I have a deviating understanding of the simulation hypothesis. If I look at our universe as simulation, then it seems to me irrelevant whether this simulation is actually executed or not. In my understanding of simulation hypothesis, we exist because we can be simulated; not because we are simulated.
My argumentation:

Say our universe can be described by a finite set of rules and initial states. Then every possible future sequence of states of the universe can in principle be calculated. Even if some rules are stochastic, something can in theory run a simulation long enough to cause any possible sequence of states.

Then what makes our universe real? Is it real because it is executed on some mysterious "real" computing device? If we run our universe on more real computers, will it be more real? But from the viewpoint of the universe, it makes no difference if it is executed on 1, 2, 3, ..., N computing devices. No sequence of states of the universe depends on this information of how many computing devices it runs on.

Then what if N=0? If the universe is not executed on a single computing device, then does it not exist?
Executing the universe is something that makes the universe observable by someone running the simulation. But is an execution needed for the sole purpose of the universe to exist?
Is it possible that every possible universe must exist simply because it theoretically can exist?
Can the simulation of our universe exist without a simulator?

Am I right to assume that if we accept the simulation hypothesis we must accept this 'simulation without simulator hypothesis'?

  • 1
    Can this assumption be wrong? "Say, your universe can be described by finite set of rules and initial state." Physics is an approximation of nature; who says the universe can be exactly described by a finite set of rules? Is your argument valid if that assumption is wrong?
    – nir
    Commented Jan 30, 2015 at 15:33
  • 1
    @nir: Good question. Maybe that assumption is not needed. Classical simulation hypothesis needs this assumption but simulation without simulator = simulation not limited by finitelly fast simulator. :) Commented Jan 30, 2015 at 19:34
  • 1
    @user4894: If we are part of simulation then we are part of generated output. And we see output of simulation everywhere and call it Universe. We think that output is what must exist in some form. But that can be ilusion and what matters is only description how to generate that output. Output can be in principle discarded and reregenerated infinitelly many times from initial state and set of rules. If description how to generate output is logically consistent and computable, then its output (our universe) must exist even if does not exist actual computing device capable of doing it. Commented Jan 30, 2015 at 20:40
  • 1
    @user4894: Its unintuitive. But be able to observe existence of something is not the same thing as existence of something. With executing Microsoft Word we are making its output observerable to us and we know it exists. But opositely we simply cannot say that thing we cannot observe does not exist - we only expecting that. When no outside state depends on existence or not existence of running Microsoft Word - say, until you look at it - menu bar can exist. Commented Jan 30, 2015 at 21:17
  • 2
    Yes, its nonsense to common sense. We evolved to think like that. Today we know that at very deep level universe behaves in weird way. I think truth about existence of universe will be some sort of this weirdness. I think i explained it. :-) Commented Jan 30, 2015 at 21:38

4 Answers 4


If you are going to weave together simulations with nihilism and zero, expect to have to define very strict meanings for several terms:

  • Simulation
  • Simulator
  • Execute
  • Zero
  • One
  • Finite
  • Real
  • Exists

Your theory does balloon quickly into "every possible universe exists, because none of them exist," which isn't inherently wrong, per se, but it does limit its usefulness as a descriptive theory unless those strict definitions produce some very interesting side effects.

As a starting point, consider that "the natural numbers" starts from 1, not 0. That doesn't mean it can't start from 0, but it does mean there is some qualitative difference between the starting points that have meaning to human philosophers.

  • 2
    Zero is the most unnatural thing there is. (isn't) (whatever)
    – user16869
    Commented Apr 24, 2016 at 15:53

This idea of 'simulation' without simulator is worth exploring. Thinking along the lines proposed, one is quickly confronted with the notion of 'real' and what supports (if anything) what is 'real'? I suspect it is the notion of 'real' that is causing us difficulty.

I think the notion of 'simulation' without simulator has some merit and it's easy to think of an example.

1) Imagine -- for instance -- an infinite (or large 'enough') GAME OF LIFE GRID that has been initilized with a set of 'ON' cells that corresponds to a 'computer' (we know these can be built on the LIFE GRID) that 'simulates' the mechanics and all components of a 'universe'. We can make the LIFE GRID as large as we like so that memory is inexhaustible. Last step: WE NEED NOT ACTUALL BUILD SUCH A GRID INITIALIZED AS DESCRIBED -- we need only acknowledge that such a grid-initialization exists (Platonic). To wit: somewhere in the GAME OF LIFE UNIVERSE there IS a grid that corresponds to a computing device that simulates a universe -- and perhaps one complete with thinking creatures. There is no simulator -- there is simply a Platonic Form (so to speak) that corresponds to something we identify with a universe with various features. No creator. No simulation. Just the logical unpacking of rigid logic on the LIFE grid.

2) Same as (1) but said LIFE computer is a simulation of a single MIND and is fed a tape that gives 'it' (the sentience simulated) the experience of being a 'person' (See S Lem's short story addressing this scenario in fiction).

3) Forget the LIFE grid for a moment -- imagine a cellular automata (N dimensional, infinite or very very large) of sufficient complexity that when randomly initialized (or very specifically initialized perhaps) YIELDS complex, sustained objects (atoms, chemistry, stars, etc -- emergent structures). This happens naturally on the LIFE grid, it's easy to imagine (not so easy to create from scratch) automata with rules that yield sufficient macroscale (big chunks of the grid, perhaps the grid cells are Planck Scale) structure to support molecular formation, evolution, etc. All of this from an automata that DOESN'T EXIST IN SOMEONE'S COMPUTER -- ie, an automata from the set of ALL automata -- a Platonic item. This (as with (1) and (2)) eliminates a creator/simulator and any need for an external computing device. This hypothesis is appealing as it immediately manages some inscrutables: apparent need for a creator, paradox of 'real', an unreasonably 'large'/infinite universe.

  • 2
    Thanks for the answer. All caps feels like SHOUTING and makes for an uncomfortable read. Can you update your answer using one of the other formatting options like italic or bold?
    – jeroenk
    Commented Aug 4, 2015 at 13:39
  • "It's all an illusion. It's all... a tape." - from Mulholland Drive
    – user16869
    Commented Apr 24, 2016 at 15:58

Suppose you have genuine random elements in the simulation. Then every instantation will realise a different history which cannot be predicted beforehand. So unless you regard the entire ensemble of possible unfoldings of history as your universe you really have to run the simulation to see what you get.

Running the simulation on multiple devices will give different universes, though with common features implicit in the models.

  • It does not matter what value is N (number of instances). Value of N is not information that distinguishes betwen existing and nonexisting universes. If N will be infinite then with infinite number of instances of randomly driven universes there will also by infinite amount of duplicit instances. Commented Jan 30, 2015 at 19:48
  • @user3123061: You obviously understand the statistics of existence of simulated universes so much better than I do I must bow to you argument (whatever it was) Commented Jan 30, 2015 at 19:52
  • Absolutely no. I am amateur philosopher and programmer. Of course i can be wrong and probably i am wrong. It will be nice if someone explain that what i say is nonsense. :-) Commented Jan 30, 2015 at 19:56
  • "Music is a kind of nonsense that everyone understands." (I can't locate the source right now, but it was some famous person)
    – user16869
    Commented Apr 24, 2016 at 16:02

Not only can you not have a simulation without a simulator, but also without something to simulate.

Take for example a painter who makes a painting of a tree. When you look at his canvas you see a picture of a tree, a simulation of tree.

But how could he paint a tree if there was no actually existing tree to paint?

The point I’m making is that the notion of simulation is conceptually incoherent and more likely signifies a kind of philosophical fantasia that says more about the people promulgating such an idea and the times that we live in than it actually says anything about the world.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .