relation to it. This is the direction in which original research has been tending but has not wholly attained, owing to its attachment to the false conceptions of historical "facts" and groups of facts, and to its dependence upon the source material for form as well as for matter.
In short, the historian should take the measure of everything he has to deal with, just as the scientist takes into account every factor affecting his experiment. That he can not measure as accurately and completely as the chemist or the physicist is the very worst reason for his not measuring at all. If periods, movements and human motives are so uncertain—as historians sometimes tell us by way of excuse—how can one venture to assume them instead of trying to remove to some extent that uncertainty or of frankly recording it as an element of error in the investigation? Much could be measured that never has been. While some of the researches in which historians are engaged give no promise of sure results and offer rather a broad choice of plausible or ingenious theories, extensive past literatures teeming with human prejudices, motives and ideas are waiting for accurate measurement and estimation. Were extant Greek literature, for instance, sifted thoroughly and its utterances on different topics collated and evaluated, many existing notions about Greek civilization might be modified, many new truths about it and its relations to the culture of other periods might be revealed. Or at least we should gain firm ground to stand on. For while classical scholarship is intent on minute points of ancient customs and language, for ancient ideas we have no statistics, only opinions. Such opinions have been formed no doubt as a result of acquaintance with the source material, but that fact alone is not enough. Science is not satisfied by a sort of alchemistic process in which various ingredients of source material are thrown by the historian into the seething kettle of his intellect, whence, after long subjection to the fires of unconscious cerebration and the juices of ripe reflection, they are supposed to emerge fused and transmuted into historical truth. There must be a complete and-accurate analysis and measurement of that material and a sound process of deducing historical truth therefrom. This may be illustrated in more detail.
The necessity of qualitative measurement of statements not only per se, but with reference to the sources from which they are drawn, is generally recognized: but historians as yet seldom heed the twin requirement of quantitative measurement of our data in comparison with the sources whence they come. It is well understood that one must take into account the reliability of the source from which the statement comes, the circumstances under which it was written, the attitude of its author—whether superstitious or sceptical, rhetorical or sober, gossipy or official, spiteful or eulogistic, contemporary or hearsay, doctrinaire or unconscious and objective. On the other hand, while the scarcity of source material is often noticed, the amount of relevant material is seldom discounted to affect the worth of a particular point supported by
