Essays in Experimental Logic

by John Dewey

I

To my mind, Professor Perry rendered philosophic discussion a real service when he coined the phrase "ego-centric predicament." The phrase designated something which, whether or no it be real in itself, is very real in current discussion, and designating it rendered it more accessible to examination. In terming the alleged uniform complicity of a knower a predicament, it is intended, I take it, to suggest, among other things, that we have here a difficulty with which all schools of thought alike must reckon, so that it is a difficulty that cannot be used as an argument in behalf of one school and against another. If the relation be ubiquitous, it affects alike every view, every theory, every object experienced; it is no respecter of persons, no respecter of doctrines. Since it cannot make any difference to any particular object, to any particular logical assertion, or to any particular theory, it does not support an idealistic as against a realistic theory. Being a universal common denominator of all theories, it cancels out of all of them alike. It leaves the issue one of subject-matter, to be decided on the basis of that subject-matter, not on the basis of an unescapable attendant consideration that the subject-matter must be known in order to be discussed. In short, the moral is quite literally, "Forget it," or "Cut it out."

But the idealist may be imagined to reply somewhat as follows: "If the ubiquity were of any kind other than precisely the kind it is, the advice to disregard it as a mere attendant circumstance of discussion would be relevant. Thus, for example, we disregard gravitation when we are considering a particular chemical reaction; there is no ground for supposing that it affects a reaction in any way that modifies it as a chemical reaction. And if the 'ego-centric' relation were cited when the point at issue is something about one group of facts in distinction from another group, it ought certainly to be canceled from any statement about them. But since the point at issue is precisely the most universally defining trait of existence as known, the invitation deliberately to disregard the most universal trait is nothing more or less than an invitation to philosophic suicide."

If the idealist I have imagined as making the foregoing retort were up in recent realistic literature, he might add the following argument ad hominem: "You, my realistic opponent, say that the doctrine of the external relation of terms expresses a ubiquitous mark of every genuine proposition or relational complex, and that this ubiquity is a strong presumption in favor of realism. Why so uneven, so partial, in your attitude toward ubiquitous relations? Is it perchance that you were so uneasy at our possession of a ubiquitous relation that gives a short cut to idealism that you felt you must also have a short cut to realism?"

If I terminate the controversy at this point, it is not because I think the realist is unable to "come back." On the contrary, I stop here because I believe (for reasons that will come out shortly) that both realist and idealist, having the same primary assumption, can come back at each other indefinitely. Consequently, I wish to employ the existence of this tu quoque controversy to raise the question: Under what conditions is the relation of knower to known an intelligible question? And I wish to show that it is not intelligible, if the knowledge relation be ubiquitous and homogeneous.

The controversy back and forth is in fact a warning of each side by the other not to depart from their common premise. If the idealist begins to argue (as he constantly does) as if the relation to "mind" or to "consciousness" made some difference of a specific sort, like that between error and fact, or between sound perception and hallucination, he may be reminded that, since this relation is uniform, it substantiates and nullifies all things alike. And the realist is quite within the common premise when he points out that every special fact must be admitted for what it is specifically known to be; no idealistic doctrine can turn the edge of the fact that knowledge has evolved historically out of a state in which there was no mind, or of the fact that knowledge is even now dependent on the brain, provided that specific evidence shows these to be facts. The realist, on the other hand, must admit that, after all, the entire body of known facts, or of science, including such facts as the above, is held fast and tight in the net of relation to a mind or consciousness. In specific cases this relation may be ignored, but the exact ground for such an ignoring is precisely that the relation is not a specific fact, but a uniform relation of facts. And to call it an external relation makes no practical difference if it is universal and uniform. So the idealist might reply.

Imagine a situation like the following: The sole relation an organism bears to things is that of eater; the sole relation the environment bears to the organism is that of food, that is, things-to-eat. This relation, then, is exhaustive. It defines, or identifies, each term in relation to the other. But this means that there are not, as respects organism and environment, two terms at all. Eater-of-food and food-being-eaten are two names for one and the same situation. Could there be imagined a greater absurdity than to set to work to discuss the relation of eater to food, of organism to the environment, or to argue as to whether one modifies the other or not? Given the premise, the statements in such a discussion could have only a verbal difference from one another.

Suppose, however, the discussion has somehow got under way. Sides have been taken; the philosophical world is divided into two great camps, "foodists" and "eaterists." The eaterists (idealists) contend that no object exists except in relation to eating; hence that everything is constituted a thing by its relation to eating. Special sciences exist indeed which discuss the nature of various sorts of things in relation to one another, and hence in legitimate abstraction from the fact that they are all foods. But the discussion of their nature an sich depends upon "eatology," which deals primarily with the problem of the possibility, nature, and extent (or limits) of eating food in general, and thereby determines what food in general, überhaupt, is and means.

Nay, replies the foodist (realist). Since the eating relation is uniform, it is negligible. All propositions which have any intelligible meaning are about objects just as they are, and in the relations they bear to one another. Foods pass in and out of the relation to eater with no change in their own traits. Moreover, the position of the eaterists is self-contradictory. How can a thing be eaten unless it is, in and of itself, a food? To suppose that a food is constituted by eating is to presuppose that eating eats eating, and so on in infinite regress. In short, to be an eater is to be an eater of food; take away the independent existence of foods, and you deny the existence and the possibility of an eater.

I respectfully submit that there is no terminus to such a discussion. For either both sides are saying the same thing in different words, or else both of them depart from their common premise, and unwittingly smuggle in some relations between the organism and environment other than that of food-eater. If to be an eater means that an organism which is more and other than an eater is doing something distinctive, because contrasting with its other functions, in eating then, and then only, is there an issue. In this latter case, the thing which is food may, of course, be proved to be something besides food, because of some different relation to the organism than that of eating. But if both stick consistently to their common premise, we get the following trivial situation. The idealist says: "Every philosophy purports to be knowledge, knowledge of objects; all knowledge implies relation to mind; therefore every object with which philosophy deals is object-in-relation-to-mind." The realist says: "To be a mind is to be a knower; to be a knower is to be a knower-of-objects. Without the objects to be known, mind, the knower, is and means nothing."

The difficulties attending the discussion of epistemology are in no way attendant upon the special subject-matter of "epistemology." They are found wherever any reciprocal relation is taken to define, exclusively and exhaustively, all the connections between any pair of things. If there are two things that stand solely as buyer and seller to each other, or as husband and wife, then that relation is "unique," and undefinable; to discuss the relation of the relation to the terms of which it is the relation, is an obvious absurdity; to assert that the relation does not modify the "seller," the "wife," or the "object known," is to discuss the relation of the relation just as much as to assert the opposite. The only reason, I think, why anyone has ever supposed the case of knower-known to differ from any case of an alleged exhaustive and exclusive correlation is that while the knower is only one—just knower—the objects known are obviously many, and sustain many relations to one another which vary independently of their relation to the knower. This is the undoubted fact at the bottom of epistemological realism. But the idealist is entitled to reply that the objects in their variable relations to one another nevertheless fall within a relation to a knower, as long as that relation is regarded by both as exhaustive or ubiquitous.