There’s a lot of ‘big talk’ going around suggesting that our universe may be some sort of computer simulation. The problem with taking such an idea seriously is that most of this, ‘big talk’ is undermined by, ‘little science’. In this article I take a more balanced view based upon a logical analysis and our understanding of computational sciences to answer some of the basic questions surrounding this thought provoking idea.
A Simulation of What Exactly?
The problem is that whenever we think of a computer we imagine some sort of monolithic box sitting on a desk somewhere when in reality a computer is anything that computes and its structure and mechanism can be anything that does the job. Our physics tells us that the universe consists of the interaction between energy and matter, but we don’t actually know what these are, or for that matter what time, or space are. Also physics has reached a point where we are investigating things at such small scales that the shear complex behaviour of what’s going on means that individual events are inherently unpredictable and we can only take a gross statistical view and this clouds what lies beneath. This means that with all the physics we have we still have no idea what the nature of our universe’s sub-structure is, or indeed if there is one. So there seems no reason why the fundamental substrate cannot be inherently computational and this would mean our existence has nothing to do with little green programmers in another universe.
If however the universe is generated by a computational process then it is likely to be a very poor simulation of the host universe because, like us, their universe is the only one they know and hence they would base a simulation on elements of their own universe. It would be an extremely poor simulation because by definition all the information contained in our universe would be a piddling small portion of that in the host universe (as we are a tiny part of it). The vast increase in information in the host universe over ours would likely mean it is hugely more complex than ours with maybe more dimensions and completely different physics etc. Given this we have absolutely no chance of comprehending the structure of the computational architecture that generates our universe. The best we can do is what the little green programmers may be doing, which is to create a very basic simulation that generates some sort of recognisable physics and conscious patterns within it. If they can do this which they obviously have, then they can surmise that they also could be generated in the same manner.
What Does the Program Look Like?
Well for starters it is very unlikely to be what is called a functional based program which is the type of programming that we experience every day in our apps and games. In general a functional program would have to program every rule, every type of fundamental object and the logic for every possible interaction between them. Basically if we haven’t programmed the construct of a tree into the simulation then it will not appear. Programming a simulation like this is of course impossible, unless your physics is totally accurate and totally complete and you can write such a vastly complex program. The other big problem with this approach is numbers. We have horrible numbers like the seemingly infinite pi ( 3.141592…). Doing calculations with these numbers would always produce tiny errors because a computer needs infinite bits to fully represent an infinitely long number. These tiny numerical errors would quickly generate huge instability in our simulations and they would fall apart.
The solution to a simulation is more likely to be a very simple rule based program that acts upon simple objects recursively to create ever more complex dynamic patterns. We know from the scientific study of complexity that such complex behaviour can cause patterns of behaviour to just appear, or emerge seemingly out of nowhere. Most researchers including myself are investigating the application of simple computational objects called cellular automata and if you want to see just how powerful they can be then search the web for, ‘Conway’s game of life’. This is a very simple cellular automata that can produce stable or pulsating patterns and patterns that move and eat other patterns and that also emit other patterns. All this behaviour is generated from a simple program that has four very simple rules.
If the program was a cellular automaton then our computational substrate would look like a network of nodes with each node being an independent processing unit whose state not only depends upon its own current state, but that of all its neighbouring nodes. This dynamic network would generate the universe including the emergent laws of physics that we perceive from inside the network, and all this without ever programming any of the laws of physics into the system.
In my own research these nodes have structure inside them that also computes and as we are a pattern of nodes we can never comprehend the structure within a node, although through observation we may create a model which includes lots of attributes of such nodes like spin and quantum number, charge and such like.
To us this network of nodes would look like a continuous space in which our pattern transfers through the network in a coherent manner. We could then imagine the structure within a single node containing what we would see as vast amounts of energy in an infinitesimal space which we may call a ‘singularity’. To the program this ‘energy’ is in fact just a pattern or the integer number of states in a pattern and it is this number of states that interacts with the rules to generate a change in pattern that we attribute to some characteristic of energy. To me this is one of the appealing aspects about this proposition; that it actually can make many of the incomprehensible ideas in physics make sense.
What all this would also mean is that physics is not fundamental, but that there are laws below it that generate the physics. This may be why physicists have for the last few decades been promising a grand theory of everything without very much success, because they are trying to use something that isn’t fundamental to explain something that is.
Are there an Infinite Number of these Simulated Universes?
No! As I said before each universe that is simulated from the previous one must have less information in it and therefore we would get to a point where the host universe is not complex enough (not enough physics) to generate a simulated universe that has enough complexity to be interesting. So we would find that each simulated universe becomes less complex. An interesting point is that if we are at the end of the line then although we may be able to create some kind of simulated universe, it may not be complex enough to have a lot physics that is similar to our own, or create intelligent patterns. In this situation we may decide that we are unlikely to be a simulation ourselves when in fact we are, just not a very complex one.
How can we Prove it?
I’m not convinced that we can and I’m not sure it’s easy for the little green programmers to help prove it to us either. We are in essence a pattern of information contained in billions of network nodes. So the idea of having a, ‘programmer’s back door’ to let them enter out universe is rubbish because they would have to control a highly complex pattern at the nodal level and complexity science tells us that at this level the state of an individual node is completely indeterminate. This may be why we perceive quantum mechanics because we are trying to apply science at a fundamentally indeterminate scale.
So once this program has been kicked off the little green programmers would be unlikely to be able to intervene because the vast complexity would mean that any external change they made would have completely unpredictable consequences; this is most certainly true of any simulation that we created. Therefore we are unlikely to be play things, but more likely to be a highly important individual experiment.
However I have often thought that if we could produce a successful simulation on a vast scale and suddenly identify propagating patterns that we could infer were generated by intelligent patterns then the rest of the simulated universe becomes a waste of computational effort. In this case we may be able to replace the complex and vast patterns of nodes that are far off stars or galaxies with simpler functional programed objects that ‘simulate’ the behaviour of the emergent rules, or physics. So maybe in the far distant future we will send a satellite to a distant star and watch in amazement as a completely unexpected reaction occurs when our nodal pattern interacts directly with one of these functional objects. This would be as fundamental as finding out that the moon really is made of cheese.
The other option along these lines is that the little green programmers could generate specific and unexpected behaviour in astronomical objects that are far enough away from us to have no serious influence on the earth, other than us noticing what was a pulsar now blinking in some sort of inexplicable but repeating mathematical sequence that is a code for, “hi we are the little green programmers”.
The primary problem with us proving for ourselves that the universe is a computation is that our physics is based upon our interactions from within the network and so we have natural limits on our ability to uncover structure at nodal scales. This limit may equate to what physicists’ term the Plank scale; the scale at which our model of the universe using quantum mechanics breaks down and unfortunately this is the scale we need to investigate to see the computational processes.
Is Anyone Taking this Seriously?
Yes there is a small group of prestigious and highly talented scientists who believe that this is a worthy area of study. The list includes Professor Edward Fredkin, Stephen Wolfram, Nick Bostrom, Tommaso Toffoli and Norman Margolus. I obviously don’t include myself on the list but for what it’s worth I believe we all share the view that there is currently little science to offer in support of this idea, however there is enough of a strong scientific argument to make it a valid proposition that needs serious consideration.
Finally, one other scientist who did take this proposition seriously was one of the greatest minds of the twentieth century, Richard P. Feynman, a Nobel winner for physics, who wrote (with my comments in brackets), “So I have often made the hypothesis that ultimately physics will not require a mathematical statement (it’s not fundamental but emergent), that in the end the machinery (maybe a computer) will be revealed, and the laws will turn out to be simple, like the chequer board (like a cellular automata) with all its apparent complexities (simple rules generating complex physics)”. I feel neither talented nor qualified enough to follow that.