David Deutch - The Fabric of Reality

Now, it turns out that the type of oscillations of space that would make an omega point happen are highly unstable (in the manner of classical chaos) as well as violent. And they become increasingly more so, without limit, as the omega point is approached. A small deviation from the correct shape would be magnified rapidly enough for the conditions for continuing computation to be violated, so the Big Crunch would happen after only a finite number of computational steps. Therefore, to satisfy the Turing principle and attain an omega point, the universe would have to be continually ‘steered’ back onto the right trajectories. Tipler has shown in principle how this could be done, by manipulating the gravitational field over the whole of space. Presumably (again we would need a quantum theory of gravity to know for sure), the technology used for the stabilizing mechanisms, and for storing information, would have to be continually improved — indeed, improved an infinite number of times — as the density and stresses became ever higher without limit. This would require the continual creation of new knowledge, which, Popperian epistemology tells us, requires the presence of rational criticism and thus of intelligent entities. We have therefore inferred, just from the Turing principle and some other independently justifiable assumptions, that intelligence will survive, and knowledge will continue to be created, until the end of the universe.

The stabilization procedures, and the accompanying knowledge-creation processes, will all have to be increasingly rapid until, in the final frenzy, an infinite amount of both occur in a finite time. We know of no reason why the physical resources should not be available to do this, but one might wonder why the inhabitants should bother to go to so much trouble. Why should they continue so carefully to steer the gravitational oscillations during, say, the last second of the universe? If you have only one second left to live, why not just sit back and take it easy at last? But of course, that is a misrepresentation of the situation. It could hardly be a bigger misrepresentation. For these people’s minds will be running as computer programs in computers whose physical speed is increasing without limit. Their thoughts will, like ours, be virtual-reality renderings performed by these computers. It is true that at the end of that final second the whole sophisticated mechanism will be destroyed. But we know that the subjective duration of a virtual-reality experience is determined not by the elapsed time, but by the computations that are performed in that time. In an infinite number of computational steps there is time for an infinite number of thoughts — plenty of time for the thinkers to place themselves into any virtual-reality environment they like, and to experience it for however long they like. If they tire of it, they can switch to any other environment, or to any number of other environments they care to design. Subjectively, they will not be at the final stages of their lives but at the very beginning. They will be in no hurry, for subjectively they will live for ever. With one second, or one microsecond, to go, they will still have ‘all the time in the world’ to do more, experience more, create more — infinitely more — than anyone in the multiverse will ever have done before then. So there is every incentive for them to devote their attention to managing their resources. In doing so they are merely preparing for their own future, an open, infinite future of which they will be in full control and on which, at any particular time, they will be only just embarking.

We may hope that the intelligence at the omega point will consist of our descendants. That is to say, of our intellectual descendants, since our present physical forms could not survive near the omega point. At some stage human beings would have to transfer the computer programs that are their minds into more robust hardware. Indeed, this will eventually have to be done an infinite number of times.

The mechanics of ‘steering’ the universe to the omega point require actions to be taken throughout space. It follows that intelligence will have to spread all over the universe in time to make the first necessary adjustments. This is one of a series of deadlines that Tipler has shown we should have to meet — and he has shown that meeting each of them is, to the best of our present knowledge, physically possible. The first deadline is (as I remarked in Chapter 8) about five billion years from now when the Sun will, if left to its own devices, become a red giant star and wipe us out. We must learn to control or abandon the Sun before then. Then we must colonize our Galaxy, then the local cluster of galaxies, and then the whole universe. We must do each of these things soon enough to meet the corresponding deadline but we must not advance so quickly that we use up all the necessary resources before we have developed the next level of technology.

I say ‘we must’ do all this, but that is only on the assumption that it is we who are the ancestors of the intelligence that will exist at the omega point. We need not play this role if we do not want to. If we choose not to, and the Turing principle is true, then we can be sure that someone else (presumably some extraterrestrial intelligence) will.

Meanwhile, in parallel universes, our counterparts are making the same choices. Will they all succeed? Or, to put that another way, will someone necessarily succeed in creating an omega point in our universe? This depends on the fine detail of the Turing principle. It says that a universal computer is physically possible, and ‘possible’ usually means ‘actual in this or some other universe’. Does the principle require a universal computer to be built in all universes, or only in some — or perhaps in ‘most’? We do not yet understand the principle well enough to decide. Some principles of physics, such as the principle of the conservation of energy, hold only over a group of universes and may under some circumstances be violated in individual universes. Others, such as the principle of the conservation of charge, hold strictly in every universe. The two simplest forms of the Turing principle would be:

(1) there is a universal computer in all universes; or

(2) there is a universal computer in at least some universes.

The ‘all universes’ version seems too strong to express the intuitive idea that such a computer is physically possible. But ‘at least some universes’ seems too weak since, on the face of it, if universality holds only in very few universes then it loses its explanatory power. But a ‘most universes’ version would require the principle to specify a particular percentage, say 85 per cent, which seems very implausible. (There are no ‘natural’ constants in physics, goes the maxim, except zero, one and infinity.) Therefore Tipler in effect opts for ‘all universes’, and I agree that this is the most natural choice, given what little we know.

That is all that the omega-point theory — or, rather, the scientific component I am defending — has to say. One can reach the same conclusion from several different starting-points in three of the four strands. One of them is the epistemological principle that reality is comprehensible. That principle too is independently justifiable in so far as it underlies Popperian epistemology. But its existing formulations are all too vague for categorical conclusions about, say, the unboundedness of physical representations of knowledge, to be drawn from it. That is why I prefer not to postulate it directly, but to infer it from the Turing principle. (This is another example of the greater explanatory power that is available when one considers the four strands as being jointly fundamental.) Tipler himself relies either on the postulate that life will continue for ever, or on the postulate that information processing will continue for ever. From our present perspective, neither of these postulates seems fundamental. The advantage of the Turing principle is that it is already, for reasons quite independent of cosmology, regarded as a fundamental principle of nature — admittedly not always in this strong form, but I have argued that the strong form is necessary if the principle is to be integrated into physics. {1}