Monday, October 29, 2007

Do You Mostly See Double? (Yet Again)

The more people I ask, the more people seem okay with the idea that most of what we see, most of the time (roughly, everything not in the region of optical focus), is double. Whoa! I get the heebie-jeebies. Either common sense is badly wrong -- if common sense is what I take it to be! -- or a substantial number of introspectors (including such eminent ones as Helmholtz and Titchener) are badly wrong. Grist for my skeptical mill. But not happy grist. Now I'm walking around thinking maybe I'm crazy for not seeing the persistent doubling that so many say is always there!

I take comfort, though, that Stephen Palmer, whose 1999 textbook Vision Science is generally considered standard in the field, analyzes the phenomenon much as I would:

One question that naturally arises from all this talk about disparity between the two retinal images is why don't we normally experience double images?... The answer has at least two parts. One is that points on or near the horopter [roughly, points at the same distance as the point on which you are focusing and on which your eyes are converging] are fused perceptually into a single experienced image. The region around the horopter within which such disparate images are perceptually fused is called Panum's fusional area. The second part of the answer is that for points that lie outside Panum's area, the disparity is normally experienced as depth. You can experience double images if you attend to disparity as "doubleness," however, or if the amount of disparity is great enough, as whenyou cross your eyes by focusing on your nose (p. 209).

Palmer seems to be saying that normally we don't see double, unless we attend to disparity as doubleness. He might then say -- as I would say -- that the reason so many people seem willing to attribute doubleness to their daily experience, when prompted to attend to double images created on the spot, is that they illegitimately infer that their normal visual experience is like their experience during such doubling exercises. But if so, that suggests an interesting instability and suggestibility among people in their judgments about ordinary visual experience!

Conversely, someone might argue against me, and against Palmer, that our experience when we think about doubling is our typical experience: We just ordinarily miss it in ordinary experience because we don't really think about or register the actual double-experiences we have of things off the horopter (or outside Panum's area) in the everyday run of life.


nathan said...

In the past I have used Coren, Ward, and Enns for my S&P classes. I just looked at it again briefly and it may not be quite as clear as I made it out to be. As I read it, though, it suggests that although we don't normally notice it, images from stimuli outside Panum's area are normally crossed. We don't notice this because we don't normally attend to it, but if/when we do attend to it, it is readily apparent. Whether this is C, W, & E's final analysis, it is the one I think I would lean toward.

In the passage you quote from Palmer he states that we can learn to notice the double images. But in my own experience, and in what seems to be the experience of my students, whatever learning is required is extremely easy and quick. That is, as soon as I/they am/are asked to attend to what's going on outside Panum's area, the doubling is there (as though it were there all the time just waiting to be noticed). I've tried to check this as I am walking around just doing stuff, too. That is, I've tried to check whether the doubling is there during my daily in-the-flow-of-life perceiving and it does seem to me to be there.

Next time I teach S&P, I'm going to use a text by Wolfe, Kluender, Levi, et al. They have a different take on this. They say that normally diplopia outside Panum's area is suppressed in the way that artificially induced binocular rivalry may be suppressed. So we are not normally consciously aware of doubling. However, like Palmer, they say that we can make the diplopia appear by shifting our attention.

So it seems to me that the question is what effect attention is having in this case. Put very roughly, the question is, "Does attention change the information that it focuses on or does it just illuminate it?"

Anonymous said...

Hi Eric,

I wonder if the wording "most of what we see, most of the time, is double", is the right one.

What are the conditions for seeing something? Say, that we are looking at some puzzle where we need to find hidden objects. Do we say that we see the hidden object just because it is there in the picture that we are looking at?

I take it that would be weird usage of "see". But it seems that kind of mistaken usage of "see" is in the statement "most of what we see, most of the time, is double".

The other point is that even in the case where while our eyes are focused on one object, we put attention on some object in the background, we can't say that the thing is doubled, just that under specific willful abstracting from the conditions, we can interpret it as seeming double. Because, I agree with Palmer, we will be inclined to interpret this as depth, and not doubling of the thing.

So, with this cleaning up of the language used, I don't think that there is a contradiction between common sense and those phenomena.

Eric Schwitzgebel said...

Thanks, Nathan, for your help and references on this! I agree with your analysis of this issue, though my own hunch and introspective judgments go the opposite direction from yours.

Those are nice points, Tanasije! On "seeing": I agree it's ambiguous, but I think that unless one accepts a very "thin" view of consciousness, there's a sense in which we "see" the background even if there's also a sense in which we don't see particular objects that remain unparsed, as it were. The lamppost in the background (or more accurately the visual experience it creates) remains part of your visual experience even when you're not thinking about the lamppost -- doesn't it? Then isn't there a fact about whether that background lamppost is visually experienced as double or not?

Or am I making some Cartesian-theatery / analogy-to-photography kind of mistake, in thinking that there must be a fact of the matter about this...?

I guess that really expresses my response to your second point, too.

Let me again express appreciation for your very helpful comments on this post and the last, Nathan and Tanasije. There's a lot to think about here!

nathan said...

I actually think there is an important challenge to Palmer's claim that diplopia is normally experienced as depth. For objects within Panum's area, we say that the brain computes depth by finding corresponding retinal points. So when, where, and how is the brain supposed to compute depth for images outside of Panum's area where the disparity between what would otherwise be corresponding points (i.e., if we were focusing our eyes and attention there) is great enough to disrupt normal fusion?

Eric Schwitzgebel said...

Interesting thought, Nathan! It seems to me that the "computation" would be fairly simple, though. Once you know the distance at which the eyes converge, it should be simple trigonometry to convert retinal disparities into distances.

Anonymous said...

Interesting. I'm totally unfamiliar with the literature, but I have an unusual perspective on this one. I've got a lazy eye with central occular suppression. When I was young I never noticed anything that I would characterize as “seeing-double,” although my left eye points slightly up and to the left of my right eye’s line of vision, I noticed some time around high school that if I was attentive, I could “bring out” my left eye’s doubled image of the object my right eye was focusing on, and then, if I wanted, move it around “inside” the right eye’s visual field by independently moving my left eye. The difference between the "double-vision" experience and the ordinary visual experience is pretty vast. Could it be that this is partially a language problem? When someone reports instances of "seeing-double," don't have in mind the specific phenomenon of having a doubled image of the object on which they are focusing? My guess is that most of the experiments that get people to observe "doubling" are really just experiments that split the focus of the two eyes between different objects, which, while it could bring out normally suppressed peripheral fuzz, has very litte overlap with what we would ordinarily refer to as "double vision." Just a thought.

Eric Schwitzgebel said...

Thanks, Blinn. Interesting report!

I myself incline to the view that doubling experiments do more or less what you say -- they cause us to attend, and thus experience, differently. But then that's exactly the point at issue: Are our experiences during such demonstrations really different from our ordinary daily experience? Proponents of doubling will say that they're not -- that the change in attention doesn't so much change the experience as cause us to realize that there's lots of doubling all the time!

Anonymous said...

Very interesting! My experience on this is that the dubble vision is a "rest mechanism". I practise yoga, and when I am deeply de-tensed i get dubble vision. The same when I am tired. Just a thought..

Eric Schwitzgebel said...

That's interesting, Emilie. I can imagine that going along with a de-focusing of the eyes....

Anonymous said...

I have double vision - but I can't move my eyes separately (well I haven't tried much...)

I guess everything is double to me, but on a day to day basis I pick the eye to use and don't see double.


Eric Schwitzgebel said...

Thanks, GNZ! Are you saying there's no point of convergence, even, where you see things single? If so, that's pretty different from what my students reported (not that there's anything wrong with that).

When you talk about the "day to day basis", do you mean to say that most of the time you're really depending primarily on one eye and you don't see double, and that the doubling occurs only when you make the effort to use both eyes equally? Or is your experience that everything is double all the time but you more or less ignore that doubling in favor of one eye? (Or is there -- as Dennett might say? -- really no difference between these two options?)