Roger Penrose is a remarkable mathematician and physicist who opens his book with a most challenging question: can a computer have a mind?
It has long been believed by philosophers and many scientists too, that the mind is something fundamentally separate from the brain, that the brain simply housed the mind. This tied in comfortably with religious notions of some essence of a person which could persist after death, some location in which consciousness itself would sit.
Against this notion, materialist philosophers have always argued that consciousness is simply a property of the brain itself. Some philosophers such as Daniel Dennett and Douglas Hofstadter have argued that it is the very fact of self-referentiality that enables the brain to become self-aware, able to monitor its own processes, and that this is the basis of what we call consciousness.
But Penrose has a different approach. His purpose in the book is to dispose of the idea that we could ever produce a computer that could duplicate the workings of the brain. His argument against the so-called “strong AI” position is based on a simple idea. All computers are based on determinate algorithms, precise rules which the computer simply follows. But we, as humans, know that there are mathematical results that cannot be calculated using algorithms, there are insights that we know to be true, and which cannot be derived algorithmically within the formal mathematical system. Therefore, he concludes, we cannot make computers which could duplicate what our brains can do – they simply cannot show “insight”.
In developing his argument, he provides an exposition of the nature of algorithms, Gödel’s Theorem, and Turing machines, and starts to investigate the relationship between mathematics and reality leading to a consideration of the problems of indeterminacy, which of course leads into quantum mechanics. He has to go through this long process because he needs to attack the certainty on which conventional understanding of mathematics rests.
The problem of the nature of reality posed by quantum theory concerns the collapse of the wave function, that point at which a wave function collapses from an indeterminate state, to a single state as a result of the interaction with an observer. Manifested by the fact that we cannot know both momentum and position at the same time in quantum physics, this fundamental indeterminacy is related by Penrose to the problem of actually knowing reality. What is perceived of as reality is therefore seen as intrinsically related to quantum state changes.
Such a collapse of the wave function has always been an object of debate, whether it is a fundamental physical phenomenon or simply an epiphenomenon of another process. Is it a fundamental aspect of reality, or simply an artifact of the measurement process. Nevertheless, Penrose takes us through the notion that such quantum changes might have something to do with the brain’s ability to make creative leaps, those moments of insight which allowed mathematical properties to be found which cannot be derived.
Just as quantum events can suddenly settle on one of many possible states, so the brain by a superposition of computational states would jump across the space of possibilities and reach its conclusion. For this reason, Penrose suggests that rather than an algorithmic principle, a computer model of the brain would be a quantum computer.
Whether or not the brain can be adequately modelled with a quantum computer, Penrose seems to have a view of evolution which involves a predefined purpose. He says for example “There seems to be something about the way the laws of physics work, which allows natural selection to be a much more effective process than it would be with just arbitrary laws.” This belief is bordering on the creationist illusions of intelligent design.
By maintaining the position of the separation of the mind from brain, the mystery of consciousness, the indeterminacy of quantum mechanics, the reliance on “insight”, Penrose tries to build up a non-materialist case. But even after all his wonderful mathematical exposition and flair, his case remains very thin.
There is no need to worry about the Kantian “thing-in-itself”, and there isn’t any huge problem of perception except based on the premises chosen by philosophers. Although without mentioning it by name, he flirts with the notion of “qualia”, this superfluous concept has been disposed of very effectively by the work of Dennett.
Neuroscience has advanced greatly since the appearance of The Emperor’s New Mind and the Blue Brain project has already simulated a neocortex column and is on track to provide a simulated neocortex of the human brain within the decade. With the exponential advance of computing power, no doubt including the development of quantum computers, modelling the human brain is increasingly likely.
An alternative theory of consciousness is in the public domain – that of consciousness being a property of a self-referential neural network. As the simulations increasingly conform to observed functionality, we will understand more about the development of consciousness.
Penrose is successful in the limited claim that strictly algorithmic computation cannot model human insight, but it is still likely that the developments of neuroscience and computing will put an end to the quasi-mystical, quasi-religious notions of the mysterious nature of consciousness. We may well find that we are all just meat machines after all.