“It is even harder to predict the possibility (or the timescale) for building a device whose action depends on a physical theory that we do not even know at present. Such a theory would be needed, I am claiming, before we could understand the physics underlying a non-computably acting device-'non-computably', that is, in the Turing-machine-inaccessible sense that I have been using in this book. According to my own arguments, in order to build such a device we should first need to find the appropriate physical (OR) theory of quantum-state reduction-and it is very hard to know how far we are from such a theory-before we could begin to contemplate its construction. It is also possible that the specific nature of that OR theory might itself provide an unexpected complexion on the very task at hand.

At least, I suppose that we should need to find the theory first, if we are to construct such a non-computational device. But conceivably not: in actual practice, it has often been the case that surprising new physical effects have been discovered many years before their theoretical explanation. A good example was superconductivity, which was originally observed experimentally (by Heike Kammerlingh Onnes in 1911) nearly 50 years before the full quantum-theoretic explanation was eventually found, by Bardeen, Cooper, and Schrieffer in 1957. Moreover, high-temperature superconductivity was discovered in 1986, cf. Shent et al. (1988), also without prior good reason to believe in it on purely theoretical grounds.”