Page:The Limits of Evolution (1904).djvu/145

84 by unquestionable presence in sensible experience. All the refinements of scientific method — the precautions of repeated observation, the probing subtleties of experiment, the niceties in the use of instruments of precision, the principle of reduction to mean or average, the allowing for the “personal equation,” the final casting out of the largest mean of possible errors in experiment or observation, by such methods, for instance, as that of least squares — all these refinements are for the single purpose of making it certain that our basis of evidence shall be confined to what has actually been present in the world of sense. We are to know beyond question that such and such conjunctions of events have actually been present to the senses, and precisely what it is that thus remains indisputable fact after all possible additions or misconstructions of our mere thought or fancy have been cancelled out. Such conjunctions in unquestionable experience, isolated and then purified from foreign admixture by carefully contrived experiment, we are finally to raise by generalisation into a tentative expectation of their continued recurrence in the future; tentative expectation, we say, because the empirical method in its rigour warns us that the act of generalisation is a step beyond the strict evidence, and must not be reckoned any part of science except as it continues to be verified in subsequent experience of the event