In the LRB of 23 March, 1995, John Leslie wrote a review of Frank Tipler's book The Physics of Immortality: Modern Cosmology, God and the resurrection of the dead. The book is about the possibility that, at some distant future time, knowledge may have advanced sufficiently for history to be tracked back in the minutest detail so that, for example, the dead could be resurrected. Not surprisingly, the book and the review considered various impediments that nature places in the way of such a project, including the possibility that the chaotic behaviour of some of the systems involved might make things a bit difficult. . .
In his piece on Frank Tipler's The Physics of Immortality John Leslie tries to adumbrate chaos theory thus: `The classic example is the butterfly which, if given incredibly precise information and access to a gigantic computer, could deliberately induce or prevent a hurricane some weeks later, by fluttering this way or that.' He completely misses the point. A chaotic system is one in which an infinitesimal change in input causes an arbitrarily large change in output. Therefore, in order to predict how the system will behave, it is necessary to know its starting conditions with infinite precision - a finite approximation to the starting conditions, no matter how good, has no predictive power whatsoever. The information would not need to be incredibly precise; it would need to be infinitely precise. And the computer would not need to be gigantic; it would need to be infinite. Chaotic systems are thus both deterministic and unpredictable. John Leslie does mention Freeman Dyson's infinite computer based on asymptotic cooling towards absolute zero, but - even though only requiring finite energy - such a computer would necessarily take an infinite time to finish the butterfly's infinite computation; a fact that rather makes one wonder about the meaning of the word `finish'.
To move away from the dud outline of chaos in order to try to iron out one or two of the creases in Tipler's woolly book: suppose that quantum mechanics does obviate the need for infinite precision in the chaos calculations by virtue of the granularity that it superimposes on the universe. Then - if finite - the universe itself is a finite-state machine. All that can happen is rearrangements - swapping pieces between different squares on the universal chessboard, if you will. If such a universe only lasts for a finite time Tipler's idea of infinite progress is shot immediately. If it lasts for ever, after a while all possible arrangements will have been exhausted, and the whole thing would have to go round the loop again - a grand version of Tipler's `horror of the Eternal Return'; the idea of progress seems a trifle elusive in this case too. If the universe is infinite in extent, quantum-mechanical granularity doesn't eliminate chaos as the infinities are back in the sums and the computer would need to work for ever; any progress (whether infinite, or just trivial stuff like raising the dead) that depended on answers to the sums would never even get off the ground. You can't beat the house by having an infinite number of computers either, because it would still take an infinite time to distribute the initial data amongst them before the - very fast - computations could begin. Q. (as we used to write in non-woolly school geometry lessons) E. D.
It is rather important when dealing with this sort of thing carefully to distinguish between the very big and the infinite: something that is infinite is different in kind - not merely in degree - from something that is just very big.