Thus while the impossibility of adequate experimentation, the difficulties of scientific observations of phenomena so vast in scope and so intricate in their relations, make the student of sociological subjects more dependent upon printed records for his material than is the case in most other sciences, these printed records induce a sequence of thought antagonistic to the grasp of a living and moving unity. This cause is primarily responsible for the failure of many of the ablest and subtlest economic treatises to impress upon the reader a clear conception of the industrial world as a single "going concern." Each piece of the mechanism is clearly described, and the reader is informed how it fits into the parts which are most closely related to it, but no simultaneous grasp of the mechanism as a working whole is attained. When we graft upon the idea of a mechanism that character of continuous self-development which transforms it into an "organism," the synthesis of the changing phenomena is still more difficult to comprehend. These difficulties can only be overcome by a recognition that the scientific imagination must play a larger part here than it does in those sciences whose subject-matter is more amenable to direct observation. In the latter the chief function of the imagination will be the increase of knowledge by means of hypotheses which tentatively transcend the region of known facts.
In economic science, as Cairnes has ably shown, the use of hypothesis is much wider, serving in large measure as a substitute for experiment.[2] But the scientific imagination has another constant service to perform. Its exercise is constantly required by the economist, and in general by the sociologist, to gather into true relations of time, space, and causality those intricately connected phenomena which, though individually amenable to sensuous presentation, are not able to be thus presented as an aggregate in their right organic order.
The attempts to construct a deductive economic science upon a piece-meal basis by framing special and separate theories of wages, rent, value, the functions of money, and so forth, are now recognised to be in large measure failures precisely because they involve the fundamental scientific fallacy of supposing that the several parts of an organic whole can be separately studied, and that from this study of the parts we can construct a correct idea of the whole. As in economic theory so in the comprehension of industrial history, no detailed investigation of a number of different heaps of facts laboriously collected by intellectual moles will suffice for our purpose. To understand the evolution of the system of modern industry we must apply to the heaps of bare unordered facts those principles of order which are now recognised as the widest generalisations or the most valid assumptions derivable from other sciences, and endeavour without slavish conformity to the formulæ of these other sciences to trace in the growth of industrial organisms those general laws of development which seem common to all bodies of closely-related phenomena.
