Humans have grappled with the enigma of our conscious experience for millennia, attempting to untangle its complex threads. The so-called "Hard Problem of Consciousness," first articulated by philosopher David Chalmers, asks why certain physical processes give rise to subjective experience or consciousness and how these experiences can be scientifically explained.
But perhaps the Hard Problem of Consciousness is not merely hard; but is, in fact, impossible.
Rice's theorem is a result in the theory of computation that tells us, roughly, that anything interesting about the behaviour of a computer program is, in general, unknowable. You might be able to know certain things about the behaviour of specific programs, but the general question of looking at a program and predicting its behaviour is not only hard, it is impossible. More formally, Rice’s theorem tells us that no non-trivial property of the behaviour of Turing machines is decidable. (A Turing machine is a theoretical model of computation that can simulate any algorithm, and a property is non-trivial if it holds for some Turing machines but not for others.)
Attending to the Hard Problem of Consciousness from a purely computational point of view, we suggest that understanding consciousness requires discerning specific, non-trivial properties of the human mind's computational processes. If, in the limit, we seek an algorithm that could decide whether a given physical process, such as neuronal activity, corresponds to a conscious experience, are we not in a state of sin?
Consider, for example, the task of designing an algorithm that could decide whether an arbitrary Turing machine produces a conscious experience. Such an algorithm would have to be capable of analyzing the machine's behaviour and determining the presence of subjective experience.
If any machine (including the very complicated and baroque organic one in my skull) can be conscious, then the Church-Turing thesis implies that there exist Turing machines which are conscious. It cannot be otherwise unless one is prepared to reject the Church-Turing thesis. At the same time, there exist examples of Turing machines that I would be comfortable labelling as evidently non-conscious (though the panpsychists might object). Consciousness is a non-trivial property of Turing machines, thus Rice's theorem states that no such algorithm can exist, as non-trivial properties of Turing machines are undecidable.
Are our own subjective experiences, the very things that make us conscious beings, fundamentally unexplainable through scientific inquiry?
I hasten to note that this query does not invalidate the study of consciousness or the pursuit of understanding the nature of subjective experience! The truth of Rice’s theorem does not stop computer scientists from proving many interesting things about Turing machines, furthering our understanding of the nature of computation.
There is still much to learn about the neural correlates of consciousness and the cognitive mechanisms underlying our experiences. Rice’s theorem is not cause for despair, but simply a reminder that we must be humble; we must accept that some aspects of consciousness may forever elude our comprehension.