This is rather more interesting from a machine learning point of view than philosophy. For instance, it seems a blind person does know about space from touch, but cannot innately identify space from visual information. What can we infer about the way our brains process sensory information from this fact?

There's an interesting account here: http://www.richardgregory.org/papers/recovery_blind/recovery... which contains the following passage about a man who recovered his sight:

"In view of his depressed state, we felt it best not to undertake formal tests. We did, however, ascertain that he was able to find his way about without the use of his eyes, and that he could detect the presence of houses and doors by the echoes from his footsteps. He was still fascinated by mirrors, and he still noted improvement in his ability to see. In particular, he said that he noted more and more the blemishes in things, and would examine small irregularities and marks in paint work or wood. Quite recently he had been struck by how objects changed their shape when he walked round them. He would look at a lamp post, walk round it, stand studying it from a different aspect, and wonder why it looked different and yet the same."

I think the last line is interesting. There is a conflict between the spatial reconstruction and the optical appearance.

This reminds me of trying to localize yourself using a very narrow field of vision, where you see things but it's hard to paint the picture of the environment

(It may even be possible that the person has some kind of shortsightedness, I didn't see the article)

Yes. In fact there is a tradition in art theory of making a distinction between close vision (involving movement of the eye across a surface) and distant vision (in which the whole scene can be taken in at once).

It's actually a fascinating bit of intellectual history, though I am not sure whether it counts as objective scientific knowledge.

It started with the sculptor Adolf von Hildebrand and was developed by Alois Riegl into the notions of "haptic" and "optical" vision. Wolfflin then expanded that distinction into five polarities (linear vs painterly, plane vs recession, closed vs open, multiplicity vs unity, clarity vs complexity). In mathematical terms haptic seeing is concerned purely with manifold surfaces whereas the optical can also represent non-manifold geometry.

Even more significance was later hung on that structure by the philosophers Walter Benjamin and Deleuze and Guattari. In particular the haptic is associated with surface, with smooth spaces of the mechanical, of trajectories and motion, while the optical is associated with volume, with striated spaces of the organic, of hierarchy and stasis. "parametric" vs. "implicit" if you like.

It's a bit of a philosophical rabbit hole!

Ooooh. I never knew that there was this model of these different types of vision! I am a congenitally blind person who can (occasionally) find a way to make myself see a tiny bit better, and a lot of what I do is allow my eyes to wander up and down different surfaces. I basically pretend that they are fingertips. I am much less good at the other types of vision, but I've played around with them a bit; it's like I am hurling my body through space. I can't access the New Yorker article but I suspect my experience is different because I have always felt sighted even though I've never had vision. I really suck at being a blind person and thus am incredibly motivated to figure out how sight works, because I can already tell that I'm a visual learner (and an artist) and sight makes so much intrinsic sense to me.

For the blind, space and time are identical. The reason a blind person would imagine a typewriter 100 feet away as the same size as one 1 foot away is because the only time a blind person has ever experienced a typewriter at all is when it was at arms length. The idea of perspective is totally foreign to them.

The connection between space and time comes from this extreme locality of the sensory world of the blind. If walking down a hallway, they can not perceive the end of it, and can only perceive the region they are currently in contact with. As they walk down the hallway, they learn that 2 seconds down there is a light fixture, 8 seconds down there is a bump in the carpet, and at 14 seconds the hallway ends. Intuitively, the concepts become very strongly linked. Oliver Sacks discusses this in some of his books.

I think the most informative thing about this is that it is completely consistent with a view of the brain as an associative engine - stimulus which occurs together gets associated with each other, and presence of one stimulates an expectation of the others.

Wow, I am blind, and I am very very not good at this way of thinking, as evidenced by the fact that it is very difficult for me to learn routes from point a to point b based on this type of info. What's more helpful to me is imagining my line of travel and focusing how my body aligns with different landmarks.

I don't understand why this article so entirely avoided an extremely important factor - the youth of those being cured. If you give sight to a 3 year old or a 10 year old or even a teenager, their chances of developing normal sight are good. If you give it to someone after puberty, their chances drop quickly, getting worse the older the brain is. Neuroplasticity never goes away entirely (as was believed before), but it is still radically reduced compared to during childhood and especially adolescence.

I do think that's true, but if you're like me and never identified with blindness or fully adapted to it, I think your potential for sight restoration is still good. (I'm 28).

I don't think these experiments answer the philosophical question. The philosophers were assuming that the blind would have perfect vision once they were restored. But it turns out that the brain needs a long period to learn how to see. During that learning period the brain will slowly learn how to see roundness. A better, though unethical, experiment would be to somehow prevent a restored sight person from seeing any spherical objects for a long period and then show them a spherical object.

> But it turns out that the brain needs a long period to learn how to see.

I'm surprised that anyone's surprised by that since it's the same thing with hearing and cochlear implants. Even after 12 years, my hearing is still getting better in some ways.

That defeats the purpose of the experiment. Now you are asking a different question; "does the brain learn features good enough to generalize to a class of objects it's never seen before".

I would guess it does, because all the time we see and hear things we've never seen or heard before. For example, you can show a child a picture of an elephant, they will then know what an elephant looks like and recognize them, even if they've never seen an elephant before.

Exacy! It also works the other way around! My eye got seventy damaged and the retina ripped. The doctor managed to stitch it back together but I saw the fold (a 'dent' in my vision, everything had a fold in it. Took 4 weeks for my brain to re-calibrate and the fold's gone.

I can't see anything unethical in that, it's just highly impractical (which is why it might be difficult/expensive to find test subjects).

But my guess is that there isn't actually any "spherical" vs. "non-spherical" distinction at the level of processing that is (presumably) untrained in blind people, but rather some more general recognition of sharpness of objects' edges of which spheres are just an extreme case, and that would be pretty much impossible to control for, as you'd need the test subject to live in a "uniform sharpness environment", which would have to include their own body's shape(s) ...

> I can't see anything unethical in that

You don't know how it will affect the life of the test subject.

Same reason why experimenting with raising a child by never letting him hear anything for ten years (to test how brain would work) is unethical.

Informed consent is impossible to obtain with children, if it was then it won't be a problem.

> I can't see anything unethical in that

Really?

"We're going to give you sight but not let you see anything round for 2 years. So we need to keep you in this cage unless you are blind folded."

Yeah. No ethical issues here ...

That's why it was done on cats, half a decade ago, and even that was controversial.

David Hubel and Torsten Wiesel won a nobel prize for their research where they sewed kittens eyes shut and then unsewed them when they were adults. It was landmark research.

Please explain why it would be unethical to do that with someone who agrees to it for a payment of 10 Mio. USD and with the option to leave at any time with only partial payment. And while you are at it, please also explain why testing of new drugs with unknown effects on human subjects is not unethical (or is it according to your opinion?).

They couldn't possibly consent to such when they have no idea what they are actually giving up. If someone who was previously able to see then went blind was given the same choice, then they would have a basis from which to value sight and make an informed decision of consent; however, I don't think a person who has never seen before can possibly give valid consent to such an experiment.

I must say, that argument is better than what I had expected anyone to come up with - but I'm still utterly unconvinced.

For example, it would pretty much follow from your argument that we would have to force affected people to undergo surgery, as they could not possibly give informed consent to staying blind.

You don't need first-hand experience with something in order to make an informed decision. Someone who has a haptic understanding of sharpness and an understanding of the finiteness of their life and of the general economical power of 10 m USD and possibly even an interest in philosophical questions surrounding perception [...] should very well be able to make an informed decision whether they think it's worth it for them.

People also don't really know what it's like to live on a space station until they've been there, and yet astronauts very well are able to consent to being brought up there. Or for that matter, we allow people to never have that experience instead of forcing them to become astronauts, even though they can't value the experience (supposedly).

Some decisions in life are difficult, but we absolutely do routinely allow people to make difficult decisions, and I think that's the right thing to do.

That's a strange way to put it.

Suppose I'm deadly allergic to all fish and fish-tasting foods. Now someone comes up with a way to cure that, so I can eat fish for the first time. And then they say, "actually, we'd like to see what happens if someone first tastes salmon several years after they first taste haddock", and they offer me a lot of money not to eat salmon for two years. Can I consent to that, even though I don't know what I'm giving up with salmon? What about if I started with no taste at all?

(N.B. This isn't meant to be a knock-down argument against your position. I disagree with your position, and this is an argument against it, but it's not my true rejection. My true rejection is along the lines of "you have no right to tell people what they can and can't agree to", but I don't care to get into that.)

"You have no right to tell people what they can and can't agree to"

I know you said you don't care to get into this, but I'm honestly curious; where do you draw the line (if at all) on that position? Because the concept of enforcing an inability to consent to things is the basis behind minimum wage laws, age-of-consent, the unenforcibility of draconic EULAs and non-compete clauses, and probably some other laws that I can't think of which are wholly designed to protect people from being able to make terrible choices that hurt not only themselves, but the rest of society by validating that choice as a potential option.

I'm not trying to slag on you and I recognize that you don't want this to turn into a debate where everybody shouts at you, and that's probably what's going to happen if you respond; I'm just curious about the boundaries of your position.

A fair question.

I try not to be dogmatic. Sometimes, bad decisions have externalities that justify forbidding them. I don't think that's the case in this particular instance. To briefly address your other examples-

Minimum wage: in this case, I think the economic factors probably outweigh the freedom-of-contract principle. (I'm on the fence about what the economic factors actually support.)

Age of consent: I think there's a meaningful sense in which children are not "people" for the purpose of entering contracts.

EULAs and non-competes: I haven't spent much time thinking about this. Off the top of my head, I feel like these should be enforceable, but not as currently implemented. Like, suppose before I let you buy my software I get you to mail in a signed form saying that you have read the EULA and agree to its terms. Then I think I should be able to enforce the EULA against you.

As long as the person is participating by free will.

That's a slippery slope argument, that's not particularly convincing, in particular given that experimentation on humans with informed consent is very much a routine thing to do, both in psychological and in medical research.

All new drugs are tested on humans at some point, when it's not really known all that well yet what the reactions will be, that's experimentation on humans right there, and very crucial experimentation at that, as that's how we figure out what's safe and what is not. The one thing that you can not do without, though, is informed consent.

Sure, but might end up on a slippery slope if you let any random person say "hey hospital, please treat me with that experimental drug; I give my informed consent and will pay you $100k for that".

That's why we usually put the burden of proof on the person conducting the experiment that there actually was informed consent. And you don't give informed consent by stating "I give informed consent", but by demonstrating that you understand the (potential) consequences, and then consenting to what you have demonstrated to understand.

What this experiment shows is there is no such thing as a perfect vision without knowing spheres and cubes. Or at least without knowing curved surfaces and corners

I recently read Robert J Sawyer's "WWW Trilogy", in which the main character has her blindness cured via technology [not a spoiler; reader knows this going into the first book].

The main character's acclimation to and approach to sight throughout the trilogy was interesting, and I recommend in view of this article.