Saturday, November 13, 2004

Limits of scientific explanation?

Following a discussion at Sean Carroll's blog-post on science and god :

Physical color (mixtures of wavelengths of light) maps into perceptual color (what I perceive). We can compare the maps between physical to perceptual color between two individuals. For instance, the perceptual differences between a "normal" person and a "red-green colorblind" person would be objectively established by their ability or lack of ability to distinguish between two different physical color stimuli.

If we simplify and say color is defined by Red, Green, Blue color coordinates that are in the range (0,255) and as displayed by a particular computer monitor, then we can do some easy computer experiments to construct the map of the physical to perceptual. For instance, I could have the computer display a solid area colored with one color (r1,g1,b1) surrounded by a solid area of another color (r2,g2,b2), and have it ask me whether I perceive a difference. Based on what I can and cannot distinguish, the nice 256 x 256 x 256 cubical lattice of Red, Green, Blue color coordinates will map into possibly some smaller set of points, representing the adjacent colors that I can distinguish. (I'm leaving out complications like the possibility that my ability to distinguish two adjacent colors may be affected by the presence of a third, distant color in my field of view.)

But speaking for myself, because I do not know how to pose this as anything but a subjective question, I can imagine two different physical-to-perceptual maps. I can do this via a physical transformation. The second digital image is derived from the first simply by exchanging the red and green channels.


Tulips by the fence


Tulips by the fence(2)


Subjectively, I can imagine someone, who sees the world like image 2 and not image 1. Identical neural nets in my brain and his brain may be firing, our physical-to-perceptual maps may be identical. Nevertheless, subjectively, I can ask whether what he sees is like image 2. I cannot see how to turn this into a scientific question, but I think this is a valid question nevertheless.

My reality is subjective. I'm the product of an enormous amount of natural selection, and so my reality corresponds closely to the physical reality that we can measure, capture in our equations, reason about. Nevertheless my reality is different from that physical reality. Scientifically, we can construct the physical - subjective reality map, verify that the maps are more or less homologous among all humans, and are different between, say, humans and parrots. With such maps, we may be even able to manipulate or enrich subjective reality. E.g, if today I can distinguish between 256 shades of physical red but only 180 shades of physical blue, perhaps some tweaking of some neuronal circuit or gene might make my blue perception expand to 256 shades of blue. But some subjectively essential ingredient eludes the scientific grasp, I think.

All that is required in order for humans to be able to communicate amongst themselves is that the physical-subjective reality maps for almost all humans are homologous to each other.

------

At the core of the Indic Traditions (which include the so-called religions of Hinduism and Buddhism), stripped of its sacred stories, deities, rituals, etc., is speculation about subjective reality, and the use of meditation as a tool to unlock its secrets. If we insist on calling this "religion", then it is very different from the Biblical or Quranic religions. If anything like "God" exists, this is the only place in the world left for him.



2 comments:

Travis Garrett said...

"Subjectively, I can imagine someone, who sees the world like image 2 and not image 1. Identical neural nets in my brain and his brain may be firing, our physical-to-perceptual maps may be identical. Nevertheless, subjectively, I can ask whether what he sees is like image 2. I cannot see how to turn this into a scientific question, but I think this is a valid question nevertheless."


- If the neural nets and mappings are identical for two brains (which probably isn't exactly the case, although they may be quite similar), then I conclude that they have the same perceptual experience, i.e. if one saw #1 then so would the other. All of the information, including the perceptual sensations, are contained in the neural nets. Sure, this isn't testable yet, but I bet it will be in the decades to come - with sufficient advances in computing power and direct neural implants (the optic nerve only carries ~ 10^6 bits/sec!). As you noted in another blog entry it does appear that just ramping up processor speed may be coming to an end as a method for improving computer performance, but we still have a long way to go in improving parallel machines. See www.clearspeed.com for instance - they have a 25 gigaflop vector math processor working on only a couple watts. There's only 3 orders of magnitude between that and the human brain. Once we reach that stage, we can then experiment with adding on different neural nets and seeing what kind of conscious experiences they produce...

Arun said...

Travis,

I hope we get there in my lifetime!

-Arun