connected with the known facts of observation. How is such a statement reached? and when it is reached, how do we know that it is right? To answer such questions fully would demand a whole treatise on scientific method, here impossible; our intention is simply to point out that an introductory understanding of scientific method, much better than none, can be gleaned by a careful second reading of Kelvin's and of the other scientific essays in this collection, with the constant effort to learn how the announced results have been attained.
Notice, first, that for an active mind, it is "impossible to avoid theorizing" (p. 281). The lesson from this is to beware of those so-called practical persons who say they do not theorize; what they really do is to theorize in an unsafe, unscientific manner; for they, like everyone else, wish to understand more than they can see. The desire to theorize should not be resisted, but theorizing should be carefully cultivated and its results should be carefully held apart from those of observation. Notice, second, that, some facts of observation having been gained, the inquisitive mind at once sets about inventing schemes that may possibly include the mental counterparts of the unseen phenomena, or facts of inference, and then proceeds to determine the correctness of the inventions by certain logical devices or tests. That particular scheme is finally adopted as true which stands all possible tests. The tests are mostly experimental in the study of light; they are largely computational in the study of the tides. Notice, third, how ingenious the scientific mind must be to conceive the extraordinary schemes by which the unseen phenomena are supposed to combine with the seen, so as to make a reasonably working total process; how far these mental processes must go beyond the mere determination of visible facts by observation; how active the imagination must be to picture the invisible processes of the invented scheme; and also how free from prepossessions, how docile the scientific mind must be, in order to follow the experimental or computational demonstrations wherever they may lead! Still more important, notice how large a share of the standard content of science, as illustrated by the essays on light and tides, is made up of what are here called "facts of inference," and not simply of facts of observation.
