Hume on Scientific Law

Extract

For many years now a “principle of uncertainty” has played a major role in all discussion of the problem of scientific law as description of nature. That this principle had its origin in the efforts of science to describe nature is entirely appropriate; that it has had so immediate an effect on philosophic thought is inevitable.

It is also inevitable that questions should be raised concerning the meaning of certainty and its reference to descriptive law. Such questions are bound to occur to one who is concerned about the kinds of law in terms of which nature might be described. If these kinds can be considered to be two, one “mechanical,” the other “statistical,” it appears to be an easy step from the assertion of a principle of uncertainty which requires that the description of elemental processes and facts of nature be statistical—in terms of probabilities, that is—to a further assertion that after all a strictly mechanical description of nature is impossible, since the observational data which constitute its material are approximate and probable, and because the laws which provide the formal element of that description can never be more than approximately or probably true. It is concluded, then, that no element of certainty attaches to the so-called mechanical laws, any more than to others that are avowedly statistical. When a prediction is based on either sort of law, it is taken to be a prediction that an event will probablyoccur.