The simulation argument
Nick Bostrom, Are We Living in a Computer Simulation? (2003) · also simulation hypothesis, Bostrom’s simulation argument, simulation theory
Suppose a civilisation could one day run vast numbers of detailed simulations of minds like ours. Then simulated minds would vastly outnumber biological ones, and nothing in your experience would tell you which kind you are.
What it tests If advanced civilisations would run many simulations of their ancestors, should you believe you are probably living in one?
The case
Imagine that some civilisation reaches a “posthuman” stage, with computing power so great that it can run detailed simulations of its ancestors, simulations fine-grained enough that the simulated people have minds and experiences like ours. If it ran many such simulations, the number of simulated minds would far exceed the number of biological ones that ever lived. And a simulated mind’s experience would be just like yours.
Nick Bostrom’s 2003 paper, published as “Are We Living in a Computer Simulation?”, turned this into an argument.
The argument
Bostrom concludes that at least one of these three propositions is true:
- The fraction of human-level civilisations that reach a posthuman stage is very close to zero: almost all go extinct first.
- The fraction of posthuman civilisations that run significant numbers of ancestor simulations is very close to zero: almost none choose to.
- The fraction of all people with experiences like ours who live in a simulation is very close to one: we are almost certainly simulated.
The reasoning for the third horn relies on a principle of indifference: if almost all minds with experiences like yours are simulated, and you have no evidence about which kind you are, you should believe you are probably one of the many rather than one of the few.
Argument versus hypothesis
The simulation argument is often confused with the simulation hypothesis. The hypothesis is the claim that we are simulated. The argument does not assert it; it says the hypothesis is likely true unless civilisations almost never get there or almost never bother. Which horn to accept depends on assumptions about the future of technology, about what would motivate posthumans, and about whether computation can produce consciousness at all.
The objections
Can computation produce minds? The argument assumes that a simulation of a brain, run on a computer, would be conscious. If the Chinese room or the zombie argument is right that functional organisation does not guarantee experience, simulated people might not have experiences like ours, and the reasoning stalls.
The reference class. The principle of indifference asks you to treat yourself as a random sample from all observers with experiences like yours. Which observers belong in that class, and whether you can reason this way about yourself, is contested in the theory of probability.
Feasibility. Critics question whether the computing resources required are physically possible, or whether a simulation could be detailed enough without simulating the whole universe.
What would it change? David Chalmers argues in Reality+ (2022) that even if we are simulated, the objects around us are real, digital objects, so the hypothesis is not a sceptical one: most of our beliefs about our world would still be true.
How it relates to other cases
The simulation argument is the probabilistic heir to Descartes’s evil demon and the brain in a vat. Those cases ask whether you can rule out that your experience is artificially produced; Bostrom’s asks whether, given certain assumptions, you should think it likely. Its oldest ancestor is Plato’s cave, where prisoners mistake shadows for the whole of reality.
Terms
Questions people ask
What is the simulation argument?
Nick Bostrom’s 2003 argument that at least one of three claims is true: almost all civilisations like ours go extinct before they can run detailed simulations of minds; almost none that could run such simulations choose to; or we are almost certainly living in a simulation. It does not say which of the three is true.
Is the simulation argument the same as the simulation hypothesis?
No. The simulation hypothesis is the claim that we are living in a simulation. The simulation argument is Bostrom’s trilemma, which concludes only that either that hypothesis is very probably true or one of two other claims about the future of civilisations is. Bostrom himself has not claimed that we are simulated.
Has the simulation hypothesis been debunked?
No, and it is hard to see how it could be refuted or confirmed from the inside, since any evidence would itself be part of the simulation. Critics instead attack the argument’s premises: that consciousness can be simulated by computation at all, that such simulations are feasible, and that you should reason as if you were a random sample of all minds.
How is the simulation argument different from the brain in a vat?
The brain in a vat raises a sceptical doubt: can you rule out that your experience is generated artificially? The simulation argument goes further, offering a probabilistic reason to think that, under certain assumptions, it is likely that it is.
Sources and further reading
- Bostrom, N. (2003). Are we living in a computer simulation? Philosophical Quarterly 53(211): 243–255.doi.org/10.1111/1467-9213.00309
- Chalmers, D. J. (2022). Reality+: Virtual Worlds and the Problems of Philosophy. New York: W. W. Norton.
- The Simulation Argument (Bostrom''s site)simulation-argument.com/