This is a painting done by a congenitally blind artist named Esref Armagan.
The most obvious question is how his brain is able to perform such feats of perspective, but as the article mentions, it is well understood that “blind people… understand and can draw in three dimensions”
With that said, I think this particular “how” question is easily answered with modern paradigms of neural plasticity/pruning etc
I believe the more puzzling question to ask is not how he can perform such tasks, but rather, how his developing brain learned to generate high-order representations of three-dimensional space to such a phenomenal degree of accuracy.
Well, “I was taught, he says. Not by any formal teacher, but by casual comments by friends and acquaintances.”
This remark, combined with the fact that “it is impossible to know if he had some vision as an infant”, makes it difficult to extract any conclusive insights about whether his “mind’s eye” is purely mapped out in non-visual sensory-terms (the primary contributors likely being kinesthetic and proprioceptive), and his incredible “accuracy” is simply the result of his early peers subtlety nudging him back and forth until he got it “right”.
An alternative answer is that his brain got some “extra” reinforcement by crude visual data before his eye completely degenerated, thus making his inner conceptual space not in purely non- visual terms as is suggested by the article. This would also explain why his degree of accuracy is greater than most other congenitally blind people.
These are difficult, but fascinating problems in the psychology of perception, but regardless of Mr. Armagan’s unique skills, there seems to be an emerging consensus from all psychological disciplines that whatever “perception” is, it is realizable across many different modalities.
Minds and Brains 
This question of how his brain learned high-order representations of three-dimensional space might be answered most simply by asking yourself the same question. Now for you and I, it’s obvious that our own maps are the combinations of sensory experience, which allow us to get a visual picture. But one might ask if our visual organs our even sufficient enough to make up the whole picture. I know, for me, that I get a far better spatial picture with the use of sight, sound, and sensation together. If I’m remembering a picture in my head of a place I have been, then it usually gets distorted with time. Sounds do not seem to fade as quickly, neither do taste or smells. Perhaps his ability to have a clear representation of a 3D space lies in a sonar-like ability.
Thanks for the post, although its a bit dated. I am going through and reading all of your stuff.
The issue of whether depth perception was innate or learned was a major debate in the 1800’s between the British Associationist philosophers for nurture, and the Gestaltists arguing for nature. The issue was resolved in my mind with an experiment on newborn chicks who were trained to peck at the smaller of two objects. When the chicks were then presented with a large object farther and a small object nearer, such that the smaller one projected a larger image on their retina, the chicks instinctively corrected for perspective and pecked at the smaller nearer object.
But the issue runs deeper than that. Like perception, mental imagery requires a mental representation of space. That representation would end abruptly at the boundaries of the representation UNLESS it were bounded with a perspective warped grid like THIS
http://cns-alumni.bu.edu/~slehar/epist/epist6.html
I would argue that a congenitally blind person’s mental image space is warped, as is our perceptual space, such that objects in the distance appear smaller by perspective, but the scale of space shrinks with depth so that the objects are perceived both as smaller, and at the same time undiminished in size.