Friday, July 20, 2007

Making Sense of Dennett's Views on Introspection

Dan Dennett and I have something in common: We both say that people often go grossly wrong about even their own ongoing conscious experience (for my view, see here). Of course Dennett is one of the world's most eminent philosophers and I'm, well, not. But another difference is this: Dennett also often says (as I don't) that subjects can no more go wrong about their experience than a fiction writer can go wrong about his fictions (e.g., 1991, p. 81, 94) and that their reports about their experience are "incorrigible" in the sense that no one could ever be justified in believing them mistaken (e.g., 2002, p. 13-14).

But how can it be the case both that we often go grossly wrong in reporting our own experience and that we have nearly infallible authority about it? I recently pubished an essay articulating my puzzlement over this point (see also this earlier post) to which Dennett graciously replied (see pp. 253ff and 263ff here). Dennett's reply continued to puzzle me -- it didn't seem to me to address the basic inconsistency between saying that we are often wrong about our experience and saying that we are rarely wrong about it -- so I had a good long chat with him about it at the ASSC meeting in June.

I think I've finally settled on a view that makes sense of much (I don't think quite all) of what Dennett says on the topic, and which also is a view I can agree with. So I emailed him to see what he thought, and he endorsed my interpretation. (However, I don't really want to hold him to that, since he might change his mind with further reflection!)

The key idea is that there are two sorts of "seemings" in introspective reports about experience, which Dennett doesn't clearly distinguish in his work. The first sense corresponds to our judgments about our experience, and the second to what's in stream of experience behind those judgments. Over the first sort of "seeming" we have almost unchallengeable authority; over the second sort of seeming we have no special authority at all. Interpretations of Dennett that ascribe him the view that there are no facts about experience beyond what we're inclined to judge about our experience emphasize the first sense and disregard the second. Interpretations that treat Dennett as a simple skeptic about introspective reports emphasize the second sense and ignore the first. Both miss something important in his view.

Let me clarify this two-layer view with an example. People will often say about their visual experience that everything near the center has clearly defined shape, at any particular instant, and the periphery, where clarity starts to fade, begins fairly far out from the center -- say about 30 degrees. Both the falsity of this view and people's implicit commitment to it can be revealed by a simple experiment suggested by Dennett: Take a playing card from a deck of cards and hold it at arm's length off to the side. Keeping your eyes focused straight ahead, slowly rotate the card toward the center of your visual field, noting how close you need to bring it to determine its suit, color, and value. Most people are amazed at how close they have to bring it before they can see it clearly! (If a card is not handy, you can get similar results with a book cover.) Although this isn't the place for the full story, I believe the evidence suggests that visual experience is not, as most people seem to think, a fairly stable field flush with detail, hazy only at the periphery, but rather a fairly fuzzy field with a rapidly moving and very narrow focal center. We don't notice this fact because our attention is almost always at the focal center. (See section vi of this essay.)

Now when people say, "Everything is simultaneously clear and precisely defined in my visual field, except at the far periphery" there's a sense in which they are accurately expressing how things seem to them -- a sense in which, if they are sincere, they are inevitably right about their experience of things -- that's how things seem to be, to them! -- and also a sense in which they are quite wrong about their visual experience. When Dennett attributes subjects authority and incorrigibility about their experience, we should interpret him as meaning that they have authority and incorrigibility over how things seem to them in that first sense. When he says that people often get it wrong about their experience, we should interpret him as saying that they often err about their stream of experience in the second sense.

Dennett's view on these matters is complicated somewhat by his discussion of metaphor in his response to me, because metaphor itself seems to straddle between the authoritative (it's my metaphor, so it means just what I intend it to mean) and the fallible (metaphors can be objectively more or less apt), but this post is already overlong....

Update, February 28, 2012:
As time passes, I find myself less convinced that Dennett should endorse this interpretation of his view. Unfortunately, however, I can't yet swap in a better interpretation.


Anonymous said...


Nice post . . . as usual.

I wonder if we are typically greatly mistaken about the nature of our own perception when it comes to isolated time-slices, such as your example about what point in time the card becomes recognizable to us. Experiments like this get us at facts about the mechanics that are, in a sense, behind our conscious awareness – it seems to me. But getting at the mechanics of how it works does not get us to the final product, that ‘what it is like to be’. I think Nagel was right to say that this latter reality is an important kind of unity in time and through time. The unity of our perception through time does seem to be in some important sense “a fairly stable field flush with detail, hazy only at the periphery”, even though I would agree that it is “a fairly fuzzy field with a rapidly moving and very narrow focal center”. Perhaps the narrow focal center could be seen as a moving brush, constantly painting an every changing scenery, the visual part of a resulting what it is like to be.

If this is the case, we are mistaken primarily by taking the unity of consciousness as a simple given, instead of the final result of immensely complex workings of the unconsciousness mind. . . . and in this case the working of the conscious mind (?).

I would like to see more on how metaphor plays into this interesting discussion.

Anonymous said...

Hi Eric,

It seems to me that Dennett' view is not very problematic, and that might be put in the following way:

One can be wrong about what he sees, but can't be wrong about what he thinks he sees.

Justin Tiwald said...


Wow, this is very interesting! Sounds like a real breakthrough in Dennett interpretation.

I wonder how you'd characterize what goes wrong in cases of the first (fallible) sort of seeming. Here are a couple of options. (I'm sure there are others as well.)

1. The error is in misunderstanding which object the seeming pertains to. For example, I think my seeming is about the moment-by-moment experience of objects in my visual field, but in fact it's about how those seemings are reconstructed by my consciousness after a certain amount of time has passed (this could be Michael's nice point, but I'm not familiar enough with the "unity through time" language to be certain). By analogy, it's like when an apple "seems" green to me, when the apple is in fact a pear.

2. There is no mistake about the object. But a space for error opens up simply because it's a seeming—albeit a seeming of a higher order. That is, for an experience X, there's (a) how it seems to me, and (b) how it seems to seem to me. There is no confusion about what object (b) pertains to. But it, like any seeming, can be wrong—just in the same way that "the tower seems square to me" can be wrong about the tower.

Justin Tiwald said...

it's about how those seemings are reconstructed by my consciousness after a certain amount of time has passed

That should read "in my consciousness."

Anonymous said...

What the...?

Did Dennett explicitly endorse that there are two senses of "seems"? Or is that your way of putting what he did endorse?

Here's how I think one should, and a Dennettian would, describe the case of the peripheral playing cards:

It seems to the untutored observer that he or she has detailed information about the color and suit of peripherally presented cards. However...

And here's the crucial part: what should follow the "However"?

I say DD oughta say " reality one does not have detailed information about the color and suit of peripherally presented cards" and *not* say " another sense of 'seems' it does not seem to the untutored subject as though cards presented peripherally are determinately colored and suited."

Now, I'm pretty sure that DD is not anywhere near as big a fan as you are of the Chisholm/Jackson distinction between phenomenal and epistemic seemings, so I'll eat my hat or something if it turns out that he literally endorses this "two senses of 'seems'" stuff.

Say it ain't so, Eric!

Eric Schwitzgebel said...

Thanks for such thoughtful comments, folks! I'll respond to the first 4 comments tomorrow morning, but I just wanted to quickly reply to Pete's comment while my son's getting ready for bed!

Here's what happened: In conversation with Dennett at the ASSC, I was still confused. Afterwards, I gathered thoughts along the lines described here, and emailed those thoughts to him. Dennett said that I was right on the money. I explicitly asked him if it was okay to communicate that he endorsed this reading of him and he agreed (though I promised not to quote him, thinking that he might not want words expressed in an email posted into the blogosphere).

So these are not his words; they're mine. As far as I know, no sentence like "the word 'seems' has two senses" or "there are two ways of interpreting statements about experience" has come from his mouth or keyboard. But he has pretty explicitly communicated to me his approval of this view.

I hope I'm not misrepresenting him! I agree it's not the most obvious interpretation of his earlier work -- and maybe it represents a change or development of his view; but I do think the two-senses reading (or, I'm thinking now, maybe it would be better to say two ways of interpreting or approaching statements about "seemings" or experiences, rather than two senses of "seems") nicely explains the tension in his work between the incorrigibility/dictatorial authority stuff and the we-often-go-badly-wrong stuff.

Anonymous said...

Thanks for the clarification, Eric. In spite of my complaint, I do think you are on to something here. I try to expand on this over at my Brain Hammer blog:

Eric Schwitzgebel said...

Thanks for the comment, Michael! I was just thinking that I hadn't heard much from you lately.

I agree that the move from the mechanics of vision to the "what it's like to see" is not straightforward (though I also think looking at the mechanics can be illuminating and suggestive).

I'm not sure I fully understand the second half your post, though. There's definitely something appealing about the moving brush metaphor you use; but I don't see how you reconcile the apparently inconsistent descriptions of our experience as stable and broadly detailed vs. as having a rapidly moving narrowly focal center. Is your idea that the latter is a description only of the mechanics and not at all of the experience? Or...?

Eric Schwitzgebel said...

Tanasije, you put your point beautifully and succinctly! The view you describe is my own view (though I think one can occasionally be wrong about what one thinks one sees), and I'm interpreting Dennett as fairly close to that view.

However, the bit of nuance that your characterization misses, and which I think is present in Dennett, is that there is a sense in which what you say about what you think you see is itself a kind of seeming (maybe even something like a "notional world", to use an earlier idea of Dennett's) that is integrally related to "what it is like to be you" in some fundamental way.

Eric Schwitzgebel said...

Hi, Justin! I'm inclined to think one can go wrong in both the ways you describe, if I understand you correctly. "Seemings" isn't my favorite language, though, and that may be throwing me off a bit. (Yes, I know I used it in my post, borrowing from Dennett!)

Let me try to put your points in language more comfortable to me:

(1.) Your judgments about your experience can be mistaken because you might be asked to reach a judgment about your moment-by-moment experience, and report yourself as doing so, and really instead be giving a characterization that is more accurate of a later reconstruction of that experience.

(2.) Your judgments about your experience can be simply wrong -- not right about anything (not even the wrong thing, as in (1) above) -- you might think your visual experience is purplish when actually is it reddish and there is nothing purplish going on at all (except in your judgment about your experience), in much the same way you might think a coffee cup is purplish when actually it is reddish.


Eric Schwitzgebel said...

Thanks for the link and discussion, Pete. I recommend that Splintered Mind readers visit the Brain Hammer early and often!

Kevin Winters said...


I would highly suggest you read Sean Kelly's Seeing Things in Merleau-Ponty and get your hands on Alva Noe's Action in Perception. Though I would suggest getting your hands on Taylor Carman's Merleau-Pontian critique of Noe and O'Reagan's enactive theory of perception in the last section of his "On the Inescapability of Phenomenology" in Phenomenology and Philosophy of Mind, 67-89 (I probably wouldn't feel bad emailing you my copy if you send me a request). In that work, Carman also addresses Dennet's view, so it's certainly in line with this post.

Basically the argument is thus: seeing is much more than mere sensations rationally constructed into meaningful pictures through 'cognitive' structures. Rather, it includes our embodied motile grasp of the world, including our grasp (prise) of how to work with various objects. This motile aspect is why the world looks like it is rich in detail: because we possess the bodily skills to access things should we want to see better and, in most cases, our body acts spontaneously (i.e. without conscious control) to get such a better grasp if it is desired.

From everything else you've posted and the few papers I've read, I think this isn't too far from how you seem to be viewing things.

Kevin Winters said...

P.S. I just looked at Carman's website and apparently he has a forthcoming article titled "Dennett on Seeming." Interesting...

Eric Schwitzgebel said...

I've ready the Noe and Kelly, and I'm largely sympathetic with the idea that embodiment is important -- but I'm not sure that Noe and Kelly have established that, really. So much comes down to questionable introspective reports. (I criticize Noe and Kelly's descriptions of the phenomenology of visual perspective in an essay called "Do Things Look Flat?".) The two Taylor articles you mention sound very interesting. Thanks for the tip!

Anonymous said...

Hi Eric,

It sounds like you have Dennett endorsing a higher-order thought theory of consciousness! The way you put his view makes it sound just like the way that Rosenthal puts it in 'Content, Interpretation, and Consciousness' and 'First-Person Operationalism and Mental Taxonomy' this really what Dennett said he agreed to or did I miss something?

Eric Schwitzgebel said...

Hm -- that's an interesting comparison! I wasn't think of it as a HOT theory, but you're right there are similarities. One difference is that the authoritative judgments about experience aren't assumed always to attend the underlying experience. Also, I don't see Dennett as committed to anything like the view that the judgments are what make the underlying experience conscious.

Anonymous said...

Thanks for the response.

It seems to me that whay you have said does commit him to saying that the authoratative judgement makes the experience's why.

Grant that a conscious perceptual state is (whatever else we say about it) one that there is something that it is like for the subject that has it. Now, the judgment, if I understand your version of Dennett's view, is responsible for determining what it is like for me and so is respinsible for making the experience conscious.

Unless you are thinking that the fallible kind of seeming has some what it is likeness for the subject independent of the judgment? That would mean that you do have Dennett endorsing the distinction between qualitative seeming and epistemic seeming, which is suprising and, indeed, shocking!

Eric Schwitzgebel said...

I do think maybe it has to go in the direction you suggest in your last paragraph. I think that's shocking as an interpretation of Dennett if you emphasize his "authoritative" statements about our no more being able to be wrong about how things seem to us than Doyle could be wrong about the color of Holmes's easy chair.

Right now, I don't see any way to reconcile such statements with his equally prominent statements about the possibility, even the commonality, of error, other than to assume that there are two ways to understand our claims about how things seem and about "what it's like".

I'm not entirely comfortable with it either, I must confess. Suggestions welcomed!

Anonymous said...

I've read the post in 2012... five years later his pubblication... and I want to tele I see nothing wrong in dennett's view: see also Bowers interpretation of "introspection"... it is (referred to the "why?") necessarly a deduction... and like other kind of deduction is based on memory and deduction capacity... so people are more efficently to evaluate some connection than an "external" observer because they have memory of facts. They can go wrong , of course, but like inifnite kinds of deduction we know that the more information you have...the less you go wrong.

matteo , sociology student italy