Ian Phillips --- Lack of Imagination

[New Waves in Philosophy of Mind]

Lack of Imagination

Ian Phillips

Variation in the capacity for visual imagery is one of the most striking individual differences in psychology. At one end are super-imagers: subjects able easily to bring scenes before their minds’ eyes with the apparent richness and vivacity of normal vision; at the other end are non-imagers: subjects who entirely lack the capacity for visual imagination. Such variation is puzzling since non-imagers and super-imagers do not seem to exhibit substantially different levels of performance even in paradigmatic imagery tasks (e.g., mental rotation, scanning, or memory tasks). Theorists have historically responded to this puzzle in one of two ways. Some have concluded that there really are no substantial differences in imagery across individuals, only differences at the level of introspection or report. Others allow that such differences are genuine, but infer that imagery plays no useful function. In short, theorists have seen themselves as facing a stark choice between treating imagery as inscrutable or epiphenomenal.

I develop an account which avoids both inscrutability and epiphenomenalism. At its heart is a distinction between the conscious imagery enjoyed by imagers, and the non-conscious imagery shared by imagers and non-imagers alike, and put to use in their comparable task-performances. It is natural to object that this account deprives consciousness of any distinctive explanatory significance. I reply first that we may have to look beyond imagery tasks to appreciate the importance of conscious imagery. I further argue that, even if conscious imagery does lack explanatory significance, this does not imply that consciousness in general does. I show how, on one attractive account of the significance of conscious experience, perceptual consciousness has a significance which imagistic consciousness lacks.

A PDF of the paper is ready to view and download in the attachment below. 

A direct link to the PDF: http://goo.gl/sBQMg

[Declan Smithies]

Hi Ian, I very much enjoyed the paper and was pleased to discover the connections between your chapter and mine on issues concerning the epistemic role of consciousness. As you probably know, I take the internalist line that unconscious perception in blindsight and super-blindsight does not provide a source of immediate or non-inferential knowledge of the external world. So, I found it very interesting to think about whether it’s plausible for me to adopt an analogous line for the alleged “non-imagers”.

If we take the non-imagers at their word, and suppose that they do not experience conscious imagery at all, then I’d be inclined to push the analogies with super-blindsight. For instance, even if they give the right answer to questions about, say, the congruence of Shepard figures, they have no introspective access to the information that they’re using in answering these questions, so presumably it would seem to them as if they’re just guessing. Perhaps over time, they can become confident in their reliability in these sorts of tests, and so acquire a kind of inferentially mediated knowledge that is equally available to the veteran blindsighter. But even in that case, there would be no epistemic role for the unconscious imagery itself, as opposed to the background beliefs that they’re relying upon concerning their own reliability.

Having said this, I do feel some resistance to this taking this line in the case of the non-imagers, but not for the reason you give, i.e. that their performance in the relevant task is suitably reliable, but rather because I’m inclined towards pessimism about the reliability of these introspective reports for many of the reasons that Nigel Thomas gives in the post linked below:

http://www.imagery-imagination.com/non-im.htm

Moreover, I don’t share the view that there is any kind of incoherence involved in introspective unreliability – don’t the examples of blind people who think they can see and sighted people who think they’re blind put pressure on views that forge such a tight connection between consciousness and introspection?

It seems to me that the problems posed by non-imagers are similar to the problems that arise in connection with chicken sexing. It’s hard to deny that chicken-sexers know the difference between male and female chickens, although they’re unable to explain how they know the difference. One explanation is that they’re using unconscious perceptual cues, but this makes it hard to see how they’re in any better position to acquire non-inferential perceptual knowledge than the blindsighter or the super-blindsighter. An alternative explanation is that they’re using conscious perceptual cues, but they’re very bad at articulating them. This makes it easier to see how they can acquire non-inferential perceptual knowledge, since the reasons for their judgements are introspectively accessible, although they’re not very good at accessing them. Maybe something similar is true in the case of the non-imagers?

[Ian Phillips]

Hi Declan

Thanks so much for this! I'll certainly be looking to your work when I write the final paper up.

I do take the things that Thomas says very seriously: I doubt any philosopher knows more than him about imagery. However, I'm sceptical that this is all that is going on here, and that the kinds of diagnoses he offers account for all the variation in subjective reports. Two things do seem plausible. First, that people might deviate in their understanding of folk psychological language and so describe themselves differently (just as notoriously someone might wrongly think of themselves as having arthritis in the thigh). Second, that people (perhaps like Watson) can self-deceive about, or even actively suppress, their inner lives. But I don't think these factors can explain all the variation. So it needs to be that two people might really conceptualize the same inner life in very different ways for the diagnosis to be fully convincing. Yet I find it very hard to make sense of our inner life as having a lay out in this way independent of our conceptualization of it.

I agree that subjects can in certain cases (e.g. through self-deception or irrationality) fail to exploit their introspective knowledge/evidence and for a variety of reasons profess that their inner life is a way it isn't. But it seems to me an unattractive option to convict a substantial percentage of the population with no other obvious cognitive deficits or clear motivation of being in this predicament with respect to their mental imagery.

The analogy with chicken sexing is v interesting but accounting for the variation in imagery reports in that way would still require explaining why those who report weak/no imagery struggle to articulate why they are able to do imagery tasks whilst the majority of the population finds it so obvious it was their imagery.

Cheers

-- Ian

[Robert O'Shaughnessy]

Hi Ian

I tremendously enjoyed the paper and I found it provided much food for thought on  issues I am working on (how perception/simulation relates to belief/judgement). I definitely agree with you that it is unappealing to think people are often very mistaken about their inner lives as Schwitzgebel argues. But I was wondering what you thought about an alternative explanation that also makes people right about their inner lives.

I guess that I am probably a quasi-pictorialist when it comes to perceptual and imagistic experience. On this basis one can distinguish between thinking imagistically and the richness and vividness with which the detail in the quasi-pictures are ‘filled in’.

Consider blindfold chess: the very best players can (miraculously!) simultaneously play between 10 and 30 players whilst blindfolded. It seems plausible that to achieve this feat they must be able to imagistically (i.e. isomorphically to the spatial layout in the world) set out the board and what squares the pieces are currently on for each game. But what they would not need to do is render the images of the board and the pieces in any great detail (e.g. faithfully reproduce the little crenelations at the top of a rook). It could be that in the limiting cases what was laid out were something as minimal as dots (in their proper relative locations) that the system was able to interpret. In these cases the player might report no imagery.

According to the quasi-pictorialist, despite there being very little imagery, this would still count as imagistic thinking, as opposed to propositional thinking (the image would simultaneously enact many different propositions). And of course the player could be fully conscious of their reasons for moves (e.g. “because I could see that the piece was undefended”).

The same would go for tasks like rotating Shepard figures or counting the windows at the back of your house. To do the task well one must be able to layout the components according to how they would be in space (e.g. a few dots at salient points of the Shepard figure). But one would not need to fill in all the detail such as line edges or bricks. Again one might report having very little or even no imagery in such cases but still be fully conscious of the spatial layout one uses as evidence for one’s judgements.

This alternative would be confounded if there were cases of tasks that required high levels of detail in the imagery to succeed and people reported experiencing little or no detail. For example Stromayer 70 reports on ‘Elizabeth’ who supposedly could reproduce the detail in stereograms in imagination a day after having seen them. High levels of detail might also be required to see the alternative view in the duck/rabbit figure just from imagining it. I’m afraid that I don’t know the literature well enough to know if there are any reports of non-imagers succeeding in tasks which require detailed imagery. If there were that would support your view.

In any event, nothing I say contradicts your view but it might be another alternative worth putting on the table!

[Ian Phillips]

Hi Robert

Thanks a lot for your comments. You're quite right that determinacy of content is clearly relevant to these issues, and so we need tasks that intuitively benefit from more determinate content to even expect a correlation between reported imagery and performance. I'll have to go back and see if there are clear examples of such tasks. But in general I simply don't think enough work has been done to settle all the possible explanations (or combinations of explanations) of this kind (including for example the hypothesis I mention that it's motor *or* visual imagery which correlate with task performance). So my paper was really only trying to make a limited conditional move: if it's true -- as many think -- that reports and performance (in intuitively relevant tasks) aren't related, we still need not reject the reports or deny the existence/function of imagery.

To my knowledge, the case of Stromeyer's (later wife) Elizabeth which you mention, whilst fascinating, is both unique and scientifically v dubious. Thomas, as always, has a helpful note here: http://plato.stanford.edu/entries/mental-imagery/quasi-perceptual.html

Thanks again

-- Ian

[Bence]

Hi Ian, thanks for the excellent paper – I am afraid, I have very little to disagree with here. I would worry more about people’s reluctance to give up the assumption that mental imagery is necessarily conscious. Maybe it would be useful to provide some kind of big picture motivation for this stance – something along the lines of ‘Perception can be conscious or unconscious. Given the various similarities between perception and mental imagery, on the face of it, mental imagery can also be conscious or unconscious’. What do you think about this strategy (just for showing that it’s not a crazy claim)?

 

I give an independent argument for the claim that mental imagery can be (and is often is) unconscious in my Amodal perception paper (2010), so your project here and my project there would tie in nicely. I also think that Neil van Leeuwen’s concept of mental imagery (in his 2011) is not necessarily conscious – so there may be some momentum in that direction…

 

ps. A potentially very misleading typo on p. 17, first line: I assume what you want to say is that you do NOT want to endorse this view here. 

[Liz Irvine]

Hi Ian,

Thanks for your paper - I really enjoyed it! A few random comments/questions...

So clearly everyone has imagery in the sense that they can perform various imagery tasks better or worse (they have it in the functional sense). But some people report being imagers, and some don't. 

- The assumption seems to be that reports of imagery have to have something to do with performing imagery tasks. Otherwise there would be no puzzle. But I guess I don't necessarily see the puzzle. Are we sure they're tapping into the same processes? Is it just a good guess that they do? Maybe the differences between reports and performance suggests that we really shouldn't treat them as measures of the same thing in the first place?

- One of the questions is whether the ability to report imagery has a functional role, right? I just don't see why it must have a functional role (or at least a really obvious one, it might happen to help out in unexpected ways, as you suggest). Maybe this just ties into your comments that what we say about the ability to report imagery doesn't affect the potential functional role of consciousness in general. The ability to report things in general clearly has a function, and particular cognitive processes have functional roles, but if a conjunction of the two doesn't have a function, this doesn't seem problematic to me. 

Cheers 

Liz

[Justin Fisher]

Hi Ian,

I enjoyed the paper, but wanted to raise a potential worry.

On page 10, you quote Dennett as saying that his scientific approach "proceeds by defining mental images as the normal causes of these β-manifolds" and you even note his emphasis on "normal".

However, on the next page you say:

Non-imagers do not form the beliefs characteristic of Dennett’s β-manifolds.  As a result, such subjects appear to lack imagery in the scientific sense: they have no β-manifold concerning their imagery, and imagery in the scientific sense is by definition the cause of such a manifold. 

Notice that you left out the crucial "normal" here.  And it really makes a difference.

Here's an analogy.  Feet are the normal cause of footprints.  But the mere fact that  there are no footprints in my house does not entail that there are no feet in my house.  Similarly, Dennett defines imagery as the normal cause of B-manifolds, but the mere fact that there's no B-manifold in someone's head does not therefore entail that there's no imagery there.

None of this is to suggest that the view you defend is wrong, just that it may not be all that different from Dennett's after all.

-Justin

[Ian Phillips]

Dear Liz

Thanks for v helpful questions.

-- I agree we might think of reports of imagery as having nothing to do with the imagery involved in imagery tasks. But I think this is pretty revisionary and I'd prefer to be as unrevisionary as possible. I say that in large part because people who have imagery typically say that they solved the relevant tasks using their imagery. Maybe we should give up this piece of our ordinary psychological self-conception, but I think we should wait until we're forced to.

-- I absolutely agree with the thrust of your second set of comments. For example, making a piano transparent would allow you to report on the mechanism which produces the sounds, but need that be of any use to the pianist? My main aim at the end of the paper was to show how we can explain the differential in significance between imagistic consciousness and perceptual consciousness given a specific claim about what the significance in the perceptual case is. But, at least insofar as we can think of the difference between imagistic consciousness and perceptual consciousness, as the difference between imagistic reportability and perceptual reportability, I'm absolutely in agreement with your general point. As you nicely put it, just because the ability to report in general has a function and a particular cognitive mechanism does, it doesn't follow the conjunction does.

Cheers

-- Ian

[Ian Phillips]

Dear Justin

Thanks for raising this worry, which you're also quite right in seeing I'd already recognized but then hesitated over. I entirely agree about the difference "normal" might make, and I'm happy to simply defend a Dennettian line if that's what it is. No shame in that to my mind! 

Cheers

-- Ian

[Ian Phillips]

Hi Bence

Thanks v much for your comments! Not to mention the typo correction -- always grateful not to completely mislead my reader...

I totally agree that we might motivate unconscious imagery on analogy with unconscious perception. (I suppose one can imagine distinctive forms of resistance in the imagination case, e.g. due to the active nature of imagining, but the basic motivation seems compelling.) However, I was mainly aiming to address the concern not with the existence of unconscious imagery but the idea that distinctively conscious imagery has no obvious explanatory work to do. Only someone who thinks that distinctively conscious perception has no explanatory role to play can motivate that on analogy with perception, and I think that is a much less obvious view about perception.

I just looked again at your amodal perception paper. Since the conscious/unconscious thing wasn't central to it can I just check if you thought of the argument there for unconscious imagery roughly as follows? (1) When we represent parts of objects that are not visible to us, this representation involves mental imagery; (2) When we represent parts of objects that are not visible to us, this representation is often unconscious; (3) So: mental imagery is often unconscious.

Cheers

-- Ian

[Mark Sprevak]

Hi Ian!!

I really enjoyed your paper!! I'm very sympathetic with the line that you develop. 

Just a couple of friendly comments.

I guess I think that our conscious experience is only a very fine dust across a very thin slice of our cognitive life, so it wouldn't be too surprising if for sometimes that dust doesn't fall on our off-line manipulations of visual data. 

-- "Unconscious imagery" sounds like a funny thing to posit because there seems to be something like a conversational implicature that "imagery" must involve conscious imagery (it sounds a bit like the “unconscious conscious”). I was wondering if it could be made more palatable by saying that the relevant posit is just manipulation of visual data---this is sometimes that if done unconsciously doesn't sound odd at all. The imagers just get the extra dust of conscious experience and reportability on top.

-- I agree with Liz that there doesn’t seem a need for conscious report of visual imagery to have a function at all.