uniformity of nature a state of affairs thought to be required if induction is to be justified. For example, inductively strong arguments, such as ‘The sun has risen every day in the past; therefore, the sun will rise tomorrow’, are thought to presuppose that nature is uniform in the sense that the future will resemble the past, in this case with respect to the diurnal cycle.
The Scottish empiricist Hume was the first to make explicit that the uniformity of nature is a substantial assumption in inductive reasoning. Hume argued that, because the belief that the future will resemble the past cannot be grounded in experience – for the future is as yet unobserved – induction cannot be rationally justified; appeal to it in defense of induction is either question-begging or illicitly metaphysical. Francis Bacon’s ‘induction by enumeration’ and J. S. Mill’s ‘five methods of experimental inquiry’ presuppose that nature is uniform. Whewell appealed to the uniformity of nature in order to account for the ‘consilience of inductions,’ the tendency of a hypothesis to explain data different from those it was originally introduced to explain. For reasons similar to Hume’s, Popper holds that our belief in the uniformity of nature is a matter of faith. Reichenbach held that although this belief cannot be justified in advance of any instance of inductive reasoning, its presupposition is vindicated by successful inductions.
It has proved difficult to formulate a philosophical statement of the uniformity of nature that is both coherent and informative. It appears contradictory to say that nature is uniform in all respects, because inductive inferences always mark differences of some sort (e.g., from present to future, from observed to unobserved, etc.), and it seems trivial to say that nature is uniform in some respects, because any two states of nature, no matter how different, will be similar in some respect.
Not all observed regularities in the world (or in data) are taken to support successful inductive reasoning; not all uniformities are, to use Goodman’s term, ‘projectible.’ Philosophers of science have therefore proposed various rules of projectibility, involving such notions as simplicity and explanatory power, in an attempt to distinguish those observed patterns that support successful inductions (and thus are taken to represent genuine causal relations) from those that are accidental or spurious. See also CAUSATION, GRUE PARADOX, LAW- LIKE GENERALIZATION , PROBLEM OF INDUC — TIO. J.D.T.