Go figure...

The computational architecture of the systems that process sensory information in human beings is that of a neural network. In particular, most of our networks have feedback paths that can lighten the weight of computing repetitive information. To illustrate the process imagine you are given a text to translate (the input). After you done your work (you have processed the information), you are asked to translate yet another piece whose first part is exactly the same as the previous, but has a different ending. Of course, you will take the first part of your translation (the output) and put it as the beginning of the new one. You’re easing up the weight. Consider now the most notable structures of a neural network: the beginning –the part that receives the input—, the middle—the part that processes the information—, and the end—the part that produces the output. What a “feedback path” in a neural network does is to send part of the output to the middle of the network, easing up the load of computing again the same old sensory stimulation. But what is fascinating about this is that in the absence of regular stimuli, these paths have a tendency (probably induced by the presence of gradient potentials in neurons) to self-stimulate the whole network. We are in fact much familiarized with this process, since it’s the responsible for producing vivid dreams, illusions, fantasies and hallucinations. But as with any other computational system, the information will eventually degrade: some details are lost each time the result is send back again to the middle, producing a different result that loses some definition when send back, and so one. Now, there is really no evidence to suggest that we don’t treat different percepts in the same way we treat sensory stimulation: a social situation is a very complex input that we process in the same good way. So, if you want to sit down and think over and over again a particular state of affairs, you must be certain to have a very good record of what happened (to work as stimuli), because if not, you’re leaving yourself to the merciless attitude of entropy and the third time you run your argument you will meet a very different picture of the original situation. This is the reason why “figuring out things by yourself” simply doesn’t work.