Imre wrote: >On Fri, 31 Mar 2000, Mark Peterson wrote: ...... >> Take the simplest flash-an-LED PIC code you can come up with. If just one >> bit in just one simple ASCII character byte is flipped, the program will >> not run. We as on-lookers to the error would say "Of course it will not >> work. All of the code, characters, bytes, and bits must be just right for >> the program, device, and circuit to work. Someone must correct it. It >> will not fix itself." Yet, a single retina cell and in its functional >> relationship to the rest of the optical and nervous system, is many, many >> times more complicated and interdependent than the most complex devices and >> systems created by human beings. > >1. Ashby has proven it is possible to build such a system. >2. Imagine you use 1000 wires with 10x10x10 OR gates to detect whether >there is ONE input, instead of a single one. That system would not mind if >some of wires or some of gates drops out. This is a simplified way how >human (and other biological) beings does work. Using this analogy, you >will understand what happens if the last gate (the nervus opticus) is cut, >contrary to a small locus on the retina. > As Imre correctly alludes to, there are entire areas of science devoted to the study of neural organization - neurophysiology Hubel & Wiesel, Lettvin, etc), neural net studies (Rumelhart, etc), cybernetics (Ashby, etc). Models of the nervous system have been built which incorporate extreme amounts of redundant circuitry, in order that they will function correctly, even after many connections are cut. These are based upon observations that, in certain cases, the brain can compensate for severe nervous system damage, by virtue of built-in redundancy. The nervous system is more analogous a TV screen that is still easily readable, even after 1000s of CRT phosphor points have been burned, than to a single LED that lights or not. In fact, you don't need "someone to correct it". It is to some extent built to "fix itself", or at least to compensate. ------------- To expand on what I wrote previously, regarding "modification of the internal organization of the central visual system of mammals based upon early visual experience", a) Retinas are more or less hard-wired. b) Cortical centers are probably more or less hard-wired in adult animals, but of course, are still able to learn new things throughout life, via some as yet not fully understood mechanism of learning/memory. Probably the growth/efficacy of synaptic contacts rather than structural reorganization. c) However, during early development stages, cortical centers are "maleable", and can actually change their "internal organization". This is different from learning. The work begun by Hubel & Wiesel shows that young animals presented "biased" visual input are subsequently and irreversibly unable to ever perceive inputs of the opposite bias as well. Cells which would have become opposite bias cells instead develop as same bias cells, and can't go back. Instead of a 50-50 situation, you end up with 70-30 or 80-20, etc. The basic idea is that, during a "critical" period of early development, nascent cells can go either way, and the way they go is in large part due to what they experience. This all has potential ramification in terms of visual acuity, astigmatism, perception of lions in the savanna or monkeys in the trees, etc. Possibly also in the creation of society viewpoints in a potential "Brave New World" environment. Indeed, it almost looks as if there is a "plan" here. Lower animals are more hardwired, and less able to learn. Intermediate animals have some hard wiring, some ability for post-natal reorganization, and some for learning. In higher animals, "reorganization" may be present at a more abstract level, in terms of having a supreme ability for learning. [sorry if this isn't as much fun as running through stop lights at high rates of speed wearing a Ronald Reagan mask and with a James Bond revolving license plate].