[78] Scientific Method in Philosophy, p. 57.
[79] The analytic realists have shown a peculiar disinclination to discuss the nature of future consequences as terms of propositions. They certainly are not identical with the mental act of referring to them; they are "objective" to it. Do they, therefore, already subsist in some realm of subsistence? Or is subsistence but a name for the fact of logical reference, leaving the determination of the meaning of "subsistence" dependent upon a determination of the meaning of "logical"? More generally, what is the position of analytic realism about the future?
[80] Supposing the question to be that of some molten state of the earth in past geologic ages. Taken as the complete subject-matter of a proposition—or science—the facts discovered cannot be regarded as causative of, or a mechanism of, the appearance of life. For by definition they form a closed system; to introduce reference to a future event is to deny the definition. Contrariwise, a statement of that past condition of the earth as a mechanical condition of the later emergence of life means that that past stage is taken not merely as past, but as in process of transition to its future, as in process of alteration in the direction of life. Change in this direction is an integral part of a statement of the early stage of the earth's history. A purely geologic statement may be quite accurate in its own universe of discourse and yet quite incomplete and hence inaccurate in another universe of discourse. That is to say, a geologist's propositions may accurately set forth a prior state of things, while ignoring any reference to a later state entailed by them. But a would-be philosophy may not ignore the implied future.
[81] Philosophical Essays, pp. 104, 105.
[82] Sixth Meditation.
[83] Principles of Philosophy, p. 90.
[84] Treatise of Human Nature, Part III, sec. iii.
[85] It is perhaps poor tactics on my part to complicate this matter with anything else. But it is evident that "passions" and pains and pleasures may be used as evidences of something beyond themselves (as may the fact of being more than five feet high) and so get a representative or cognitive status. Is there not also a prima facie presumption that all sensory qualities are of themselves bare existences or occurrences without cognitive pretension, and that they acquire the latter status as signs or evidence of something else? Epistemological idealists or realists who admit the non-cognitive character of pleasure and pain would seem to be under special obligations carefully to consider the thesis of the non-cognitive nature of all sensory qualities except as they are employed as indications or indexes of some other thing. This recognition frees logic from the epistemological discussion of secondary qualities.
[86] To readers who have grasped the thought of my argument, it may not be meaningless to say that the typical idealistic fallacy is to import into the direct experience the results of the intellectual or reflective examination, while that of realism is to treat the reflective operation as dealing with precisely the same subject-matter as the original act was concerned with—taking the good of "reason" and the good of immediate behavior to be the same sort of things. And both fallacies will result from any assimilation of two different acts to one another through giving them both the title "knowledge," and hence treating the difference between them as simply the difference between a direct apprehension and a mediated one.
[87] Analytic realism ought to be favorable to such a hedonism; the fact that present-day analytic realists are not favorable would seem to indicate that they have not taken their logic seriously enough, but have been restrained, by practical motives, from applying it thoroughly. To say that the moral life presents a high degree of organization and integration is to say something which is true, but is also to say something which by the analytic logic calls for its resolution into ultimate and independent simples. Unless they accept the pleasures and pains of Bentham as such ultimates, they are bound to present acceptable substitutes. But here they tend to shift their logic and to make the fulfilment of some organization (variously defined) the standard good. Consistency would then admit the hypothesis that in all cases an eventual organization rather than antecedent simples supply the standard of knowledge. Meanwhile the term "fulfilment" (or any similar term) stands as an acknowledgment that the organization in question is not something ontologically prior but is one yet to be achieved.