Taylor, A. E. “Truth and Practice.” The Philosophical Review 14, no. 3 (1905): 265. https://doi.org/10.2307/2176580
THE question whether there is a determinate and direct connection between the truth of a proposition and its practical utility is one which, besides being interesting in itself, has of late been very prominently brought before the notice of students of philosophy by the sharply controversial articles of Mr. F.H. Bradley and Professor James and Mr. Schiller. Hence, some reflections on the precise nature of the logical problem involved can hardly fail to be opportune at the present moment. More particularly will they be opportune, if, as I almost venture to hope, the suggestions I am about to offer are adapted to serve to some extent as an eirenicon between the contending parties. That either side will accept my suggestions in their entirety is, indeed, much more than I dare to expect; had equine talkie dig nor honoree; but at least I believe they may be found serviceable as a basis for future discussion. May I add, that I shallot any rate try to pitch my remarks in the key appropriate to equable philosophic argumentation, and to avoid making any addition to the stock of sub pleasantries and personalities which the discussion of this particular problem is already so rich? My main object, then, in the present paper is less to offer a positive solution of the problem than to urge the necessity, as a preliminary to any solution, of a careful delimitation of the logical issue at stake from irrelevant extra-logical associations of a psychological kind by which it is in danger of being confused.
In discussing the problems involved in the truth or falsehood of any statement or theory, three quite distinct questions arise which need to be carefully discriminated, but appear to be often confounded, especially by the advocates of the 'Pragmatist ‘doctrines. We have to ask: (I) What is the meaning of the contrasted predicates true and false? (2) To what propositions as subjects is each of these predicates correctly ascribed? (3) How have we come to make the ascription in any given case or class of cases? Any confusion of one of these questions with another is bound to lead to a serious ignoration elenchi, particularly when, as commonly happens, it is the third which is confused with either or both of the others. A theory of the steps by which true convictions are arrived at can, however true, manifestly be no answer either to the question what we mean by calling a conviction true, or to the question what convictions in particular are the true ones.
Now, of these three questions, it is the third, and the third only, which introduces psychological matter, or makes any reference to the existence or the properties of 'states' or 'processes' of consciousness as the means by which the individual mind comes to be aware of truth and falsehood. The other two questions are entirely extra-psychological, (I) belonging to the theory of pure formal logic, and (2) being coextensive with the whole field of the sciences. It would thus appear that any appeal to genetic psychology, whether of the race or of the individual, must be irrelevant, if introduced into the discussion of either of these problems. In particular, it would seem that no theory as to the meaning of the terms true and false or the distribution of true and false propositions can be relevantly assailed or maintained on the ground of any special evolutionist view as to the influence of practical needs on the development of our cognitive faculties. Considerations of this kind are relevant and important in connection with our third question; if intruded into the discussion of the other two, they constitute a serious ignoration elenchid. This becomes even more apparent, when we reflect that even a philosophy which denied the existence of such things as the 'states of consciousness' which psychology assumes as its subject-matter, would still have to face both our other problems. For in asserting that it is true that there are no such things as 'states of consciousness,' it would lay itself open at once to the questions: What is meant by the truth ascribed to this assertion? And on what grounds is it maintained? But even without this reflection, it should surely be evident that the questions of what I mean by calling a thing true, what reason I have for calling it so, and how have come to call it so, are distinct and separate. In precisely the same way evolutionary considerations of the same type may help to explain, e.g. how we come to have a color sense, and again, how it comes to be more readily sensitive to some hues than to others, by dwelling on the practical advantages secured by such a means of recognizing one's enemy or one's provender from a distance. They are quite irrelevant if the question is what we mean by red or blue, or again, what things are red things or blue things. It is probably only the traditional and persistent modern error of regarding logic as somehow concerned with the subjective processes of cognition which gives vitality to so elementary fallacy. When logic is defined with strict relevancy as the doctrine of the implication of propositions, or the science of the estimation of evidence, and thus purged at the outset of psychological accretions, the confusion disappears of itself. Unfortunately few philosophers have with sufficient firmness grasped, as Aristotle appears to do among ancient thinkers and Mr. Bertrand Russell among modern, the simple and important principle that though true propositions are, so far as we know, only thought by individual minds, yet the notion of an individual thinking mind is, as we shall see more clearly in the sequel, absolutely irrelevant to the explanation of what we mean by their truth. In other words, the truth of a proposition is a function of its meaning, the content which it asserts, not of its character as a psychical event or process. That this is so, we can see by simply asking ourselves whether the truth of a newly discovered theorem is created by the fact of its discovery, or that of a forgotten one destroyed by its disappearance from men's minds. Did the doctrine of the earth's motion become true when enunciated by the Pythagoreans, false again when men forgot the Pythagorean astronomy, and true a second time on the publication of the book of Copernicus? Or, if not, must we assume as a postulate the existence of an unbroken chain of 'Copernicans before Copernicus' from the days of Pythagoras, or perhaps of Adam onward? I can see no way out of the difficulty for an obstinately psychologizing logician better than that traditional refuge of philosophers in distress, the postulation of a Deity framed, like the educational authority of English radicals, ad hoc, as a convenient receptacle for deceased and yet unborn states of human consciousness.
In dealing, then, with the logical bearings of conflicting theories of the relation between truth and practice, we must carefully confine ourselves to our first question, that of the meaning of the concepts of truth and of practice. And, in order to give an answer of any kind, we must first of all define the terms with which we are dealing. Unless our terms are well-defined to begin with, it is as good as certain that any theory we may propound is to the relation between them will be highly ambiguous. It would be impossible to exaggerate the mischief that has been done and is still being done in philosophy by that slip-shod habit of mind, self-styled 'intellectual flexibility' and 'freedom,' which systematically shirks the task of definition. “Space and time, why everybody knows what they are." "Things? Anyone can tell what we mean by things, so long as he does not trouble about pedantic refinement in expression." It is excuses of this kind for mental indolence that have done more than anything else to hinder the general attainment of clear consciousness of what space and time and things are. Yet there are not a few reputable philosophers who owe a good deal of the consideration they enjoy to this genteel indifference to accuracy, while a Socrates gets the credit of 'irony' for his protestations of his own ignorance, precisely because the bulk of mankind are incapable of sharing the Socratic conviction that the definable is not really known until you can define it. To be sure, if you insist on precise definition wherever it is possible, you are likely at first to define many things unsuitably; but then there is always the hope that the very imperfection of the definition may, by leading to manifest errors of inference, bring about its own detection, whereas, where you are dealing with terms of which the precise import has never been assigned, you may go on forever heaping inconsequence on inconsequence without being brought to a halt.
First, then, as to truth. Truth and falsehood have sometimes been held to be simple indefinable, just like such elementary sense-qualities as 'white' or 'bitter.' This position, however, appears to me inconsistent with the actual results of pure formal logic. To begin with, we may note that it is only when the word 'definition' is used in a very special sense, which is not that employed in pure logic, that such simple sense-qualities are rightly said to be 'indefinable.' If by definition you mean analysis of a complex object of thought into its constituent parts, of course, there can be no definition of what is ex hypothesis a simple concept. But in another way, you can define such terms as white or sweet quite easily, or, at least, you could do so were our detailed knowledge of the physical world rather more exact than it is. You can specify or could if our knowledge of anatomy and physiology were more advanced, the precise nervous processes with which the sensation in question is correlated. Thus, given the general concept of psychophysical concomitance, and the more special concepts of the processes aroused by stimulation of particular known arrangements in the central nervous system, any sensation may be defined as the correlate of a specific nervous process. In a similar way truth, or rather the class of truths, may be unambiguously defined, if we can find, as we know from formal logic we can, a relation to the whole class of propositions which is satisfied by all truths and by them only. But though it is thus an error to maintain that the concepts 'true,' 'white,' 'sweet,' because unanalyzable, must be indefinable, there is an element of truth confusedly expressed by the mistake. What is true is that you cannot recognize truth or whiteness or sweetness when you meet with them; in other words, you cannot identify the members of the class determined by the definition, unless you are already acquainted with at least one example of a truth or a white surface or a sweet taste. Thus, a person born blind, and subsequently endowed with sight by a miracle or by the less sensational methods of modern surgery, might during the period of his blindness be perfectly aware that white is the sense-quality which is experienced when a certain specific complex of nervous processes is aroused; but, on his acquisition of sight, he would nonetheless be quite unable to identify this quality until he had once at least been shown a white thing and told ' that is white.' In this respect the definition of the complex by analysis enjoys psychologically a certain advantage. If I know what black is and what a man is, and further that a negro is a black man, I am in a position to identify a negro, if ever I come across one, without needing first to be shown an example of an actual negro. Perhaps this may in part explain why the attention of philosophers has been chiefly bestowed upon a type of definition which is neither the only one nor the most important. It also explains why it is that logical theory can escape in the end from recognizing the existence of genuinely self-evident truths. Truths as a class, then, are definable, and, as is well known, their definition is effected in modern formal logic by the statement that true propositions are the class of propositions which are implied by all propositions, and false propositions the class of propositions which imply all propositions. When we remember that in formal logic ‘a implies b' simply means that ' the conditions under which a is true are included in the conditions under which b is true,' or ' whenever is true, b is true also,' we readily see that this definition is simply tantamount to saying in a highly condensed and artificial form that true propositions are those which have an unconditional claim on our recognition, a right to be affirmed, and false propositions the denials of them. In other words, true propositions are those which are always, false propositions those which are never, entitled to recognition. That any proposition whatever, when so stated as to contain no time-variable or other ambiguous element belongs to one and only to one of these two classes is, of course, the gist of the laws of contradiction and excluded middle when interpreted of statements.
Now arises, as in the case of all definitions, the need of proving by an existence theorem that the class thus defined contains at least one member and is thus not identical with the null-class or ‘identical zero.' We have, in fact, to show that all propositions cannot be false, that there is at least one absolutely certain true proposition. This can easily be done by the citation of any one simple, self-evident statement, such as, e.g., 'when I feel hot, I do feel hot, ‘when I deny, I do deny,' or better still, " 'all propositions are false' implies 'at least one proposition (viz., That all propositions are false) is true. ‘This last example is clearly identical in its essential meaning with the famous Cartesian Cogito, of which the real function is merely to prove in this fashion that the class of truths is not an empty one; and one may be allowed to doubt whether Descartes's example would not have been more happily chosen, if he had, like ourselves, taken care to keep it free from every vestige of irrelevant psychological assumption about ‘consciousness.'
Having got so far, we can easily go on to show that, since the class of true propositions contains one member, it contains an infinite number of members. This has been shown, after Dedekind, by Professor Royce, on the ground that if a is true, it is true that a is true, and again true that it is true that a is true, and so on. If the objection should be raised that the successive members of this infinite series are really only verbally different forms of the same original assertion, the criticism might, I think, be met by a slight modification of the argument. We may in any case reason thus. If a is a true assertion, then at any rate ' there is one true assertion ' must be allowed to be also true, and to be a second true assertion, different in meaning from itself. Then it follows also that there is also a third truth, viz. 'that there are at least two truths,' and a fourth 'that there are three truths,' and so on. All this is interesting as showing that Dedekind and Royce are demonstrably right in finding in the class of truths! a proof of the existence of infinite classes but will not further concern us here.
Next as to the meaning of practice. Practice is a term which obviously might be taken either in a narrowly physical or in a wider sense. We might conceivably define it exclusively in terms of physical motor reaction; but with such a definition we should either stand committed to the as yet unproved hypothesis that all intelligent self-adaptation to changes in the objective environment, and consequently all voluntary control of the sequence of our own thoughts, consists in modification of motor responses and in nothing else, or should have to refuse the name of practice to those self-adaptive processes which have not been proved to consist exclusively in motor changes. The latter alternative would clearly involve us in serious difficulty, when we go on to deal with ethical questions as to the respective functions of 'practical' and 'theoretical' reason and to ask whether the one ought morally to be subordinated to the other; the former implies acceptance of a far-reaching psychological theory which is neither fully demonstrated nor yet self-evident. It is better than to define practice, independently of any special theory as to the psychophysics of the matter, in purely psychological terms as the self-initiated alteration by individuals of some datum of presented fact, thus following closely in the path laid down by Mr. F. H. Bradley in his recent articles on "The Psychology of the Will."
If we accept either of the above accounts of what is meant by practice, it follows at once that we cannot acquiesce without further ado in the undemonstrated and indemonstrable identification of truth with practical efficacy. For we cannot admit without further proof, as a logically necessary inference from the definitions of truth and of practice, that my recognition of propositions having a right to acceptance is in actual fact followed in every case by consequent alteration of presented data. The conclusion, in fact, would only follow, if we admitted as certain the principle that all change in consciousness involves change in motor reaction. Still less should we be justified in assuming, even if we allow this principle to pass unchallenged, that these alterations of presented data are the sole thing which gives truth a higher value for us than falsehood. To warrant such an inference, we must further assume not only that the effects of recognizing propositions as true always include motor change, but that they include nothing else. For if they include motor change and something else as well, it may be that this something else is in part, at least, the reason for our preference of truth to falsehood.
But the real issue lies still deeper. Even if we admitted both the references rejected in the last paragraph, we should still not have shown that the truth of a statement is identical with the practical consequences of belief in it. For, as we have seen, the truth of a statement means not the actual fact of its recognition, but its rightful claim on our recognition. There may be, indeed, unless we assume that mankind must sooner or later know all that there is to be known, there must be truths which are never actually recognized by us, and which therefore, though true, exercise no effect upon our practice. It is thus imperative to recognize the logical distinction between the truth of a statement., its claim on our acceptance, and the alterations of fact which issue from the actual acceptance of it. For other sciences this distinction may possibly be negligible, for logic it is fundamental. The distinction turns, in fact, upon the principle that practice is essentially alteration of facts by individuals, and is thus relative to the individual, whereas truth, the rightful claim to recognition, if relative to actual consciousness at all, is relative not to the consciousness of individuals, but to an ideal or universal consciousness, a universal impersonal Burstein bureau, which is not identical with any individual consciousness, and of which the actual existence is at least problematical. Hence considerations which are practically important for one man or set of men may be of no practical importance for another, and we might even ask whether we have a right to speak of any considerations as practically important for all individual intelligences without exception. But there is no sense in speaking of a proposition as true for one mind, but not true for another.
To take a single example, when Mr. F's aunt interrupted a conversation with the observation that " there's milestones on the Dover road," she was guilty of an irrelevance, because the remark, whether true or false, had no practical bearing on the behavior of any person in the company. True, any one of those present might happen to find himself on that road, and it might thus become practically important to him to know his distance from London and the direction in which he was travelling. But the truth of the statement remains quite unaffected by the question whether anyone ever would actually need to take it into account. If true at all, it was true when it was uttered, and would remain equally true if no one ever should have occasion to travel between Dover and London.
Another way of putting the same point would be to say that, while all practice consists in making something, all truths are in the end found or accepted, not made. We may or may not formulate certain proposition, but once formulated its truth means right or claim to admission which is entirely independent of individual volition. Either it has the right, or it has not, and individual choice can neither confer the right where it does not exist nor destroy it where it does. All that lies in our power is to grant or withhold our actual individual recognition, and thus the right to recognition inevitably remains unaffected by our action. To deny this result would be in effect to deny that there is any such thing as a truth as previously defined, i.e., to assert that the class of truths is identical with the null class. And this would be suicidal, for we see at once that, if there is no truth, then the statement, 'There is no truth,' cannot itself be true. With Professor Rickert I cannot but regard this consideration fatal to the attempt to extract from the doctrine of the primacy of the practical reason a defense of the now very properly discredited fancy that 'all knowledge is relative.' Obvious as these reflections are, they are so often lost sight of, and the neglect of them leads to such inextricable confusion, that I venture to repeat them. We can to some degree voluntarily accelerator retard the discovery of truth; we can put ourselves into more or less fit state to recognize it; we can make increasingly successful approximations to it; but we are absolutely powerless, and God, if he exists, must be equally powerless, to make truth false or falsehood true. In none of the instances which have been cited in support of the opposite view to our own dowel get a genuine example of a belief 'making itself true.’ For instance, if I am cured of a complaint by 'faith-healing,' you might at first sight suppose that the belief, ‘I am now well,' has thus made itself true. But you suppose this only because you neglect to observe that the proposition, ‘I am now well containing a time-variable. (In the language of Mr. Bertrand Russell, the sentence is not a genuine proposition but a prepositional function, a general scheme from which an infinity of different propositions may be derived by causing the variable to assume succession of values.) For some values of the variable (the ‘now') the proposition is false, but for subsequent values true. For the same value of the ' now 'it is once for all either finally true or finally false.
While I am upon this question whether a belief can in any sense 'make itself true,' I should like to point out a confusion which, unless detected and avoided, is almost certain to make havoc of our reasoning on the subject. The confusion I refer to toys, indeed, simply one more instance of the obstinate tendency of the psychologist which is apt to beset all our reflections on logic. We cannot too carefully discriminate between the 'propositions' of logic, the relations which are asserted or denied to subsist between the objects of our thought, and the psychological ‘judgments,' the subjective 'states of mind' by which individuals actually assert or deny those relations. We cannot too often remind ourselves that logic is concerned solely with the meaning, psychology only with the process of assertion and denial.
Now in the case before us, what is true or false is the logical meaning of the belief in question, and this, like all meanings, is outside the flow of events, and has no place in the time-series. What is effective in producing an alteration in the realm of fact, what makes the sick man whole, is not this non-temporal logical meaning of the proposition, but the processes by which he as an individual comes at a given moment to frame the corresponding judgment and to affirm its truth. There is an ambiguity about our English word 'belief' which lends itself only too readily to the psychologizing fallacy. A belief may mean either the thing which is believed in or the occurrence in me of the state of mind of believing in it. It is only the former which can be false or true, only the latter which can be operative in effecting a change in my relation to my environment; the psychologizing fallacy consists in the transference to each of predicates which are only intelligible of the other.
I would now briefly enumerate some conclusions from the foregoing reflections which seem to me important as possibly providing at any rate a basis from which to estimate the value of pragmatistic doctrine of knowledge.
I. A truth, as I have already said, may have a claim to our recognition even though no one has actually recognized it, Justas a moral precept may have a right to my obedience and yet in point of fact be disobeyed. And further, just as even universal disobedience would not destroy the obligatoriness of obedience to the moral precept, so even universal failure to recognize a truth would not make it the less true. So again, in aesthetics. Shelley's poetry was truly beautiful when his contemporaries agreed to regard it as 'driveling prose run mad’ no less than it is to-day; the improvement which time has brought with it in our literary taste is an improvement in us, not in Shelley. And if the improvement had never taken place, and we still regarded Prometheus and The Cloud as 'drivel,' our error would have no bearing on their real worth. In all three cases the question is one not of fact but of right. Hence any identification of truth as such with 'what mankind will ultimately agree to admit,' 'the beliefs which will in the long run naturally establish themselves under the stress of the evolutionary process,' seems in principle perverse. We have no more ground to suppose that all truths must someday come to be recognized and all errors discarded by mankind than we should have to suppose that the moral ideal will one day be universally attained and moral wrong-doing a thing of the past. And we know, as a matter of history, that both error and crime have played an important part in the evolution of humanity. Why, then, should we deny that they will always continue to do so?
Now, if we admit that a proposition may be true and yet a single mind (my own, for example) fail to recognize its truth, it is easy to see that truth must be equally unaffected even though all minds without exception should fail to recognize it. The mere number of the deluded can surely make no difference to the principle. To introduce any reference to actual recognition by individuals into the definition of truth, is to commit the old ignoration elenchi of Locke and others, who fancied that the validity of the ontological proof could be overthrown by simply producing an actual atheist.
But, it may perhaps be objected, at this rate truth will be hypostatized into a realm of realistic things-in-themselves, standing out of all relation to the minds which know it, and their knowledge of it reduced to the position of a lucky but inexplicable accident. I reply that the conclusion is entirely unjustifiable. Our very definition of a truth as a proposition which has a right to recognition, implies the essential correlation of the knowing intelligence to the truth known, and the impossibility of giving account of either in abstraction from the other. So much I would not only admit but insist upon. But, and thesis the point, though the recognition of every truth may be logically implied in the recognition of any, it is only a very few of these implications which are actually perceived by any of the intelligences with which we are acquainted. The great majority of them are simply overlooked. Now this may conceivably bathe the case with all intelligences, unless indeed we could prove that the existence of an omniscient God is an indispensable presupposition of thought. Hence, except in this latter case, the possibility will remain open that there are truths which are never known, precisely because no one, except God, if there is a God, can be presumed to be aware of all that is implied in what he does know. It is thus possible to reject the identification of truth with what is actually recognized by individuals without denying the essential correlation of truth and intelligence, and to hold that some truths may remain permanently unknown without affirming a crude realism in metaphysics. And even if there were reasons for believing that some or all intelligences will ultimately know everything, it would still be necessary in logic to distinguish between the propositions affirmed by such omniscient beings and the psychological process of their affirmation.
2. It also follows that, if truth is quickstep tun's non fact, there cannot possibly be any universal and infallible psychological criterion of truth. To what Professor James says of the absence of any "click of the mental machinery" by which truths might be known when obtained, I would only respond by the fullest admission of the fact. We may, however, while we admit its truth, dispute its relevancy to logic. Indeed, the general position of the famous, I had almost written too famous, essay on the Will to Believe would, it seems to me, amount in practice to almost dangerous, because in the strict and literal sense of the word, unprincipled skepticism, and to justify a very great doubt about the realization of Mr. Schiller's hope that the joint doctrine of Professor James and himself will deliver us from the blighting influences of Agnosticism. For, if there is no certain psychological criterion of truth, it surely must follow that we cannot legitimately treat our private emotional satisfaction in a belief as such a criterion. Indeed, the illegitimacy of such an appeal, unsupported by the allegation of logical grounds of conviction, becomes apparent as soon as we realize that we are being urged to infer from the premises, ‘It gives me satisfaction to believe this,' the conclusion, 'this-has a claim to be believed by all intelligences.'
One might similarly argue from 'I very much wish to act thus 'to' any responsible being in these circumstances ought to ‘act thus'; but the moralists, I fear, would disallow such a practical syllogism. I know, of course, that elsewhere Professor James expresses himself in a more guarded way and makes his appeal not so much to the individual's private emotions as to the concordant emotional aspirations of mankind as a race. But, not to raise the question whether the race as a whole are enough atone in their ethical and religious emotions to justify our talking of fundamental rein-mensch ich demands on the nature of things, a comparative study of the historical effects of salutary error in ‘assisting the mental and moral development of civilizations would surely go far to shake the conviction that what men will hereafter believe, and what it will be beneficial that they should believe, must needs be the truth. I do not for a moment deny that the salutary errors of the past have been very far from being pure and unmixed errors; on the contrary, they have usually borne deceptive resemblance to important truths, have been first approximations to truth. And I would readily concede that it is always wise to look for some kernel of truth in any widespread permanent and practically effective human belief. Only we know that such beliefs in the past have been mere first approaches to truth, have frequently contained more husk than kernel; and I see no reason to doubt that it will continue to be so in the future. And this of itself is enough to show that what we mean by the truth of a statement cannot be identical with an expectation that it will one day be universally admitted and universally efficacious in regulating conduct. Of course, our opponent may retort by whispering with Nietzsche: "But how do you know that the absolute and ideal truth of which you talk is anything more than an empty idol of the cave? "But the suggestion, which has been refuted on logical grounds, involves complete and hopeless surrender to the extremists of philosophic skepticism.
After all, there is only one way of establishing the truth of a proposition which is not self-evident, and that is either to deduce it directly from true premises by the use of self-evident principles of inference, or indirectly by deduction from such principles to show that the denial of the proposition leads to results inconsistent with already known truth. It may, to be sure, be asked how a self-evident truth can be recognized without some psychological criterion. But the answer is not far to seek. All alleged self-evident principles are such as to be capable of statement in the form 'a implies b? or 'if x is a, it is also b'; to refute this allegation, what you need to do is simply to produce an instance of an x which is a without being b. Thus, suppose it is alleged as self-evident that two intersecting straight lines cannot both be parallel to one and the same third straight line, and you desire to disprove the self-evident character of the allegation. The refutation will consist in producing a pair of intersecting lines which satisfy at once the definition of straightness and the definition of parallelism to one and the same third. So, to take another instance, let it be disputed whether the proposition that ‘if x is either b or c, then either it is b or it is c,’ is a self-evident principle of logic. That it is not, that the principle is, in fact, false, if affirmed without restriction, is shown by considering the case in which x, b, c are all classes containing more than one-member. We then see at once that from 'all the persons in this room are Canadians or Americans,' we cannot infer that they are all Canadians or all Americans. Thus, the refutation of an alleged claim to self-evidence reduces to the establishment of an existence-theorem. This might, indeed, have been inferred immediately from the consideration that a self-evident proposition must always state a logical implication, and must therefore be formally universal. Hence the denial of it must be particular, and every particular proposition, as we know, has existential import. The interest of these obvious and elementary reflections is that they go far to remove the prejudice apparently felt by some philosophers, notably Professor T. H. Green and his disciples, against the recognition of any really self-evident principles. For we see from them that there is no psychological criterion of self-evidence, and in particular that no proposition is proved to be self-evident by the fact that when I 'look into my breast' I find a strong propensity to believe it. If self-evidence cannot be directly proved, at least it can be formally disproved. They show again that, except in the case of the universal principles of all logical inference themselves, indemonstrability up to date need not be the same thing as genuine self-evidence. For the list of indemonstrable is capable of reduction in two ways. The supposed indemonstrable may turn out to be actually deducible from more ultimate premises, as has already been shown to be the case with many so-called axioms, both of logic and of arithmetic. Or, again, as in the examples previously considered, the supposed axiom may be shown to be false by the production of a class of objects which satisfy one side of it the antecedent or hypothesis but not the other the consequent or thesis. Finally, they show that, as Professor Royce has reminded us, all truths alike are ultimately truths of experience, the only logical peculiarity about those statements as to particular event of the temporal order to which that name is conventionally confined being that they involve in their enunciation as genuine propositions a time-variable.
3. That 'truth' cannot mean identically the same thing as 'usefulness for practice' seems clear from what has been said, and becomes even clearer if we ask, what would be meant by the assertion 'it is true that truth means the same thing as usefulness ‘itself. This cannot mean merely that it is useful to believe the tautologous statement that usefulness is usefulness. For it would be hard to find a less useful proposition than such a barren tautology, and I am therefore sure that those who seek to found gospel on the alleged identity of truth and utility must mean something very different. To be sure, if we go to particular instances of true assertions, the emptiness of the identification is not so immediately apparent. It looks at first sight plausible today. 'It is true that fire burns 'means' it is useful to think that fire burns.' But then it is only useful to think this because fire does actually burn, i.e., because 'fire burns' is true. The usefulness of the belief is a consequence of its truth, its truth a condition of its usefulness. To repeat an illustration which I have employed elsewhere, why is it more useful to believe 2 + 2 = 4 than to believe that 2 + 2 = 3? Because 2 + 2 are 4; if 2 + 2 were 3, it would be useful to call them 3 and to count them as 3.Apart from the logically prior truth of the former statement, it would be no more useful than the latter.
But presumably what is really in the minds of many of those who assert the identity of truth and usefulness is merely that usefulness is inseparable from, and forms a universal criterion of truth, just as what many Hedonists seem to mean by the doctrine that the good is identical with the pleasurable is that pleasurableness is a criterion of moral worth. At least, this seems to be all that Professor James is really arguing for in much of his reasoning, though he occasionally commits himself verbally to the more sweeping interpretation of pragmatism. But even in this restricted sense, is the proposition 'truth is usefulness ‘itself true? To prove it true, we need to show: (a) that all truths are practically useful, and (b) that all practically useful propositions are true. If there are grounds for disputing either of these propositions, usefulness ceases to be a universally satisfactory criterion of truth. Now both positions appear at best open to serious questions.
a. There are truths which do not appear to be of practical use. Instances of such truths can easily be produced from among the formulae of pure logical and mathematical science. Thus, e. g., take the curious logical theorem (Schroder, II., 270) as to the mutual implications of three propositions, ‘If a does not imply b, then b implies c, whatever propositions a, b, c maybe.' It is easy to prove that this is true, but difficult, not today impossible, to see what can be the practical use of such a proposition. Or, again, we may appeal to the case of the approximate values of such incommensurables as pi and e. The value of pi and be worked out to 100 decimal places are clearly truer than values calculated to 10 places only, but surely it does not follow that they are more useful in practice. In fact, the reverse is pretty obviously the case. It would be folly in the application of mathematics to a concrete problem to employ a value for which goes beyond the first few places; but does it follow that it is therefore not true to say that, e. g., the hundredth figure after the decimal point in the expression of pi in the ordinary scale of notation is a 9. Again, there are many true statements relating to present and past events which appear to be entirely devoid of practical significance. Marlowe asserts in his Dr. Faustus that the mistress of Alexander the Great had when alive a mole on her neck. This may be a truth, but I find it hard to see how believing or disbelieving it can have any influence on the practice of myself or any other human being. Of course, it is true that it would usually be rash to assert of any true proposition that it will never in any circumstances become of practical use. Even the correct computation of the hundredth figure to the right of the decimal point in the evaluation of pi might conceivably come to be of practical significance for somebody, though the contingency hardly appears likely. My point is simply that we can know the proposition to be true in advance without needing to wait to decide whether it will ever have practical importance or not.
The special case of historical truths lends itself to an ambiguity which must be my excuse for dwelling on it for a few moments. In another context I illustrated the point by saying that the practical usefulness of the belief that Cain killed Abel is no greater than that of the belief that Abel killed Cain, and that consequently your only interest in believing the one statement rather than the other would be that the one is true and the other false. It has since been represented to me that the illustration tells against myself and in favor of Pragmatism. For a practical consequence of believing that Abel killed Cain would be the inference that the Bible narratives are, in some cases at least, false, religion as a practical scheme of life. Now I am very willing to admit the force of this criticism insofar as it bears upon my assertion that it makes no difference now to the conduct of any human being whether Cain really killed Abel or not. The fact is, I suppose, I am myself so accustomed to think of the merits of Christianity as a practical rule of life as being independent of the truth or falsehood of the Hebrew mythology that I forgot tallow for the existence of persons who regard the truth of that mythology as a fundamental condition of the usefulness of Christianity. But even so, in spite of this oversight, I maintain that the illustration will serve my turn. For it will be admitted that to me personally at least, holding as I do the views I have just expressed, the question whether Cain killed his brother or not has no practical importance. My conduct and walk through life are in no way affected by the decision of such an issue. I am neither going to accept the current religion because I believe in the legend of Cain nor to reject it because I disbelieve. And it is at least possible that mankind 'in the long run' may come to the same position. There may be a day, even though I was premature in saying that it had already dawned, when no human creature’s conduct will be more dependent upon beliefs about the doings of Cain than it is now upon beliefs about the color of Pharaoh’s hair. But will it then cease to be the case that one of the two statements, 'Cain killed his brother,' 'Cain did not kill his brother,' is definitely true and the other false?
Or, again, consider the supposed case of a man who should allege his confidence in the veracity of the Biblical story as a reason for being a Christian. The practical utility of his conviction lies here in the fact that it is the cause of his performing the religious duties of a conforming churchman. But when he tells you, 'I act thus because I regard the Biblical narrative as true, ‘he manifestly does not mean, ' I do so because the Biblical narrative does lead in my case to these acts,' and nothing more. He intends to give his belief in the truth of the Bible as a reason, and not merely as an efficient cause, of his conformity; and this of itself suffices to show that he does not mean by the truth of the narrative its mere efficacy in producing conformity.
b. On the other hand, any of the familiar instances of approximation to the value of an incommensurable, such as, e.g. pi, a square or cube root, a logarithm, a trigonometrical function, an area, a volume, will furnish an illustration of the existence of propositions which are practically important and yet not strictly true. Thus pi = 3.1416 is a statement which is frequently of practical significance, though we know it, strictly speaking, to be untrue. And it is worthwhile to remember that most, if not all, physical laws of nature, in the forms in which they are available for practical application to the control of events, involve the use of such rough approximations to truth. Other examples could easily be cited from different quarters. For instance, it would not be in itself irrational to hold that God the Father, the saints, hell, or the philosopher's stone, do not exist, and yet that it was once or is now highly advantageous to the development of science or morals that men should believe in their existence. And I imagine no one would be hardy enough to meet such cases by saying, e.g., that so long as it was beneficial to morality that men should be afraid of hell, hell really existed, but has ceased to exist now that belief in it has become prejudicial rather than beneficial to conduct. Still, if there should be anyone prepared to go even this length, even so a bolder flight yet awaits him. For thought no longer seems desirable for most men that they should believe in hell and its eternal torments, we have all known a few persons who are kept out of mischief, so far as can be judged, largely by their belief in and fear of hell. Must we say then that hell atone exists and does not exist, that the two contradictory propositions expressing the fact of its existence and the fact of its on-existence are both true and both false? And if so, whereas the promised deliverance from the intellectual paralysis of agnosticism?
On grounds like these I feel constrained to hold that the truth of a statement is neither identical with, nor yet a determinate function of, its practical utility. The present popularity of the opposing view is, I believe, largely due to confusion between theological question of the nature of truth and the psychological question of the way in which we arrive at it. In a word, it is a consequence of the obstinate persistence of the psychologizing fallacy. All the most telling of the Pragmatist arguments appear to aim at proving that our individual judgment as to what statements are true is constantly swayed by considerations of practical need. But this, however true, seems quite irrelevant to logic. Against all these often very ingenious attempts to make a genetic psychology take the place of a theory of knowledge, Kant’s famous mot with respect to Locke and his "physiology of the human mind "appears to me to possess more cogency than it has as a criticism upon Locke himself, whose philosophy is after all not entirely contained in his second book. Whether my position is sound or not, it has at least, I believe, the merit of being able to admit the force of the antagonistic arguments as far as they go. What it denies is their relevancy to the logical issue. In conclusion, I may indicate two or three misconceptions which have probably helped largely to popularize Pragmatist theories of truth.
a. What I as an individual actually accept as truth, depends upon what propositions have for me the special feeling of obligatoriness, of demanding unconditionally and of right to be affirmed. The efficient cause of my acceptance of a belief as true is thus a form of emotion, and if we like to give to such an efficient cause the name of subjective criterion, we may say that the individual’s subjective criterion of beliefs is an emotion. Only we must remember the following things. Truth does not mean what is actually believed, but what ought to be believed, not what individuals accept, but what has a right to acceptance by all intelligences universally. And, again, the emotion which we have called the 'subjective criterion' is a very specific form of emotion, just as the emotion which is for me in the same sense the subjective criterion of the beautiful is also unique and specific. It is not all emotional efficacy, but only efficacy in awakening this special feeling of intellectual obligation to affirm or deny, which warrants my acceptance of a proposition. Just as the sense that I want a gratification very much and should enjoy it intensely is of itself no proof that it would be morally right for me to have it, so the consciousness that I should be made very happy by believing a thing to be true is no sufficient reason for my saying that it is true. The feeling that I should be morally better or happier, or my world more beautiful, if a certain belief were true, cannot be regarded as a justification for accepting that belief so long as it does not awaken this specific sense of obligation. Thus man might feel certain that he could check himself in his dissipated courses, if he were really convinced that his dead wife was not lost to him irrevocably, and might further wish with all his heart that he could be convinced of it; yet the longing to believe and the knowledge that belief would have a steadying effect upon conduct are not of themselves a justification at the bar of logic for belief; the specific emotion which accompanies the intellectual obligation to assert can have no substitute. This point, which has been admirably elaborated by Professor Rickert, seems to me to be too little regarded in the philosophies of Professor James and Mr. Schiller.
b. Mr. Schiller's recent (or shall I say most recent?) manifesting Mind (N.S., No. 52) appears to be based on the assumption that the only alternative to the view which turns all truth into practical postulates is the notion, which he reasonably rejects, of truth as a mental copy of an independently existing extra-mental reality. This argument appears to me to turn entirely upon a false disjunction, and I have little doubt that the philosopher against whom Mr. Schiller's polemic is most immediately directed would be in agreement with me here, though I speak, of course, here as everywhere solely for myself. Between the views that truth is somehow manufactured by a process of arbitrary postulation and that it is a 'copy' of a world of realistic 'things-in themselves' there lies yet a third doctrine, viz., that truth is just the system of propositions which have an unconditional claim to be recognized as valid, a doctrine with reference to which I should like once more to call attention to the admirable expository and critical essay of Professor Rickert, Der Greensand der Recentness. That this doctrine is the true one, seems to me tube indicated by its already noted correspondence with the formal logician’s definition of the class of true propositions.
c. The notion that truth can be manufactured by arbitrary postulation has perhaps also been furthered by illegitimate inferences from the correct reflection that it is to some extent a matter of arbitrary choice what properties of an object (e. g., the circle or the ellipse) are taken to form its definition, and again which of a system of connected truths (e.g., those of formal logic) are taken as its primitive axioms. Hence philosophers like Hobbes, who exaggerate the importance of definitions for inference, have at times gone the length of declaring that all truth is arbitrary because it is deduced from arbitrarily adopted definitions. (By the way, why have not the ' friends of postulation,' to speak after the manner of Plato, made some formal acknowledgment of their intellectual kinship with the great Oldman of Amesbury? Let me assure them the relationship is one of which to be proud.) But it must be remembered that, speaking from the standpoint of the strict theory of logical method, no definitions are truths at all.
Methodologically a definition is the mere abbreviation of a set of symbols into a single symbol; in the words of Professor Piano, “every definition expresses an abbreviation which is theoretically unnecessary, though convenient and at times practically necessary for the progress of science." To constitute a truth you require, over and above a definition, a statement that certain consequences follow from the definition taken in conjunction with the various primitive propositions of the science in question. Now in a properly constituted science, the system of implications thus arrived at remains identically the same, no matter which of the possible sets of definitions and axioms we start from. E.g., in arithmetic there is some latitude of choice as to what concepts we shall assume primitive and indefinable, but the system of arithmetical propositions to which we are conducted is in any case the same. Thus, all that is arbitrary is not the recognition of certain propositions as true, but only the selection from among them of the group which are to serve as unproved premises for the deduction of the rest. But the full discussion of this subject would clearly call for a further article on axioms and definitions.