Mind & behavior
The simulation argument is more complicated than 'we live in a computer'
Bostrom's argument links three possibilities about civilization and simulated observers. It does not choose one of them by discovering a flaw in reality.
Three doors, not one conclusion
Nick Bostrom's 2003 simulation argument is often compressed into the claim that we probably live inside a computer. The paper's actual structure is a three-way proposition. Either civilizations like ours rarely reach a sufficiently advanced stage, or advanced civilizations rarely run many ancestor simulations, or observers with experiences like ours are overwhelmingly simulated.
The conclusion depends on assumptions developed in the paper. It is not a report of an experiment detecting simulated space, a computer outside the universe or an error in physical reality. Its force comes from asking whether several beliefs about the future can consistently be held together.
Counting observers requires assumptions
One assumption concerns consciousness: a suitable computational implementation could support subjective experience. Another concerns scale: a mature civilization might possess enough computing capacity to create large numbers of simulated observers. A further step concerns how an observer should reason about belonging to the resulting population.
A simple counting analogy helps isolate the issue. If a container held a million indistinguishable simulated tickets and one original ticket, a randomly selected ticket would probably be simulated. But human observers are not literally tickets drawn from a known container. The philosophical work lies in justifying the relevant population and the rule for reasoning about one's place in it.
Accepting the argument is not selecting the hypothesis
Bostrom's 2025 FAQ explicitly distinguishes accepting the argument from accepting the simulation hypothesis. One can regard the three-way reasoning as sound while assigning substantial weight to a future in which advanced civilizations do not run such simulations. The author also declines to offer a falsely precise personal probability.
That clarification blocks a common shortcut. Rapid improvements in ordinary computing do not independently establish simulated consciousness, future civilization survival, the motives of advanced societies or the right observer-counting principle. Evidence relevant to one premise cannot automatically settle the others. The conditional chain matters more than the dramatic final possibility.
What would disagreement concern?
A careful criticism can challenge the consciousness assumption, the feasibility estimates or the move from numbers of observers to personal probability. These are distinguishable objections. Saying simply that the conclusion sounds implausible leaves the actual structure untouched, while claiming that large computers make it inevitable skips the disputed steps.
The paper is valuable as philosophy because it makes those commitments visible. It can be examined without treating everyday coincidences as hidden messages or regarding other people as unreal. Whether a world were simulated would not by itself make the experiences of its inhabitants unimportant. The archive's role is to preserve the argument in a form that allows its premises to be questioned, rather than turn a carefully conditional proposal into a declaration about the reader's reality.
Sources and further reading
- Are You Living in a Computer Simulation? ↗
Philosophical Quarterly 53(211), 243–255; Sections II–V
- Nick Bostrom, Simulation Argument FAQ, version 2.0 (2025) ↗
Questions 1–2 and discussion of observer reasoning