Tuesday, March 08, 2011

Oliver Sacks: Through One Eye, the World Looks Flat; Through Two, The World Has Depth

In his 2010 book, Oliver Sacks writes:

The only way to actually perceive depth -- to see it rather than judge it -- is with binocular stereoscopy (p. 114).
Thus, he reports on a woman ("Stereo Sue") who gains stereo vision and with it, for the first time in her life, real experience of three-dimensional visual depth. And of his own loss of vision in one eye he writes:
Stereo vision, however, now that I am mostly monocular, is quite compromised -- completely missing in the upper half or two-thirds of my visual field, though partly intact at the bottom, where I retain some peripheral vision. So I see the lower halves of people in stereoscopic depth, while their upper halves are completely flat and two-dimensional (p. 173).

With one eye occluded, I have no sensation of distance or depth whatever (p. 157).
Although Sacks acknowledges the possibility of some individual variation, he also objects to contemporary accounts of depth perception that treat binocularity as only one among several depth cues:
Such [multifactorial] views [of depth perception], while wholly consistent with a behavioral or empirical theory of vision, give no weight to the qualitative and subjective aspects of stereoscopy. Here one needs inside narratives, personal accounts of what it is like to suddenly gain stereo vision after a lifetime of stereo blindness... or to suddenly lose it after a lifetime of seeing in stereo [as in Sacks's own case] (p. 140-1, fn 14).
I wonder, though: Why is permanent loss or gain of stereoscopy necessary to the subjective evaluation? Can't the binocular among us at least temporarily mimic monocular experience simply by closing one eye? Does doing so make the world go from three-dimensional to flat?

It doesn't seem that way to me. Nor to most other people I've interviewed. Maybe depth is a little richer and more striking binocularly than monocularly -- but the difference is nothing so radical as the difference between 2-D and 3-D. Some of my interviewees, though, do characterize the monocular-binocular difference as 2-D vs. 3-D; and so do, for example, Ernst Mach (1886/1959) and Brian O'Shaughnessy (2003). Call those who report radically different monocular vs. binocular experience, like Sacks and O'Shaughnessy, the PuffOuters. Call those who report pretty similar monocular vs. binocular experience, like me (see also Chapter 2 of my recent book), the StaySamers.

Question: Is the difference between PuffOuters and StaySamers a real difference in experience? That is, does the world actually seem to puff out and go flat for PuffOuters, depending on whether it is seen monocularly or binocularly, while it varies little for the StaySamers? Or is the group difference mainly a difference in report only, with everyone having pretty much the same monocular vs. binocular experience -- and some people mistaken about or misdescribing their experience (in line with some of my conjectures in Chapter 2)?

We might start to address this question empirically as follows.

Sacks mentions that people differ substantially in how they experience three-dimensionality when presented with stereoscopic pictures -- that is, when presented with slightly offset pictures, one to each eye, producing a three-dimensional effect (like in contemporary 3D movies). For example, when presented with a stereoscopic picture of an impossible M.C. Escher tuning fork, Sacks reports one person saying it looked like the top prong rose about 3-4 centimeters from the plane of the page, while two others reported seeing the prong as 12 cm above the plane and Sacks himself saw it as 5 cm higher still (p. 134-5).

Conjecture: If the PuffOuter-StaySamer difference is a real difference in experience, the PuffOuters should, as a group, report much larger effects of that sort than the StaySamers. That could be tested empirically. Presumably PuffOuters should also better detect subtle stereoscopic effects -- which is even more appealing experimentally because it would permit catch trials to corroborate any differences in subjective report.

Yes, this is totally do-able.


djc said...

I've had a lot of experience with this, having been very short-sighted in one eye and then getting glasses when I was 10. I wrote about this in
The Conscious Mind, saying "things that had depth before somehow got deeper". The same seems to me to apply to the difference between monocular and binocular vision. So I suppose I'm like your Puffouters in that I think there's a big change in experience, but at the same time unlike them in that I don't think the difference is as radical as the difference between 2D and 3D.

I'm inclined to say that anyone who claims their monocular experience is 2D is misreporting it -- at best it is relatively 2D, but there's still a lot of depth phenomenology. (Prediction: find some way to impair other depth cues and they'll report "even flatter" experience.) I'm also inclined to say that anyone who claims that their experience is the same in the two cases is also misreporting it. (Prediction: have them walk around with one eye bandaged shut for a few days and then reopen it, and most of them will report a difference.) Of course I can't entirely rule out that different people have different phenomenologies here, or that the experience of short-term monocularity differs from that of long-term monocularity.

Eric Schwitzgebel said...

Thanks for the comment, Dave! Yes, some compromise position of the sort you mention seems pretty reasonable. Of course, there is also a tradition which says that even binocularly everything looks 2-D.

One issue to keep an eye on (or maybe two eyes): Too much an absence of depth cues starts to become a case of illusion, probably, which is importantly different than an appearance of flatness by which I am not at all fooled and which is instead just part of how things unproblematically look to me, either normally or under certain conditions. The importance of the difference between illusory shape and merely-apparent-but-non-illusory shape is important on views (like Noe's and Tye's dual-aspect views and also on traditional flat-sense-data views) on which experience has non-veridical but non-illusory perspectival properties; something similar could arise here, for certain opinions about the relationship between monocularity and depth. (This also touches on what I call the problem of known illusion -- e.g., the passenger-side mirror case -- which we have chatted about a bit.)

Jorgen said...

Maybe this is a bit off topic, but I am interested in the perception of 2D vs. 3D because, for some reason, whenever I view a 3D motion picture (even the really terrible 3D movies that you find in dinosaur museums), I am instantly thrown into an anxiety attack and feel like I am dreaming. I simply cannot view a 3D movie for more than 2 seconds or else it almost seems as if my mind starts glitching (so to speak, I don't really know any other way to explain it).

I should note also that, even though I don't have epilepsy or any such ailment (as far as I'm aware), I have a similar reaction to strobe lights, some florescent lights, and so forth. But 3D films *really* cause me some distress. I've not met anyone else who has experienced this.

Eric Schwitzgebel said...

How odd, Jorgen! I haven't heard of that before. Do you have the same reaction to still stereoscopes? Also, possibly some 3D movies exaggerate the effects -- maybe you're hypersensitive to that?

Jorgen said...

I have actually never looked at a stereoscopic image (at least not through the stereoscope) - but now I am interested in whether this would cause a similar result.

I don't think my reaction to 3D movies is only due to exaggeration, because I first discovered this phenomena when viewing a very basic 3D film at an Ancient Life Museum. And it begins with archaeologists digging in the dirt, and that's all I could handle. Literally 1 second of exposure to any 3D film causes about 2 minutes of a dream-feeling.

Another odd thing about this - namely, one eye producing less depth - is that my fiancee said she had to wear a patch over her eye for a couple of weeks when she had eye surgery, and that (as David presumed) there was never a noticeable transition from normal depth to less depth. But once the patch was removed, there was quite a drastic shift back to normal depth (something quite noticeable). Usually, it seems like the brain would adjust to make something like vision *more* accurate (e.g., whenever one loses a sense, they tend to have sharper senses after some time). But in this case, it appears that the loss of one eye yields no immediate results, but then adjusts after some time to give *less* accurate perspective. I'm confused about this. It seems that it would be vice versa (i.e., it seems that covering your eye should instantly cause the perception of something more 2D, but over time the brain would compensate and provide a more accurate perspective).

Eric Schwitzgebel said...

All very strange and underexplored!

Galen Mitchell said...

Assuming my last comment didn't go through, here's my late response (somewhat more organized now that I've thought about it for a bit).

It seems to me the entire issue might have something to do with the eye's ability to focus. When we investigate what actually happens when we look at something with both eyes open, we can see that our eyes give us two separate signals which our brains then interlace near the point of focus (the signals are made to overlap). Not only does this seem to point to our perception of depth having some sort of relationship to our ability to focus, it also complicates the puff-outer's side of the issue quite a bit.

To begin with, since this area near the point of focus is the only area that could possibly be affected by an individual's closing one eye, it would seem that the puff-outers would have to conclude that this area is the only area in our field of vision which we perceive to be in 3D. In which case, it's not that everything puffs out when both eyes are open, but rather just the thing being focused on. However, for the puff-outers it can't simply be an issue of "what we focus on is perceived as being 3D" because it is possible to focus on something to the extreme right or left in the field of vision which one eye cannot see (in which case, one eye tends to see a ghostly sort of image of the nose instead of what is being focused on). Given this, puff-outers would have to conclude that both eyes must be able to focus on something before it can be perceived in 3D, just as we would expect them to.

This conclusion has it's own problems. To start, no eye is exactly the same. One eye will inevitably have better vision than the other, and it is possible that one eye may not have the ability to focus nearly as well as the other. In which case, a puff-outer would have to conclude that there are individuals who cannot possibly perceive things in 3D due to massive variations in their eyes which prevent the "proper" interlacing of two focused signals. Issues relating to which eye is dominant etc will also complicate the issue for puff-outers.


Galen Mitchell said...


However, I happen to think it's a bit odd to say that the things to the extreme right or left in the field of vision are perceived as 2D. Looking around the room, I perceive no perceptible change when looking at something that can only be seen by one eye vs something that can be seen by both. Now, this could be simply because I'm a stay-samer to begin with, but I think there's a bit more to it.

As I keep saying, it seems to me that all this is an issue dealing with the eye's ability to focus. Given my own experience, when my vision is unfocused, it does not seem to have any depth to it—it's just a bunch of smudges of color and general blotchiness. And, given my own experience, it also seems as if my eyes focus more quickly and accurately when I have both eyes open. If this really is the case we might be able to explain why puff-outers feel as if they loose depth when they close one eye: they are experiencing a less focused visual field when they close one eye and as such report less, or even no, depth in their visual field. Likewise, we might find that stay-samers might focus faster, and more accurately with one eye open than the puff-outers do. In which case, we could see why a stay-samer would still claim that there is depth to their visual field.

While this is just a random hypothesis, it could be tested fairly easily. Take a number of individuals, test their vision, determine their dominant eye etc, and then have them attempt to focus on something with both eyes open and again with one eye closed. Assuming we were to measure the time it took for their eyes to focus along with their reported phenomenal experiences, we might find a correlation between those who report that their visual field puffs out when both eyes are open and those who have difficulty focusing when only one eye is open. In which case, we might have a physical explanation for why one person would say one thing while another would say something else.

Eric Schwitzgebel said...

Interesting thoughts, Galen! I wonder about one of your starting claims, though: Why think that it's only in the area of focus is the only area that could be affected by closing one eye? I'm also inclined to disagree about out-of-focus experience seeming flat. When I take off my glasses, everything is out of focus that's more than about 7 inches from my eyes, but the field still seems like it has depth.

Galen Mitchell said...

I suppose my reasoning around the area of focus issue is due to the fact that more often than not, it's the area of focus that benefits from having both eyes open. This is because we tend to (if not almost always) frame a scene so that the thing we are focusing on is in the center of our field of vision. The fact that the signals from our eyes overlap in that same area seems, at least to me, to point to the possibility that what puff-outers experience is an increase or decrease in focus depending on whether they have both eyes open or one eye closed. Granted, I was led to this "hypothesis" because, as a stay-samer, this was the only change I was able to detect when I closed one eye. Obviously, this limitation on my part could be somewhat problematic.

As for out of focus experience, you may be right. I don't wear glasses, and as a result I don't have much of an ability to switch back and forth between focused and out of focus vision. However, I was wondering if you at least perceive there to be less depth when your vision is out of focus? In which case, focus might have something to do with the issue, but not everything to do with the issue.

Of course, now that I think about it there's another potential hole in my little "hypothesis." It is possible that a puff-outer's perceiving the world as having depth has only to do with having two visual signals. Put another way, the sheer fact that there are two visual signals being processed by the brain of a puff-outer could cause their brain to process the information differently, and this difference could then translate into their having a different phenomenal experience when both eyes are open vs when one eye is closed. However, this seems extremely unlikely to me given the number of people who report to be stay-samers. It'd be a rather odd thing to encounter a person who's brain legitimately processed visual signals in a substantially different way.

In the end, I have to ask myself "where's the difference if it's not in focus?" Because a difference in the brain's processing of information seems somewhat unlikely, and I wouldn't even know how to identify it.

Eric Schwitzgebel said...

@ Galen: Yes, I do think many of the puff-outers think that it is the binocular disparity in the dual signal that creates the three-dimensionality of binocular experience, and dual signal is available for most of the visual field (except the far periphery on either side).

My own experience is that depth is just as vivid with and without glasses. It might even be more vivid for me without glasses for the idiosyncratic reason that my dominant eye is also the more nearsighted so that without glasses my normally less dominant eye becomes dominant, perhaps meaning that the work of the two eyes is more evenly balanced.

Galen Mitchell said...

Interesting stuff. Other than Oliver Sacks, are there other puff-outers I ought to read to learn more about the position? I'm just a lowly grad student, so I am admittedly not anywhere near an expert on these things.

Thanks for taking the time to respond, it's been enlightening.

Eric Schwitzgebel said...

You might want to look at Chapter Two of my 2011 book and the references in there. If you send me an email, I can send you the uncorrected page proofs of that chapter.

Steve Angle said...

Hi Eric, I thought I'd weigh in as someone with monocular vision. Closing one eye makes no difference to how things appear. 3-D movies are lost on me--I don't experience the difference (though things are blurry without the special glasses, so I have to wear them anyway). The optometrist's steroscopic tests don't work either. However, (1) I believe I experience the world in 3-D, and (2) I don't walk into trees. In fact I don't even ski into trees--manage the glades pretty well--but I did recently fly a toy helicopter into a tree. It was about 50 yards away from me, and I was trying to fly it around the far side of the tree and then bring it back to me. I simply couldn't tell whether it was beyond the tree or not, and ended up flying squarely into the tree. So: my experience is that there are lots of other cues that I process, leading me to experience things in 3-D (I think), though there are clearly limits imposed by the lack of binocular vision.

Congrats on the ANU offer, by the way!

musicalcolin said...

Two (partially) unrelated thoughts and a concession:

The concession: interesting topic, which I have ill-formed views on, despite continuing to return to it, I keep forgetting to explore the literature.

Thought 1: I agree with you that Sacks must be wrong. If binocularity gives depth perception, then why does monocularity (closing one eye) not immediately destroy depth perception? I know that their are claims that closing one eye for a few minutes won't destroy depth, but closing it for several hours will. If this is true, it suggests that there is a close relationship between binocularity and depth perception, but that it's not a direct causal relation such that without binocularity depth is not perceived.

Thought 2: I have a lazy eye (I've been told that it's very subtle, though it certainly isn't subtle to me), however, when I focus (e.g. for reading) I can get my eyes to converge correctly (both not focusing and focusing on objects too far away give me headaches). Converging my eyes (i.e. consciously focusing) is the only way that I can see 3D movies or video games (this is very tiring on my eyes). When I saw Avatar, if I didn't consciously focus, the movie looked 2D. Similarly, when I played the new Nintendo 3D hand held video game console (the 3DS) it would appear 2D to me unless I consciously focused. In support of the PuffOuters, I didn't realize that one of the consoles I was playing was supposed to be 3D until someone told me at which point I realized that I was not focusing on the screen with both eyes. Once I did, the images appeared 3D.

I don't know enough about the literature to conjecture as to whether or not 3D images are an adequate test of depth perception. Obviously, there are big differences between a 3D video game and the perceptual world. From my perspective (!), the biggest is that the world (as far as I can tell) always has depth (though focusing can create 'greater' depth), whereas video games and movies are totally without depth unless I focus. I'm not sure what the consequences (if any) of this are, but there it is.

Galen Mitchell said...

@musicalcolin I'm glad you mention the issue of focus. I was starting to feel like a complete idiot for suggesting that focus had something to do with it.

Bill Robinson said...

I've come across this rather late, but I hope not too late to suggest a simple experiment. To wit, look out a window and cover one eye with your hand. After about 20 seconds, uncover.

Result when I do this is as follows. When I cover, there's not much difference. Maybe a little, but i'd still say the houses I see *look* variously distant, i.e., that I'm not just judging them to be so.

But when I uncover, the restoration of "3D" vision is palpable and, for a moment, quite like the effect of a (still) stereoscope.

I don't know why there's this asymmetry between the taking away and the restoring, but that's my experience. For that reason, I don't find the PuffOuter/StaySamer distinction felicitous -- with me, at least, where I'd fall would depend on whether the comparison is made at the time of covering or at the time of uncovering.

Congrats on the ANU offer!

Eric Schwitzgebel said...

Thanks for the interesting comment, Bill!

I tried your experiment, and I share your sense that the monocular/binocular difference is more pronounced after one holds one eye closed for half a minute -- though I still don't think I would describe the difference as strongly as Sacks does. And I'm not entirely confident.

One possible explanation: Immediately after closing one's eye, one still has the binocular depth cues from immediately past input feeding one's visual experience, and the use of those cues slowly fades. The difference, then, is more pronounced between restoring after brief partial fading and restoring after long fading.

To test this idea, I tried another experiment: I closed one eye, went through my office door, looked down the hall, and opened that eye. This, too, should lead to a sudden enhancement of binocular cues and a more striking differential in depth experience than in my previous briefly-close-one-eye experiment. And indeed it did seem to me that the differential in depth experience between my original monocular view of the hall and my subsequent binocular view was greater than the differential between monocular and binocular views of the hall as a stood there afterwards opening and closing one eye.

So: Thanks again for the very interesting suggestion!

Unknown said...

Hi Eric,

Check out Koenderink's "Pictorial Relief" (Phil Trans. R. Soc. Lond. A [1998] 356, 1071-1086). There's lots of relevant stuff in it.

Eric Schwitzgebel said...

Interesting article, Robert. Thanks for the tip!