Natural Intelligence. And its 'origins'
It is difficult for the modern, educated person to imagine a non-evolutionary origin of intelligence. Immaterialism opens the door to non-Darwinian explanations of reality, but it does not require such explanations. One could naively imagine a Darwinian form of immaterialism being played out in a purely experiential realm. In fact, I am sympathetic to the view that an immaterialist form of Darwinism would give rise to a God of Love as its ultimate expression.
My strong reservation with any form of Darwinism is its dependence on a temporal dimension. Evolution implies a unidirectional flow of time. Thus evolution cannot explain the origin of such a structure. One might wish to think of time as ‘evolving,’ but clearly this would be a major and perhaps untenable extension of the notion of evolution. An immaterialism that posits an a priori space-time manifold is broaching self-contradiction.
Another way to pose this problem is that if time is a construct of mind or intelligence then we cannot invoke time to explain intelligence. For immaterialism, mind is the a priori element. The cosmic mind does not evolve or change in any essential manner. It may employ the construct of time as an essential ingredient of itself, but that is a very different state of affairs.
We are naturally appalled at this evasion of our instinct to explain everything, but any system of understanding must start somewhere. It cannot start from nowhere, logically speaking. By starting with elementary particles or fields, physics seems more satisfying in this regard, but that satisfaction cannot be taken as a substitute for truth. Furthermore, rationalism must start with the rational, that is to say a minimal system of meaning. That meaning can ever be derived from non-meaning is not part of any rational metaphysic. Rationalism, immaterialism, idealism, relationalism, organicism and the coherence theory of truth each envision an a priori system of things, none of which need be, or possibly even can be, finite. Similarly, linguistic competency is not something that is finitely specifiable, certainly not in the sense of being programmable. One may be able to simulate linguistic proficiency with a computer to a limited degree, but not emulate it. All the king’s programmers will not be able to pass the Turing test for natural intelligence.
It should not be difficult to remain skeptical about artificial intelligence or about the evolution of intelligence. Linguistic ability is a case in point. What is the minimum conceivable program that would enable a computer to ‘learn’ a language? There are two schools of thought: connectionism and nativism. The former would argue that all we need is a sufficiently robust connection between any two ‘neurons’ to enable machine language. The latter would argue that yet to be specified ‘deep’ rules of grammar are also needed.
A connectionist could argue that just a few genetic mutations might account for the development of human linguistic ability beyond the natural intelligence displayed by other primates. Too many mutations would make it very difficult to construct an evolutionary scenario for a linguistic community. But one should then wonder why it has been so difficult for us to reproduce this same small step in our machines. Is not this quantitative conflict between the explanatory and programming problems making any reductive model of language impossible to conceive? However, a non-reductive model of language would open the door to immaterialism in a major way. How much longer can reductionism bear the strain in this arena, or in any other arena for that matter? How much strain can the materialists tolerate before they are willing to swallow immaterialism? What should be surprising is that neither the materialists nor their opponents recognize this simple logistical impasse.
- | Contents |