Models & Machines
In science, you can always model the object of your interest. You can make a physical model, for instance, which can move, out of an energy supply and bits and pieces which (separately) don’t move. Or you can make a mathematical model (often of extraordinary complexity). But can you make a model which has conscious experience out of things that don’t? No, you can’t. One can safely utter the challenge: ‘Construct out of unconscious elements a machine which will have conscious experience.’ It is impossible to imagine how such a task might be performed. When I was in discussion some few years ago with an artificial intelligence expert, he seemed to admit this fact; but claimed he lived in the hope that, suddenly, for no discernible reason, one of these days, one of his complicated electronic machines might become conscious. The magician would have produced a white rabbit without himself knowing how; or perhaps we should write that some unknown source had produced the white rabbit for some unknown reason. However, if in fact (and most improbably) such a thing were to happen, and if on further investigation no material cause could be found, then, since we would have no materialist explanation of the manifestation, we should still be left with the possibility that some non-material event had occurred.
Wishful materialist though he is, the same insoluble problem is met by Hofstadter at the end of I am a Strange Loop, a book whose ambition is to make feedback seem a plausible model for consciousness:
What would make a loopy abstract pattern, however fancy it might be, constitute a locus of interiority, an inner light, a site of first-person experience?
He has, I fear, no solution to offer except wishful thinking and the persuasions of style. As is well known, Colin McGinn, in despair at the impossibility of modelling and hence grasping consciousness in a ‘scientific’ fashion, suggests that we are constitutionally incapable of understanding it.
The impossibility of modelling consciousness is a real problem for materialists. Humphrey’s difficulties, described above, are another instance of this point. An exceedingly interesting article by Fred Dretske contains ‘a recipe for making a thought’. His premisses seem to be not only ‘If you can’t make one, you don’t know how it works’ (which seems, I think, reasonable enough), but also ‘Anything that can’t be made can’t possibly exist.’ But this amounts to saying (1) ‘anything that can’t be imitated can’t possibly exist,’ and also (2) ‘Anything that I don’t understand can’t possibly exist.’ Whether the first of these propositions is true or not I am not concerned to argue here. The second, however, cannot possibly be true; and indeed the impossibility of modelling consciousness proves it.
What Dretske’s proposals really demonstrate is as follows: they show the limitations on materialism: we are well able to model material objects – for we are modelling them out of concrete matter or abstract mathematics. But we shall find it impossible to do the same with immaterial phenomena such as thoughts, hopes, fears, emotions or qualia. Still less with consciousness, in whose phenomenal field these experiences have their being. In short, you cannot either model percipere out of percipi or derive the former from the latter. For the two are the ontological opposites of each other.