In an earlier comment, I indicated that I would post something on my blog to help clarify the distinction between direct and representationalist perception.
I have now done that, in a series of four posts. The first of those is
and it, in turn, contains links to the other posts.
Your turn.
If it can be explained with inference, and if it is repeatable, then it can be explained without inference.
That’s what you seem to be missing.
The part I don’t understand is that retinas can suppress or enhance visual stimuli based on the nature of the stimulus. II would like to call that immediate modelling.
For example, frog retinas can model a meal-ready fly and pass the stimulus to the parts of the brain responsible for feeding.
The same data presented at a different rate or different pattern causes no firing of retinal receptors. I’d want to call that modelling.
Humans are not entirely without this kind of processing.
Neil,
If you have an argument to that effect, then present it.
You’ve asserted that the motion illusions I’ve cited are “fully consistent with direct perception”, meaning that they need not involve inference.
Show us. Explain the last illusion, in which the red ball appears to move behind the green square, in terms of direct perception.
I’ve been asking and asking.
petrushka,
The line between direct and indirect perception can be a bit fuzzy. For example, even direct perceptionists agree that perception involves representation, in the sense that patterns of neural firings represent something in the outside world. The disagreement is over how tightly the representation is tied to the incoming stimulus.
I argue that it’s loosely tied in the case of the motion illusions, because the representations are of moving objects, while there is no motion at all in the actual stimulus.
Here’s the problem I see.
There’s an assumption that a photon can cause the firing of a retinal receptor. I suppose one could have a philosophical discussion of whether the photon is real or not. Or that there must be something real to produce or reflect the photon.
But that is not the problem I’m thinking about.
My problem is that the retina receptor can “choose” not to fire, based on some elementary modeling of what is worthy of being passed on.
In humans there are perceptions of color that take place at this forward position in the chain of perception. I don’t see any point in the system that is free of modeling.
What’s to explain?
If you can get one form of information, and then compute a second, you can also get the second form of information directly by just changing the hardware used to get the information. In practice it is a tradeoff between the one-time cost of the hardware and the repeated cost of the computation.
I have no idea what you are looking for there.
Gibson clearly distinguished between perception and sensation. You seem to want an explanation of sensation.
petrushka,
My understanding is that the photoreceptors themselves — the rods and cones — always fire, but that there are higher-level cells that can be stimulated or inhibited in more complicated ways.
Of course, whether this counts as “modeling” depends on the particular definition of modeling you are using.
I’ve been stressing the motion illusions (particularly the last one) because I think they unambiguously show that inference is happening.
Then you’re diametrically opposed to Neil, who sees the entire perceptual process as being free of modeling, inference, or representation. In his view, any modeling/inference/representation that occurs happens only after the perceptual process is complete.
It’s an indefensible view, in my opinion, which may explain why Neil has been unable to defend it.
Neil:
If A collects the information, and B computes something from it, how is that different from a scenario in which A collects the information and computes that same thing from it? The process is still the same: collect the information, then compute something from it. You’re doing nothing but relabeling.
keiths:
Neil:
I’m looking for exactly what I asked for:
Neil:
No. Here’s what I want:
Neil,
That is your statement. You typed it in and clicked ‘Post Comment’. You are responsible for it.
If this illusion is “fully consistent with direct perception”, then show us by explaining it in terms of direct perception.
keiths:
Neil:
keiths:
Neil:
I received Gibson’s last book (The Ecological Approach to Visual Perception) this weekend through an interlibrary loan. I wanted to read about direct perception in Gibson’s own words.
On the topic of saccades, his view is clear. As I suspected, he does not doubt the idea that the eye focuses images on the retina. He just doubts that the purpose of saccades is to bring a stimulus from the periphery of the retina to the fovea:
The quote you give seems completely unrelated to the issue.
Neil,
‘Saccadic movement’ is unrelated to ‘saccades’?
I’ve been asked to continue this discussion here.
In an earlier message (just a little above here), keiths quoted J.J. Gibson, and said that the Gibson quote supported a representationalist position.
Here’s my paraphrase of Gibson (abbreviated version):
That keiths would claim the quote supports a representationalist position is stunning. Since keiths presumably knows that Gibson is a critic of representationalism, it is surprising that he would not question his own reading of that quote.
I implicitly questioned his use of that quote with: “The quote you give seems completely unrelated to the issue.” And the only response I got was that the quote was about saccades.
This came up again today. And again, keiths response is that the quote is about saccades.
At times, it is almost as if I am debating an AI bot in the style of Eliza.
Neil,
Here’s our full exchange, from just above in the thread:
keiths:
Neil:
keiths:
Neil, failing to take responsibility:
keiths:
You misunderstood Gibson, Neil.
I’ll wait for your paraphrase of what you believe Gibson was saying.
Neil,
Gibson is saying that the purpose of saccades is not to bring peripheral stimuli to the fovea.
He’s not saying that our eyes don’t focus images on the retina. If that were the case, everything would be blurry! Why do you think some of us need glasses, if not to help our eyes focus images on the retina? Why would we have focusing muscles in our eyes if we didn’t focus images on our retinas?
Strictly speaking, no. He is just disagreeing with the idea that the purpose is to bring them to the fovea. He does not say that there is a purpose to not bring them to the fovea.
In that quote, he does not say anything about focusing images on the retina. So your assertion “As I suspected, he does not doubt the idea that the eye focuses images on the retina” is not supported by that quote. Rather, that quote you gave is silent on the issue.
That’s what you say. It is not what Gibson said in that quote.
For your information, direct perception proponents who say that the eye does not focus an image on the retina, are not in any way suggesting that the eye is producing an unfocused image on the retina. When they say that the eye does not focus an image on the retina, their disagreement is with “image”, not with “focus.” Your argument misses the point.
Neil,
Yes, that’s what I’m saying. The purpose of saccades isn’t to bring peripheral stimuli to the fovea, according to Gibson — it’s something else.
keiths:
Neil:
I selected that quote because it describes Gibson’s views on saccades.
Remember, the question was whether
…as you claimed.
He didn’t. He simply thought that the function of saccades wasn’t to bring peripheral stimuli to the fovea.
Ah, now I see the source of your confusion. Direct perceptionists don’t disagree that the eye focuses images on the retina. ‘Image’ is a standard term in optics for the in-focus pattern produced by a camera, eye, telescope, microscope, etc.
What direct perceptionists disagree with is the idea that the visual system processes successive static “snapshots” of the retinal image.
Here’s how Michaels and Carello put it in their criticism of the indirect approach:
That is utterly absurd.
Gibson worked on perceptual psychology, including vision. You pick one brief statement, and somehow conclude that is the sum total of his views on saccades.
That one statement was simply his debunking of the usual representationalist position on saccades. That’s all it was.
You are just making up stuff.
Neil,
It’s your claim:
If you don’t like my Gibson quote, you’re free to provide one that supports your claim.
I think you’ll have a hard time finding one, because Gibson understood, as most educated people do, that the eye focuses images on the retina. His beef was not with that obvious fact, but rather with the idea that the visual system analyzes successive static snapshots of the retinal image.
keiths,
There’s a difference between:
“The eye focuses an image on the retina”
and
“The way that the eye focuses light results in an apparent image on the retina”.
The first of those has an implication that the eye is carrying out a purpose of producing an image. The second does not have that implication of purpose.
I disagree with the first because of the implication of purpose. I agree with the second version.
Neil,
The eye needs to focus images on the retina in order to work properly. When the eye loses that ability, we give people glasses, or contacts, or laser surgery to restore it.
Gibson knew this and didn’t object to it. His complaint was against the idea that the visual system works on successive, static “snapshots” of the retinal image.
He certainly didn’t think that saccades were evidence against the focused image idea.
No, it doesn’t. Whether or not there is an image is of no relevance to the eye.
The eye needs to get the best information it can about the external world, and it needs to adjust the focus of the lens to maximize the amount of such information. That this can be said to produce a focused image, is merely a side effect.
The apparent image would not look all that focused to us. If we were looking at the image, we would see a lot of chromatic and spherical aberration. In a camera, where the image is important, we design the lens to avoid chromatic and spherical aberration to the extent possible. But that does not matter to the eye, because the image is a mere side effect.
If you are going to repeatedly claim this, then you need to provide a citation.
Saying it once could be just an incidental remark. But repeating that assertion about what Gibson thought does make it a claim that requires evidence. Either stop repeating that claim, or provide a citation. The quote that you have already given does not address that point at all.
Neil,
If the image isn’t focused correctly on the retina, the visual system won’t work properly.
If you don’t understand this, I would suggest consulting some books on optics and eye physiology.
This seems to be a clear example of you missing the subtleties of language.
Neil,
See this, and note the repeated use of the word ‘image’.
There are many other references just like that. If you want to argue that the rest of the world is wrong,and you’re right, then I’m not interested in participating.
That’s not actually relevant. We were not discussing lasik surgery.
I repeat:
keiths,
You are insisting on using language that puts the forming of a retinal image in the causal chain for vision. But that is part of what is in dispute between direct and indirect perception.
If you are going to insist on that, then there is no discussion.
Neil,
Please reread this comment.
keiths,
You are only playing rhetorical games.
Neil,
I’m disputing your claim. The customary next step would be for you to provide evidence in support of your claim.
You said:
I looked through Gibson’s last book and found no evidence for that. I think you misunderstood him.
You’re welcome to demonstrate otherwise by providing a supporting Gibson quote. However, as I said before:
That was never a claim. It was an off-the-cuff remark.
I made that off the cuff remark in response to your earlier claim:
That’s a claim you keep repeating. You have not provided any evidence at all in support of that claim.
You have, instead, been playing rhetorical games by trying to shift the burden of evidence to me.
I see.
So if I say “direct perceptionists think X”, then that is a claim and I had better supply some evidence, pronto.
But if you say “Gibson thought Y”, then that is not a claim, it’s merely an “off-the-cuff remark”, and you are in no way responsible for backing it up.
Perfectly consistent, perfectly rational. 🙂
See you around, Neil.
Right.
Correct. That’s because nothing depends on it. I used that only to express skepticism of your claim (the one for which you have not provided any evidence). As Hitchens has said: “That which can be asserted without evidence, can be dismissed without evidence.”
All you are doing, is trying to shift the burden.
If it makes you feel better, then I withdraw that comment about Gibson and replace it with a direct dismissal of your unevidenced claim.