Sunday, December 7, 2014

The Phenomenology of Peripheral Vision (Part 2)

In the last post, I described an exercise that suggests many of us are wrong about the character of our own visual experiences. We tend to overestimate the breadth of foveal vision, even when we're reflecting without distraction on vision specifically while our eyes are open.

I find this result humbling.

I still accept a (very) weak infallibility thesis about present phenomenal experience. When we explicate beliefs about present experience, the phenomenal objects in question partially comprise those beliefs. It's not possible for me to believe that I'm having a red experience without having a red experience, else I'm referring to something besides "a red experience" when I utter the words "a red experience", and the mistake is purely linguistic. A lot of people disagree with me about that, but I'm not so sure it's worth arguing over. If it's true, it's trivially true, so it's not very helpful to know. If you're interested in this topic, I recommend the SEP article on Self-Knowledge pretty highly.

But why are we wrong? If the world is visually clear to us for only about two degrees of arc--the size of your thumbnail held at arm's length--why would we ever think otherwise, and why don't we notice our mistake before someone hits us over the head with it?

I think it's a combination of two things. 

The first is a sampling bias. If I ask myself "How much of the room can I see clearly at once?" the most natural way to find the answer is by looking at the room. Without successful fixation on a particular object, my eyes automatically move around without my conscious guidance. I imagine my brain is likely running through an abbreviated, non-conscious version of, "The wall is clear, what about the blender? Yep, that's clear. So's the lamp, and the cabinet, and the door. I can't see behind me, but pretty much everything in front of me is clear." When people know they're being asked about the phenomenology of peripheral vision, maybe they make up for that at least a little, but apparently not enough. 

It is weird to consider visual experience of things we're not looking at. Our brains evolved to to move our gaze toward the objects of our attention, and we do it so automatically that it takes a special effort to notice it happening.

The second is also a sampling bias. It's not just the case that I tend to move my gaze to the object of my attention. I also tend to keep my attention fairly narrowly focused on the object of my gaze, at least while I'm attending to vision. I bet you do to. Here, let me show you.

    1. Look at the "A" in the title "Agenty Duck". Keep your gaze fixed on that letter.
    2. Without moving your eyes, try to read the word directly below "Agenty". Can you feel your attention prying itself away from your gaze?
    3. Try moving your attention around, still looking at the A. Move your attention to different parts of the screen, then off of the screen and around the room.
    4. Ok, now bring your attention back to the A, joining vision and attention once again. 
    5. This time, keep your attention on the A, but move your eyes around the screen. Your attention wants to follow, doesn't it? Don't let it. See how quickly you can look around without losing attentive focus on the A.
    What does this mean for the phenomenology of peripheral and foveal vision? 

It means that we're primarily aware of that which we see most clearly. It is difficult to bring objects in peripheral vision into the focus of attention. One person who tried the above exercise couldn't make it past part two, because it was so uncomfortable to decouple visual and attentive focus. You have to be aware of something to notice it, so it takes effort, and possibly practice, to notice that an object ten degrees off center in your visual field is quite fuzzy and indistinct.

I'd performed the above exercise, and several like it, many times before encountering these questions of phenomenology of vision. As it happens, my initial estimation of the breadth of foveal vision was nearly right--three to five degrees of arc, instead of two. In the first instant I did feel a temptation to say something closer to thirty, but I successfully decoupled my attention from my visual gaze quickly enough that I barely noticed doing it. So my hypothesis is that I've trained myself to notice this kind of perceptual mistake upon reflection. I'm currently running a Tortoise Test on Facebook to see if others can do the same. I pre-commit to publishing the results, whatever they might be.

Results: I asked people "How many degrees of arc, would you say, are there at the center before things start going fuzzy in the periphery?". 16 people responded with straightforward numerical answers. 6 of them did the above vision/attention decoupling exercise before encountering the question. For those who did the exercise first, the average answer was 5 degrees of arc. The most common answer was 2, and answers ranged from 2 to 15. For the 10 who didn't do the exercise, the average was about 15 degrees, the median was also 15, and answers ranged from 7.5 (actually "5 to 10") to 35.

This supports both my explanations, though it doesn't distinguish between them. I could do that by having people do the same thing with just fixation and no decoupling, and then with decoupling but no fixation. I'm not sure how to do the second thing, unless I have them move attention to something besides location in the visual field (such as color, or even sound). I could also test the "training combats the illusion" hypothesis more directly by having people do these exercises once a day for three days, and then wait a week or two before asking them to estimate the breadth of foveal vision. Needless to say, I'd just like more data overall.

I concluded the last post by conceding that "Maybe the experiment itself modifies peripheral vision, rendering the foveal center artificially narrow, and people are actually correct in their beliefs about vision the whole time." I also said my priors were strongly against that, but I dismissed it too hastily. (This may have had to do with having spent much of the day reading anti-phenomenological infallibility articles.)

In fact, after writing the paragraph following this one, I now feel more than 50% sure that that's what's going on.

Time distortion may render the "illusion" true to the the phenomenology of peripheral vision even if experiments in reading and change blindness demonstrate that much of our peripheral visual experience is fabricated. Just as you might line up many pictures to make a completely in-focus, seamless panoramic, we might move our eyes to several parts of our visual field in succession, keeping the data from each saccade in memory, and then experience all that data with the same phenomenological time stamp. 

If that's so, an activity involving careful, extended visual fixation would create a poverty of in-focus data to piece together, thereby revealing the narrow range of ontological fovea compared to ordinary phenomenological fovea. 

Something like this happens when you tap your nose with your finger and experience pressure in your nose and finger simultaneously. It takes longer for a nerve impulse to travel to the sensory cortex from your fingertip than from your nose*, so the experience of simultaneity is evidence that the phenomenological "present" is a layering of recent memories.

In these past two posts, I set out to demonstrate and discuss a striking failure to hold accurate beliefs about our own ongoing visual experiences. At the end of it, I find I no longer agree with Dennett and his compatriots when they count this as a demonstration of immediate phenomenological error. 

Hopefully I'll still be able to make the points in my next post that I originally intended this to illustrate.


*I believe this because I once heard Sam Harris say it. He is a neurologist, after all, but I'm not entirely certain I trust him to ask himself whether the difference in signal arrival time is enough to be perceptible even if we don't layer recent memories to create present experience. And I don't know what the time difference is myself. Seems plausible, though. But perhaps more plausible to me than you due to my overt time dilation experiences under the influence of hypnosis and marijuana.

No comments: