Essays in Experimental Logic

by John Dewey

I

There is something a little baffling in much of the current discussion regarding the reference of ideas to facts. The not uncommon assumption is that there was a satisfactory and consistent theory of their relation in existence prior to the somewhat impertinent intrusion of a functional and practical interpretation of them. The way the instrumental logician has been turned upon by both idealist and realist is suggestive of the way in which the outsider who intervenes in a family jar is proverbially treated by both husband and wife, who manifest their unity by berating the third party.

I feel that the situation is due partly to various misapprehensions, inevitable perhaps in the first presentation of a new point of view[52] and multiplied in this instance by the coincidence of the presentation of this logical point of view with that of the larger philosophical movements, humanism and pragmatism. I wish here to undertake a summary statement of the logical view on its own account, hoping it may receive clearer understanding on its own merits.

In the first place it was (apart from the frightful confusion of logical theories) precisely the lack of an adequate and generally accepted theory of the nature of fact and idea, and of the kind of agreement or correspondence between them which constitutes the truth of the idea, that led to the development of a functional theory of logic. A brief statement of the difficulties in the traditional views may therefore be pertinent. That fruitful thinking—thought that terminates in valid knowledge—goes on in terms of the distinction of facts and judgment, and that valid knowledge is precisely genuine correspondence or agreement, of some sort, of fact and judgment, is the common and undeniable assumption. But the discussions are largely carried on in terms of an epistemological dualism, rendering the solution of the problem impossible in virtue of the very terms in which it is stated. The distinction is at once identified with that between mind and matter, consciousness and objects, the psychical and the physical, where each of these terms is supposed to refer to some fixed order of existence, a world in itself. Then, of course, there comes up the question of the nature of the agreement, and of the recognition of it. What is the experience in which the survey of both idea and existence is made and their agreement recognized? Is it an idea? Is the agreement ultimately a matter of self-consistency of ideas? Then what has become of the postulate that truth is agreement of idea with existence beyond idea? Is it an absolute which transcends and absorbs the difference? Then, once more, what is the test of any specific judgment? What has become of the correspondence of fact and thought? Or, more urgently, since the pressing problem of life, of practice and of science, is the discrimination of the relative, or superior, validity of this or that theory, plan, or interpretation, what is the criterion of truth within present non-absolutistic experience, where the distinction between factual conditions and thoughts and the necessity of some working adjustment persist?

Putting the problem in yet another way, either both fact and idea are present all the time or else only one of them is present. But if the former, why should there be an idea at all, and why should it have to be tested by the fact? When we already have what we want, namely, existence, reality, why should we take up the wholly supernumerary task of forming more or less imperfect ideas of those facts, and then engage in the idle performance of testing them by what we already know to be? But if only ideas are present, it is idle to speak of comparing an idea with facts and testing its validity by its agreement. The elaboration and refinement of ideas to the uttermost still leaves us with an idea, and while a self-consistent idea stands a show of being true in a way in which an incoherent one does not, a self-consistent idea is still but a hypothesis, a candidate for truth. Ideas are not made true by getting bigger. But if only 'facts' are present, the whole conception of agreement is once more given up—not to mention that such a situation is one in which there is by definition no thinking or reflective factor at all.

This suggests that a strictly monistic epistemology, whether idealistic or realistic, does not get rid of the problem. Suppose for example we take a sensationalistic idealism. It does away with the ontological gulf between ideas and facts, and by reducing both terms to a common denominator seems to facilitate fruitful discussion of the problem. But the problem of the distinction and reference (agreement, correspondence) of two types or sorts of sensations still persists. If I say the box there is square, and call "box" one of a group of ideas or sensations and "square" another sensation or "idea," the old question comes up: Is "square" already a part of the "facts" of the box, or is it not? If it is, it is a supernumerary, an idle thing, both as an idea and as an assertion of fact; if it is not, how can we compare the two ideas, and what on earth or in heaven does their agreement or correspondence mean? If it means simply that we experience the two "sensations" in juxtaposition, then the same is true, of course, of any casual association or hallucination. On the sensational basis, accordingly, there is still a distinction of something "given," "there," brutally factual, the box, and something else which stands on a different level, ideal, absent, intended, demanded, the "square," which is asserted to hold good or be true of the thing "box." The fact that both are sensations throws no light on the logical validity of any proposition or belief, because by theory a like statement holds of every possible proposition.[53]

The same problem recurs on a realistic basis. For example, there has recently been propounded[54] the doctrine of the distinction between relations of space and time and relations of meaning or significance, as a key to the problem of knowledge. Things exist in their own characters, in their temporal and spatial relations. When knowledge intervenes, there is nothing new of a subjective or psychical sort, but simply a new relation of the things—the suggesting or signifying of one thing by another. Now this seems to be an excellent way of stating the logical problem, but, I take it, it states and does not solve. For the characteristic of such situations, claiming to terminate in knowledge, is precisely that the meaning-relation is predicated of the other relations; it is referred to them; it is not simply a supervention existing side by side with them, like casual suggestions or the play of phantasy. It is something which the facts, the qualitative space and time things, must bear the burden of, must accept and take unto themselves as part of themselves. Until this happens, we have only "thinking," not accomplished knowledge. Hence, logically, the existential relations play the rôle of fact, and the relation of signification that of idea,[55] distinguished from fact and yet, if valid, to hold of fact.

This appears quite clearly in the following quotation: "It is the ice which means that it will cool the water, just as much as it is the ice which does cool the water when put into it." There is, however, a possible ambiguity in the statement, to which we shall return later. That the "ice" (the thing regarded as ice) suggests cooling is as real as is a case of actual cooling. But, of course, not every suggestion is valid. The "ice" may be a crystal, and it will not cool water at all. So far as it is already certain that this is ice, and also certain that ice, under all circumstances, cools water, the meaning-relation stands on the same level as the physical, being not merely suggested, but part of the facts ascertained. It is not a meaning-relation as such at all. We already have truth; the entire work of knowing as logical is done; we have no longer the relation characteristic of reflective situations. Here again the implication of the thinking situation is of some "correspondence" or "agreement" between two sets of distinguished relations; the problem of valid determination remains the central question of any theory of knowing in its relation to facts and truth.[56]