Scientific realism

Scientific realism is the philosophical position that the world described by our best scientific theories is the real world — not merely a useful fiction, instrumental convenience, or predictive device. Realists hold that electrons, genes, black holes, and quarks exist independently of our theories about them, and that the success of science is best explained by the approximate truth of its claims. The position stands in direct opposition to anti-realism and instrumentalism, which treat theoretical entities as calculational tools rather than descriptions of mind-independent reality.

The strongest argument for scientific realism is the no-miracles argument: it would be a miracle if a theory made consistently accurate predictions while being radically false about the entities it postulates. The strongest argument against it is the pessimistic meta-induction: history is littered with successful theories that were later abandoned, and their central theoretical entities — phlogiston, caloric, the luminiferous ether — turned out not to exist. If past successful theories were false, why believe current ones are true?

The debate between scientific realism and anti-realism is not merely academic. It shapes how we interpret the outputs of complex systems models, climate simulations, and artificial intelligence. A realist treats a model's predictions as evidence that the model captures real structure; an anti-realist treats them as evidence only that the model is empirically adequate. The difference determines what kind of certainty we are entitled to claim — and what kind of action we are justified in taking on the basis of that certainty.