Category: time-and-causality
Type: thought experiment
Associated thinker(s): Pierre-Simon Laplace
Original publication or approximate period: Essai philosophique sur les probabilités (1814; later editions)
Difficulty: Intermediate
Themes: determinism, prediction, laws of nature
Related experiments: The Consequence Argument; Newcomb’s Problem
Imagine an intelligence that knows the complete physical condition of the universe and all relevant forces, with enough mathematical capacity to analyze them. In Laplace’s idealized picture, the future and the past would be available to this intelligence without uncertainty.
The familiar label “demon” is a later name for the intelligence, not a claim about a supernatural creature in a laboratory. The example joins a physical assumption about lawful evolution to strong epistemic and computational idealizations. Keeping those ingredients distinct is essential: a world’s history could be fixed even if no inhabitant could calculate it.
Would complete physical information and unlimited analysis make the future predictable, and what would that imply about human action?
The picture clarifies several different claims: that events are determined, that we can predict them, and that choices do not matter. Only the first concerns whether the laws and a complete state permit more than one continuation. A decision can remain causally important even if it is determined.
Determinist idealization: for an appropriate deterministic physical model, a complete state and laws specify the subsequent evolution. The intelligence dramatizes this commitment.
Limits-of-prediction response: chaotic sensitivity, incomplete information, and computational limits can prevent practical forecasting without introducing multiple lawful futures from the exact same state.
Free-will responses: incompatibilists see determination as threatening an important kind of freedom; compatibilists argue that acting through one’s reasons and capacities can still count as free. Prediction alone does not settle that dispute.
Chaos is not synonymous with indeterminism. Determinism is not fatalism: the result can depend on what you do, even if what you do has antecedent causes. Modern physics also cannot be summarized as “quantum mechanics simply disproved determinism”; interpretations and different theories handle determination differently.
Self-referential prediction adds the problem of an agent responding to its forecast. Quantum and relativistic theories change the relevant physical assumptions. These extensions require explicit models; the nineteenth-century image should not be used as a substitute for current physical analysis.
The SEP determinism entry supports the conceptual distinctions and reproduces Laplace’s passage through a later edition. The ETH digitized edition was found to verify the 1814 publication chronology. No claim of an achievable universal computer is made.
Research check: 2026-10-30. References above were retrieved for the support described; this is an original educational explanation, not a substitute for the primary texts. Additional speculative comparisons are teaching prompts, not established historical variations.