in my response to Arnold and James.
    3. There are really interesting studies on the interpersonal variation of mental imagery. The upshot is that there is a really disconcerting amount of variation. Ian Philips has a great paper on this and its philosophical implications forthcoming in the volume ‘New Waves in Philosophy of Mind’.
    Thanks again, Bence

  16. Arnold and James, I’m very happy that your comments converged, although I am not sure that you are really talking about the same phenomenon. So you are saying that while in the case of perception, we can just reach out and grab the object we see, in the case of mental imagery, we don’t have this kind of proximity to the object. I think that the ‘if I reach out I could just grab it’ thing is true of some but not all cases of perception and it’s also true of some but not all cases of mental imagery. It is not true of picture perception: you’ll never be able to grab the depicted object. But we sure see them – we see them in the picture. (you can object that this is not really seeing, and then we’re knee-deep in the depiction literature, which is really my indoor pool, so I’d be happy to talk about that. And some cases of mental imagery we do get this, if I could reach out, I could grab it (as evidenced by Brent’s comment, but also, remember, my concept of mental imagery is much broader than just closing your eyes and visualizing – it includes imagining an object you as having a different color than what it in fact has, it also includes hallucinations, etc.
    So the ‘reaching-out-grabbing’ phenomenology, I would say, cross-cuts the distinction between perception and mental imagery. I’m guessing that you would both deny this, so please do say more. Thanks again, Bence

  17. Bence, it seems to me that in distinguishing between visual perception, visual imagery, and visual hallucination, we have to take account of visual experiences with and without similar retinal events. (1) Normally, if we have a a visual perception of an object, we have a retinal image of the object that induces the perception somewhere in the space in front of us. Whether or not we feel that we can reach out and touch it, it looks like a real thing. (2) Normally, if we imagine an object in the absence of a similar retinal event, even if it seems so vivid in front of us that we feel we could reach out and touch it, we know it is not the real thing. (3) If we imagine an object, in the absence of a similar retinal event, that is so vivid in front of us that we believe it is the real thing we usually call it a hallucination.

    So while I agree with you that the ‘reaching-out-grabbing’ phenomenology cross-cuts the distinctions between perception, imagery, and hallucination, our sense of whether a visual experience is an experience of a real thing in front of us does distinguish between perception, normal imagery, and hallucination. The question I posed earlier concerned the status of the triangle that was experienced in the SMTT experiment. In this case all subjects experienced what they believed was a real triangle oscillating horizontally out there on the screen in front of them when, in fact, their retinal image was a dot oscillating vertically above a dot directly below. If this is a hallucination, it is unlike any other hallucination because the subjects’ conscious experience of a real triangle moving side-to-side in the space in front of them can be systematically controlled for size, shape, and rate of lateral oscillation by either the subject or the experimenter according to rules derived from a theoretical brain model. What would you call this kind of visual experience?

  18. Hi Bence, I’m willing to grant that the ‘reaching-out’grabbing’ phenomenology, as you call it, might cross-cut the distinction between perception and mental imagery. In particular, I think I’d grant that it would be there in some cases of hallucination. My claim was more restricted, and one that I thought you agreed with: that in the vast majority of ordinary cases, it’s clearly there in perception, and not in visualization. I think this is a central feature that allows us to intuitively differentiate the two states in ordinary cases. I’ve found it very helpful that you’re resisting making this difference constitutive of what differentiates the two states, because I think accounting for the potentiality phenomenological similarity between them introduces important complexity into the conceptual space. My worry, however, is that by making the fact that the two states could be phenomenally similar in the ways you discuss so central, you make it challenging to account for the difference in the ordinary cases.

  19. I should add that I think Arnold introduces a further interesting distinction in his last comment, between something like the phenomenology of an imagined object appearing to be located in egocentric space and the phenomenology of the imagined object appearing to be physically real. I’ve been running these two together in my comments, but I think the SMTT experiment shows nicely how they come apart.

  20. Arnold, sorry, I forgot to address this, thanks for the reminder. This experiment is a nice illustration of how mental imagery can be of objects in one’s egocentric space – in this case, it is, as I understand it, continuous with the perception (two dots are perceived and the triangle is represented by means of mental imagery). I agree that there are a lot of post-perceptual stuff that confuses things (whether I take myself to be perceiving, to be visualizing, etc – these can all have an impact on the overall phenomenology, of course). So to return to your last question, I would say that the triangle is represented by means of mental imagery – attribution of sensory properties to something that does not correspond to any retinal stimulation. In fact, it is quite similar in my view to amodal perception (or rather modal perception if you make this distinction) (I argue elsewhere that we represent occluded parts of perceived objects by means of mental imagery.

  21. James, yes, I certainly don’t want to deny that perception normally feels different from mental imagery – but some of these differences may arise from post-sensory/post-perceptual factors (as Guilia suggested above). So whether I take myself to be perceiving or visualizing has a massive impact on my overall phenomenology. This is consistent with my claims about the similarity of content. You are absolutely right that I should make it clear right at the beginning that I don’t deny this – there are similarities and there are differences. I think that starting with Perky may have given the wrong idea to the reader – I should change this.
    But to digress a bit (and to go back to the original point you made), and to go back to perception and the ‘feeling of presence’ stuff, do you think that relationalists are committed to having this ‘feeling of presence’ as a necessary feature of perceptual experience? I take this ‘feeling of presence’ theme to be different from the particularity of perception argument (that one is not about phenomenology). So I’d be curious how you see the relation between the two. Thanks again, Bence

  22. Bence, you wrote:

    “So to return to your last question, I would say that the triangle is represented by means of mental imagery – attribution of sensory properties to something that does not correspond to any retinal stimulation. In fact, it is quite similar in my view to amodal perception (or rather modal perception if you make this distinction) (I argue elsewhere that we represent occluded parts of perceived objects by means of mental imagery.”

    Here is what puzzles me. I can understand the amodal perception of occluded parts by mental imagery based on our past experience. But in such cases we are aware that the occluded part is not visible and is not really seen. In the SMTT experience, I don’t see how amodal perception can account for the phenomenal content because what we experience is a complete object that is really there in front of us. So it must be a modal perception. But if it is a modal perception how would you account for its creation, since there is no similar modal (visual) stimulus?

  23. Hi Bence,

    Thanks for an interesting paper!
    The term ‘attention’ plays a significant role in your account. But, as you acknowledge, this is not attention in the ordinary sense, since the ordinary sense involves getting (or attempting to get) new, more detailed information causally from the actual scene. So, ‘attention’ in the case of mental imagery designates some other phenomenon, which involves utilizing information from memory or expectation. This suggests that you think of attention as a generic phenomenon (covering both the perceptual and the imagery case). As you say above: “attending to a visualized property is just representing it (or attempting to represent it) with more determinacy.”
    But if attention is no more than this, then I suspect it also occurs in cases that involve no imagery and no perception at all, such as cases of judgment or desiring. For example, I right now judge, on the basis of memory, that my cat is on my gray sofa. I can make this judgment more accurate, by adding the specific shade of gray of the sofa to the judgment (from memory). Likewise with a desire: I form a desire for a red tomato, and I can make it more precise and form a desire for a tomato with a specific shade of red (from memory or expectation). These cases seem to exemplify your generic sense of ‘attention’. But, ex hypothesi, they are not cases of mental imagery, and they do no have phenomenal character similar to perception or mental imagery. So they seem to be counterexamples to your view.

    My point is, basically, that it feels problematic, given your theory, to define ‘attention’ as simply “representing (or attempting to represent) something with more determinacy”, because this sense of attention is too broad (or so it seems). It appears, then, that you need to find a narrower sense of attention that will still be broad enough to cover both perceptions and mental imagery, but won’t cover the aforementioned cases of judgment and desire. I have no doubt that you can answer this challange, and the answer might even be implicit in your paper. But I’m curious what it is.

    All the best,


  24. Hi Bence, I think that the phenomenology of presence and the particularity of perception are often related, e.g., the presence of an object in perception can be part of what makes it rational to have thoughts that are directed towards that particular object. But I also think that you can visualize a particular object or person without taking them to be physically present. I am intrigued by Alva’s view, defended in his recent book, that phenomenal presence extends to thought. I’m not sure whether his take on these issues would support your view or cut against it, but in any case if he is right about the phenomenology of the cases he discusses, it would be interesting to see how your view would accommodate that.

  25. Hi Bence,

    I can see you’ve got lots to respond to here, and I’ve nothing much to say (though plenty to think about) on the various other issues raised above. So I’ll just comment on your response to the case I raised, for now.

    “If you visualize your grandmother’s armchair, you can shift your attention from its color to its shape and this will not change the object I visualize. But if I visualize an apple (AN apple) and shift my attention from its shape to its color, I’m not sure it’s still the same object. What do you think?”

    The first bit sounds right to me; if you visualise some particular horse, changing the focus of attention doesn’t change its object. The second bit is the bit I’m not sure about either (it might be that there’s not even an answer). Perhaps we can ask the question in three different ways, and this might be part of what is confusing things:

    (1) is the intentional object the same in both apple cases?

    (2) is the content of the experience the same in both?

    (3) is the vehicle of content the same in both?

    Plausibly, one might answer differently to each of those: (1) indeterminate, (2) no, (3) yes. I’m not saying those are my considered answers (I need to think about it more), but the fact that we could consistently give different answers to each might partially explain why the case is puzzling.

    That’s all I’ve got for now, but I’ll be back if anything else occurs to me.



  26. Arnold, thanks – it is indeed different from representing the occluded parts of the cat behind the picket fence. If we make the distinction between modal and amodal completion (arguably both are cases of amodal perception (these labels are a bit confusing sometimes), like Singh 2004 does, we can say that amodal perception is the representation of an object behind an occlude, whereas modal completion is the representation of an object in front of inducers. So the Kanizsa triangle would be an example of modal completion. And I would describe your case in the same way as the Kanizsa triangle: the way we represent the sides of the Kanizsa triangle is the way we represent the sides of the triangle in your experiment. Ie, by means of mental imagery. If I’m missing something, could you point to a potential disanalogy between the two cases? Thanks again, Bence

  27. Assaf, thanks a lot – very important question. I only intend to talk here about visual attention – when I talk about attention as the increase of the determinacy of the represented property, this is supposed to be a story about perceptual attention. And the empirical studies that support this view all come from vision science. I don’t think we have any reason to think that attention is a monolithic category or that the word ‘attention’ picks out a natural kind. Attending to what we are doing (or doing things attentively, paying attention in class and visual attention are all important and interesting phenomena, but I don’t think we should treat them as the same phenomenon. In any case, my story about attention (ie, the increase of the determinacy of the represented property) is supposed to be a story about perceptual attention. I doubt it would apply for cases where we would also use the natural language category ‘attention’.

    The case you mention (ie, attending to specific aspects of a judgment or thought) is very interesting and some people, whose who take mental imagery to be a necessary feature of conscious thought could explain these cases using the argument I give in this paper about the role of attention in mental imagery. But of course not everyone agrees that mental imagery is necessary for conscious thought. And this debate is a bit of a nightmare (made even more difficult by the well-documented variation in the strength and frequency in people’s report of mental imagery – see Ian Philips’s paper I referred to in an earlier comment). So that’s why I didn’t say much about attending to different features of thoughts. Thanks again and let me know if you have a positive story about attending to different features of thoughts, Bence

  28. James, thanks, that’s helpful. I always thought that the particularity point would in principle apply to unconscious perception as well – there is nothing inherently consciousness-y about the general insight there. But the other contrast (the ‘feeling of presence’ stuff) is clearly only about consciousness. So that’s why I was wondering how the two are related. But what you say helps clearing this up. So would the relationalist camp still push the idea of the particularity of perception in the case of unconscious perception? I’m often unclear about what the relationalist would say about unconscious perception in general… Thanks again, Bence

  29. Nick, thanks again. I’m glad I’m not the only one who is uncertain about this… And it’s a nice disambiguation. I took the question to be about the ‘sensory individual’ (to borrow Jonathan Cohen’s term), ie, the entity properties are attributed to perceptually (or in the case of mental imagery, quasi-perceptually). This seems to be your option (a), and I have to agree that I have no idea how they are individuated in mental imagery (it’s also pretty tricky in the allegedly simpler case of perception, of course). If you have some further thoughts on this, let me know, thanks again, Bence

  30. Bence, you wrote: “And I would describe your case [SMTT] in the same way as the Kanizsa triangle: the way we represent the sides of the Kanizsa triangle is the way we represent the sides of the triangle in your experiment. Ie, by means of mental imagery. If I’m missing something, could you point to a potential disanalogy between the two cases?”

    Here’s the disanalogy: In the Kanizsa triangle we are presented with the three corners of a triangle and we complete the occluding triangle by imaging the three edges connecting the corners of the figure. So Kanizsa is a case of object completion. In SMTT there is no part of the visual stimulus that corresponds to the triangle that is vividly perceived and taken for a real triangle. In fact, in my SMTT experiment the oscillating dot is moving *up and down*, whereas the perceived triangle is moving laterally *back and forth*. The only way that I have been able to explain SMTT is by the neuronal structure and dynamics of the retinoid model.

    Here’s another puzzling phenomenon in perceptual imagery. Look at Figure 3 in this paper . Now what would be your explanation of why one part of a 2D picture should be perceived as sliding over another part of the same picture?

  31. Arnold, I thought that in the SMTT experiment there is a part of the visual stimulus that corresponds to the triangle that is vividly ‘perceived’ – the two dots. Which is of course much less than the three pacmen in the Kanizsa triangle (that’s why the SMTT is such an elegant study), but there is also the movement that helps the mental imagery. So I do think that there is a deep similarity between the two cases and I would still describe the way the triangle is represented as ‘by means of mental imagery’.
    In fact, this example shows nicely the diversity of mental imagery (which, as you remember, I define as the quasi-perceptual attribution of properties to sensory individuals that do not instantiate these properties).

  32. Thanks everyone for the great comments. I learned a lot. And huge thanks for Richard to put the whole thing together. Already looking forward to next year’s edition.

Comments are closed.