Oracles and Supercomputers

The laws of physics determine what happens in the universe. That is, they mandate everything in physical reality that occurs. It is sometimes then claimed, on the basis of this, that if only we had perfect knowledge of those laws of physics and a set of conditions at some moment in time and further a sufficiently powerful computer to do the calculation, then “in principle” we would be able to predict what happens from moment to moment - at all times in the future and the past.

But this is not so. Putting aside “in practise” where all agree this cannot be done and even putting aside that we ignore what physically might be required to build such a computer that could do the job of computing the motions of every individual particle in the universe moment to moment (including, recursively, its own behaviour which would have a model within it of the universe containing the prediction generating computer, within which there is a model of a universe containing the prediction generating computer, ad infinitum) such a prediction is not possible in principle. Indeed it would not even be possible to predict what happens moment to moment just on the Earth to any great accuracy because what happens here on Earth is determined by what happens (potentially) anywhere else (and almost  any-when) in the universe.  Here is why.

Firstly predicting the motion of all particles in the universe as a whole in principle: while we can imagine what some statement of the “final laws of physics” might entail (ignoring for a moment that the laws of epistemology in terms of our fallibility rule this out) what does it mean to have a statement of the “initial conditions” (or indeed the positions & momenta at any other time) of all particles in the universe simultaneously (which is what would be needed to make such a prediction in principle). Why? There is no simultaneous time in the universe. There is no time t, (i.e: for any value of t), where we can list all the particles in the universe next to where they are now and how fast they are moving (and what their mass is, say, among other things). So at the first hurdle, thinking that *now* we can ever have knowledge of all particles *now* in order to make a prediction just in principle now - is impossible. There is no simultaneous “now” for us on Earth even if we had the most advanced laser scanning system for locating all particles in our “frame of reference” because whatever the time is “now” for us (approximately speaking) there is no equivalent time for particles in the Andromeda galaxy - much less the “other side” of the universe. 

Ok, but what about just predicting what happens locally? Never mind the rest of the universe - let's do something perhaps a little more tractable by a few orders of magnitude where the relativity of simultaneity can be largely ignored. This prediction is also impossible in principle precisely because what happens on Earth moment to moment can be (and is) directly affected by what happens in the rest of the universe. But we have already said "what happens" in the rest of the universe cannot be known in the sense we require because there is no "what happens" at any given time across the cosmos because...there can be no "given time" that applies to all places unambiguously. 

And also, for example, if a supernova occurs on “the other side” of the universe but billions of years ago just so it is detectable tomorrow then no prediction made today about Earth can account for it. But why should that supernova, extremely dim when detected billions of years later, billions of light years from its origin and only via a handful of low energy photons reaching Earth, have any effect on what happens here even in principle? Those photons have nowhere near the momenta required to alter the trajectories of any particles here. Perhaps. But what if an astronomer detects light from that supernova tomorrow and it causes her to write a paper revising what we know about the brightness of (say) Type IA supernova and this initiates a complete rewrite of our understanding of the accelerating expansion of the universe and so on? Perhaps that astronomer earns fame altering the course of their life, and texts are rewritten altering the course of the lives of many more people and eventually a new version of The Big Bang is produced changing school and university courses across the globe - altering the course of the lives of almost everyone on the planet? None of that could possibly have been accounted for by some prediction made today based on where all the particles on Earth are now and what they are doing to perfect precision alongside the exact laws of motion. Any prediction made today will not have been able to account for those photons from just that tiny patch of sky arriving tomorrow at just the right time so it is observed by exactly that astronomer and so on. Any prediction our imaginary supercomputer (programmed with the ultimate laws of physics) makes ignores the rest what is going on far beyond the Earth in this supposedly far more reasonable version of the thought experiment. But the very reason it cannot take into account what happens elsewhere in the universe that might affect what happens on Earth in the way I suggest there is because it cannot be known what happens elsewhere to the level of precision required to make such a prediction even in principle, never mind in practise. To know what might happen in the universe that just might affect what happens here on Earth would mean knowing what has happened at all other times in the past across the universe which could, in theory, convey information a person might use as part of a hitherto unknown explanation. In other words: it would require perfect knowledge of the motions of all particles in the universe right now, which would include all photons already in transit to the Earth. But those photons cannot be detected without changing their trajectories (or better: destroying at least some of the information that would have caused a particular change here on Earth).

Sometimes writers appeal to a proverbial “oracle” (or for those who have taken physics: "Laplace's demon") who knows the laws of physics perfectly and the state of the particles in the universe at any given time and therefore can make perfectly accurate predictions about the evolution of the universe over time. Or, even just a more restricted specific prediction about the future as it applies to a tiny sliver of physical reality (never mind the evolution of the entire universe). Of course new can see the oracle must be supernatural: it’s an omniscient god of a kind. So people being rational on this point reject the possibility of such an oracle. But science minded "rationalist” types will readily substitute a “super computer of the future” in place of the oracle that can do everything the oracle can. But because it’s just called “a supercomputer” giving it a veneer of scientific plausibility this cannot save such a so-called "supercomputer" from being just another appeal to the supernatural. And it is. There is no possibility for this supercomputer in actual physical reality (even if it could be built) to ever come to possess the data it would need to perform the calculation required for making any accurate prediction of such a kind. The data needed would literally be the positions and momenta of all particles in the universe…now. But, again, there is no universal “now” in a universe governed by relativity - as our universe is known to be.

And this is to entirely ignore the impossibility of knowing the precise precision and momenta of any individual particle simultaneously anyway because of prohibitions on this information given the laws of quantum theory. All we have is a range of values - and a range of values for any set of initial conditions means in principle a range in terms of the final conditions for the "prediction". But we are not after a "range" of predictions - we are after the specific prediction: what happens to us, in our universe, at some time t in the future? But say we ignored even that complication (in brief due "Heisenberg's Uncertainty Principle") even if we did have those laws and specific individual values for each observable (and not a range) this would only in principle allow us to predict the evolution of the multiverse over time - not individual universes or individual particles or anything in between. And as we have been at pains to say: we we cannot know what we would need to know about all those particles simultaneously to make any such prediction.

But if the oracle or the omniscient god could provide a perfect prediction of how “the multiverse” (i.e: all of physical reality) looked tomorrow compared to today - any individual observer - any person in that multiverse - would not know what branch of the universe they would find themselves in. To them, such a prediction would be entirely uninformative on the question of “what happens next to me?” We don’t know what consciousness is, how creativity and personal choice works or how any of this is connected exactly to knowledge creating people and their place in the multiverse. But this is all a separate issue that merely adds to the complications here for a rational, scientifically minded “determinist”. One can be rational, scientifically minded and a determinist - but it is important for such a person to keep in mind that in physics as we understand it, “determined” does not mean “predictable” - not merely in practise, but in principle.