What if someone thought an integrated circuit was an organism? computer programmer, engineer, biologist, causal connections, connectionism, reductionism, fads
What if a biologist thought or pretended to think that an integrated circuit was an organism? He would look for signs of life. And what would he find? No food intake. No breathing. No chemical excretions. No reproduction. No aging. No death. No life.
He would find responses to stimuli, but they might be different than the sort of responses that a computer programmer would expect. After trying many different levels of electrical inputs, he will note corresponding outputs. Whereas the computer programmer will think of input and output in binary terms -- either on or off -- the biologist will look for non-linear relationships. And he might just find some. Of course (no surprise to either the biologist/electrician or to the engineer), he will note that certain levels of input will be catastrophic.
What if neurons perform their instrumental role in perception etc. precisely in virtue of their CAUSAL interrelations, AND these interrelations correspond to non linear functions rather than to the all-on or all-off functions that preoccupy a computer programmer? In such a case, since the true causal relationship between each cell will be invisible to the computer programmer, so too will the physical basis of the operations (cognition, appetition, effort) associated with the brain.
IF a person whose day job is to design computer chips were made aware of the vast array of non-linear functions that characterize the input/output of a cell, he might make something that would simulate that sort of interaction. But this would be a simulation rather than a duplication. In fact, the simulation of just ONE cell may require a computer or even a supercomputer. And simulating the interactions of billion of neurons in a brain may be a bit too complicated for the designer.
Let's suppose, however, that the designer COULD make a computer chip that interacts with others in a manner analogous to how cells interact: to do this correctly, he would have to go back to school and learn cell biology. In such a case, the engineer would be able to set aside all of his pre-connectionist speculations about artificial intelligence in terms of ones and zeros. Bio-chemistry rather than good old-fashioned artificial intelligence would supply the understanding needed to create this chip.
It would seem that computational theories might have to be trashed, and they will be replaced by the effort to use nanotechnology to create artificial neurons.
***
Afterthought: one who objects that cells can be thought of in terms of output/input rather than in terms of causal relationships internal to the cell has no right to object to the Chinese Room argument, for he is treating the cell like a miniature Chinese Room. In other words, while the CRA might not defeat connectionism generaliter, it does defeat the sort of connectionism that treats individual neurons in the same way that a functionalist treated the whole computer (I have to say this more clearly, but no time for doing that now). It follows that we could characterize connectionism as a kind of retreat from a more mechanical to a more organic conception of cognition. But perhaps it is not enough of a retreat. Perhaps a surrender is called for.
He would find responses to stimuli, but they might be different than the sort of responses that a computer programmer would expect. After trying many different levels of electrical inputs, he will note corresponding outputs. Whereas the computer programmer will think of input and output in binary terms -- either on or off -- the biologist will look for non-linear relationships. And he might just find some. Of course (no surprise to either the biologist/electrician or to the engineer), he will note that certain levels of input will be catastrophic.
What if neurons perform their instrumental role in perception etc. precisely in virtue of their CAUSAL interrelations, AND these interrelations correspond to non linear functions rather than to the all-on or all-off functions that preoccupy a computer programmer? In such a case, since the true causal relationship between each cell will be invisible to the computer programmer, so too will the physical basis of the operations (cognition, appetition, effort) associated with the brain.
IF a person whose day job is to design computer chips were made aware of the vast array of non-linear functions that characterize the input/output of a cell, he might make something that would simulate that sort of interaction. But this would be a simulation rather than a duplication. In fact, the simulation of just ONE cell may require a computer or even a supercomputer. And simulating the interactions of billion of neurons in a brain may be a bit too complicated for the designer.
Let's suppose, however, that the designer COULD make a computer chip that interacts with others in a manner analogous to how cells interact: to do this correctly, he would have to go back to school and learn cell biology. In such a case, the engineer would be able to set aside all of his pre-connectionist speculations about artificial intelligence in terms of ones and zeros. Bio-chemistry rather than good old-fashioned artificial intelligence would supply the understanding needed to create this chip.
It would seem that computational theories might have to be trashed, and they will be replaced by the effort to use nanotechnology to create artificial neurons.
***
Afterthought: one who objects that cells can be thought of in terms of output/input rather than in terms of causal relationships internal to the cell has no right to object to the Chinese Room argument, for he is treating the cell like a miniature Chinese Room. In other words, while the CRA might not defeat connectionism generaliter, it does defeat the sort of connectionism that treats individual neurons in the same way that a functionalist treated the whole computer (I have to say this more clearly, but no time for doing that now). It follows that we could characterize connectionism as a kind of retreat from a more mechanical to a more organic conception of cognition. But perhaps it is not enough of a retreat. Perhaps a surrender is called for.
Comments