David Chalmers thinks we probably would not — perhaps in-principle — be able to determine if we’re in a simulation. Everything in it — even ‘the physical’ — would be simulated. No amount of reductionist physics could get to the bottom of the ontology and find the simulator and its structure.
It’s a pretty convincing idea, and I am not about to assert it is definitely wrong. It might be. I think there might be tools available in information theory (more specifically — Shannonian statistical and Kolmogorovian algorithmic) that allow a relevant hypotheses to be tested. One probably has to accede first to my principle that physical structure is necessary for information to obtain.
Of course if one wants to fight about the nature of the physical (and plenty of philosophers do) then that gets otiose and recondite. If you’re not willing to accede the in-principle ontic priority of structured physical processes as the necessary existential basis of any information, then I confess that I don’t know how to deal with that. Got information without Shannonian information sources (physical stochastic processes)? Yes?
Then you’ve got me, probably.
However I suggest there would perhaps still be a path from information-theoretic inference to the best explanation (IBE) and induction to detection. We don’t know of any kind of information processing system that isn’t existentially grounded in physical processes. All such systems have features or properties that are very difficult to dispense with in any universe that involves information. They have sources (Shannonian stochastic physical processes in this universe), signal transmission (or perhaps emission), and some kind of encoding (or else incidental or natural ‘encoding’.)
Fig 1. Shannon’s model of information transmission in The Mathematical Theory of Communication (1948)
One way to look at this is that if the simulation were an information processing-system then ‘we’ (our universe) would be tantamount to the signal/message transmitted through what is effectively a channel, and ‘we’ would also perhaps be a kind of receiver-destination (although I have more difficulty envisaging how the latter would apply.)
I propose that the simulation is probably necessarily an information processing system and that as such it requires the features of such systems in our universe. It is quite possible that — should the simulation hypothesis/argument correspond to reality — the fact that our own universe is replete with such Shannonian systems as a matter of natural necessity and natural nomic constraints (natural law) might be a message that tells us that the simulation works that way too.
Of course if there is a simulation, it might not be anything like any information processing system that we know of. Or perhaps there is something else apart from physical and non-physical categories — some other existential category — that we couldn’t possibly comprehend. (Maybe mathematical Platonists like Max Tegmark could comprehend it?) I just don’t know how one could get any information to exist at all without stochastic physical processes and the physical (dynamical) structures they reduce to. I don’t know how one could get such anywhere, at any level of simulation. I think that science is probably unable to determine any such thing too. But then: I am a very limited information processing system.
