The brain is not a computer

If a keyboard is equipped with a few sophisticated functions, we call it intelligent There are intelligent regulators, cars, buildings. Tomorrow we will have intelligent vacuum cleaners. And to call a person intelligent soon will mean that he behaves like a program. No one can stop this development. There is only one way out; we must invent a new term for what was called intelligence before the introduction of the invention of the intelligence quotient, and before McCarthy borrowed intelligence for the term Artificial Intelligence from Ashby's intelligence amplifier, the computer.

Calling the brain a machine is worse because a replacement of terms will not help. Our mind, or rather some of its functions and products, can be described by models which permit us to call the model a machine. But the mind can jump out of any model or machine. There is no justification for the hope that such jumps will be programmed one day and thereby force the mind back into the cage of the machine model.

The case of 'Automatic Translation' of natural languages has shown that one can produce only partial models of semantics - a conclusion the specialists could have drawn from the philosophical transition from Wittgenstein I to Wittgenstein II. The case of the 'General Problem Solver' has shown that the control exercised by our mind on the complicated structure of the brain, which defines the brain as well as the whole body as a unit, is beyond our programming faculty. And the case of a program simulating consciousness has shown that such attempts divert the thinker away from entities like consciousness rather than explaining them.

Of course, there are the ever-ready arguments of science - 'so far' and 'not yet'. But unless one studies the limits of the particular model and of general models, is this not like the search for speeds greater than the velocity of light, for information beyond the uncertainty principle in logic and physics, an attempt to find a coding principle which would allow transmission of signals beyond the channel capacity? Neither for the computer nor for the brain do we have a formula to calculate the limits of their power. Information cannot be measured, nor can machine intelligence, but there are limitations. Moreover, the brain is not a computer. The brain can compute, it can regulate and it can do character processing, but it does so using principles quite different to the computer programs. Calling the brain a machine not only neglects the individual character of the mind - it raises the expectation that concrete and abstract machine principles make up the essence of the brain and mind. There is nothing but a set of models embodying such machine principles, and we have no evidence that the main operation of the brain is based on these models or any generalisation of them.

It is wise to respect the boundaries and to keep them visible in our way of speaking. Considering the brain a machine is certainly useful for our technical thinking and for part of our scientific thinking, but the brain has no reason to be a machine. It is not the result of professional engineering or software design: it existed before engineering and design began.

Sometimes at this point someone puts the argument that modesty requires the admission that the planet Earth is an unimportant grain in the universe, and that man is but a possibly rare effect of the stream of physical events. It is orders of magnitude more immodest to extend notions produced by under-informed man to such general statements. We are indeed the centre of the universe because we look from this planet Earth. And we cannot call the brain a machine because we have not built it and we know very little about how it came into existence and how it works. We have a mind which remains a mystery for science. Rejecting this mystery by superficially calling the brain a machine is a bad service to science, language and mankind.