Abstract

Newton claimed his laws of motion are certainly true, and yet his justification was surprisingly weak: he merely cited a handful of experiments and the ‘agreement of mathematicians’. Surely then these laws are probable at best. I examine the experimental evidence Newton provided and argue that, while this evidence gives strong support for the laws in limited cases, and justifies their use in Newton’s mathematical system, it does not justify such strong epistemic claims. In modern Bayesian terms, we might say that Newton’s laws merit high subjective priors. This does not make them certain. I then suggest that Newton’s laws earn epistemic warrant in another way: via a process I call ‘epistemic amplification’. On this account, Newton’s laws, as the axioms of the theory, gain epistemic status by virtue of the theory’s success. In some places, this looks like straightforward confirmation: since the motions of the planets confirm Newton’s theory, they must also confirm the laws. But in other cases, Newton’s mathematical model seems to provide a crucial test of the laws. I sketch an account of this notion of epistemic gain. I then draw some conclusions about Newton’s methodology. While my account offers some vindication for Newton’s grand epistemic claims (not so far as ‘certainty’, however), it contradicts his own methodological statements. In short, the case highlights a key difference between Newton’s method of ‘deduction from phenomena’ and the popular hypothetico-deductive method.