CHAPTER XVII.
PRAGMATISM AND PHYSICAL SCIENCE.

§ 307. Science not infrequently protests against the encroachments of the philosopher who wishes to interpret its theories in a metaphysical sense. 'You keep to your province and I will keep to mine,' it says to the metaphysician; 'I am concerned only with phenomena, their sequences and their quantitative relations you are concerned with substance and accident, matter and form, cause and effect in a word, with reality itself. Let us be content to leave one another alone and to keep our provinces distinct.'

This principle might work very well, provided the province of science could be completely shut off from the realm of philosophy, and provided neither philosophers nor scientists were anxious to scale the dividing wall. Unfortunately, neither condition is completely fulfilled. The respective provinces of science and of philosophy coincide in two points. In the first place, the phenomena whose variations are measured, classified, and to some extent explained by the scientist, are precisely the same as those whose ratio essendi and ratio cognoscendi the philosopher seeks to discover; while the methods which the scientist uses presuppose the principles of sound logic, a discussion of which clearly lies within the sphere of the philosopher, quâ logician. And in the second place, often enough the scientist is not content to keep to his science. He insists upon his right to speculate metaphysically on the validity of scientific theories and on the ultimate nature of the phenomena with which he has to deal, and it is by no means uncommon to find that he mingles his metaphysical assumptions with the methods and principles of his science.

Now, if a particular scientist wishes to indulge in metaphysical speculations it is not for the philosopher to say him nay. It is only natural that one who devotes himself to the study of the laws which govern the phenomenal world should desire to know what phenomena are, and should form for himself a metaphysical theory of the universe. But if the scientist constructs a metaphysical theory, he can hardly complain should the philosopher criticise that theory which is not the less metaphysical because it comes from the pen of a scientist. Physicists such as M. Duhem place themselves beyond the reach of the metaphysician by denying that their theories have in any sense a metaphysical import; but there are others who, in discussing the methods of science and the validity of its laws, have taken up a definite metaphysical position, which, they tell us, is more compatible with, if it is not actually presupposed by, the principles of science. This attitude, which is becoming more and more prevalent both in Germany and in France, is closely connected with the pragmatic movement. The pragmatic method claims to be based on that of science, and not a few scientists seem in return much inclined to adopt as their own the pragmatic theory of knowledge-in-general and the philosophy of Pure Experience with which it is so intimately bound up. Science has given up the naive, uncritical and often materialistic Realism which was formerly its customary attitude, and in its stead many of the devotees have taken up, not the non-metaphysical position of M. Duhem, but the position of Empirical Idealism.

§ 308. Fifty years ago every scientist started from the common-sense point of view, assuming with his less educated brethren, that material things really exist independently of the exercise of mental activity. He took it for granted that his thoughts about the universe did not affect the nature of the 'facts' with which he had to deal. He did not trouble about the possibility of there being any a priori forms of the mind to which experience, consciously or unconsciously, had to conform; nor did he dream that in observing facts he was in reality making them. The aim of scientific research was to give an explanation, not only of the relations holding between phenomena, but also of the nature of the universe itself. Both mechanists and dynamists hoped to find an interpretation of the objective, real world at least in so far as it is material. Their atoms and molecules and their centres of force were real entities constitutive of material things and giving rise to those phenomena which we perceive by the senses. In fact, the complaint which the mechanist found with the dynamist was that the latter introduced into reality an unknown entity 'force,' which could neither be imagined nor defined.

Nowadays the position is changed. By many of our leading scientists the older metaphysics has been discarded and an Empirical Idealism or Pragmatic Sensationalism substituted in its stead. Mind and matter, relatively independent, are no longer the metaphysical conditions of scientific knowledge. For matter has been substituted sensation, and instead of knowledge arising through the manifestation of objective reality to a relatively passive mind, knowledge is now said to be due for the most part to the constructive activity of thought, to "l'action pensée," to ideas due, in part at least, to the creative power of mind, and striving to realise themselves in the field of sense-experience. The data of modern science are sensations; its aim is to discover the relations which hold between them; the means by which it seeks to acquire this knowledge is first of all sense-experience, in which experiment plays an important part, and, secondly, a mental activity of a higher order in which spontaneity and choice are conspicuous. Through the senses we have experience of relations between phenomena or sensation-complexes, and through the instrumentality of definitions and hypotheses created by thought we endeavour to arrange and classify these relations, to subsume them under general forms, and, if possible, to reduce them to unity by the discovery, or better, perhaps, the invention of some primary relation which holds throughout.

§ 309. Two names stand out prominently as representative of this attitude, at once metaphysical and epistemological, in regard to the scope of science. They are those of Mach and Karl Pearson; and to these we may add a third, chosen from the more sceptical school of the Philosophie de la Contingence, M. Le Roy. M. Poincaré, on the other hand must be placed in a different category, for he admits the 'objectivity' of fact, and even to 'laws' assigns a certain 'normal objectivity,' though in certain passages he seems to speak as if he were a sensationalist like Mach.^{1}

Mach distinguishes three stages in scientific procedure, the experimental stage, in which we are in immediate contact with reality, i.e., with sensation, and merely tabulate the results of experiment and observation; the deductive stage, in which we substitute mental images for facts, as in Mechanical Physics; and the formal stage, in which our terms consist of algebraical symbols, and our aim is to construct by their means the most convenient and most uniform synopsis of results. Similarly Poincaré distinguishes three kinds of hypotheses, (1) hypotheses suggested by facts and verified at least peu près in experience, (2) 'indifferent hypotheses,' which are useful in that they express under images and figures relations between phenomena, but which are neither true nor false; and (3) mathematical conventions,' which consist of definitions more or less arbitrary, and which are independent of experience.^{2} Poincaré's 'indifferent hypotheses' correspond to Mach's second stage in the development of science, and manifest a tendency eventually to disappear. Already Mach himself prefers to dispense with their service as rather encumbering than facilitating thought; while Poincaré though he considers them still indispensable for the moment, holds them to be devoid of real significance.

§ 310. Thus the second stage of scientific procedure in this view is of secondary importance. It is the experimental and mathematical stages that really constitute science. By observation and experiment we are brought into contact with reality; not indeed with the material world, for no such entity is supposed to exist; nor even with the world of sensible appearances strictly so-called -- for an appearance implies something that appears, -- but with sensations. The objective condition of scientific knowledge, the reality which in science we desire to know, is sensation. The data of experience are sensations. Mach, in his Analyse der Empfindungen und das Verhältniss des Physischen zum Psychischen, has developed this view at considerable length. Sensations and sensation-complexes -- these, he says, are reality. All science consists in the analysis of sensations. Nature is composed of elements given by the senses. From these we choose those which are most important for practical purposes and call them 'objects' or 'things.' But 'things' are really abstractions, and a name is a symbol for a complex of sensations whose variations we neglect. There are no things-in-themselves, nor are sensations symbols of things, but what we call things are symbols of sensation-complexes of relative stability. Colours, sounds, pressures, spaces, durations, these are the real things. All thought is governed by the principle of Thought-Economy. We are ever trying to save ourselves trouble. Hence we have acquired the habit of grouping sensations together in a lump and calling them by a single name. One group of sensations we call 'water,' another 'a leaf,' another 'a stone.' Smaller groups, again, are combined to form larger ones. The group 'leaf' is joined to the groups 'branch,' 'stem,' etc., and the whole, being vaguely or generically pictured, becomes a 'plant.' These larger groups, again, are included in others larger still. What we call 'the external world' comprises all those sensation-complexes which are relatively constant, i.e., which repeat themselves again and again in the same sort of way and are not subject to the control of our will: whereas 'the self' comprises that other very extensive group of sensation-complexes, some of which are always present in consciousness, though ever varying in tone, while others can be produced at any time if we so desire, and thus are directly under our control.^{3}

In the external group the relations between sensation-complexes are constant, i.e., the complexes follow one another in the same order. Thus the sensation-complex (water) is juxta-posed in a certain way to another complex (bunsen-burner), and always after a certain time the bright transparency of the former complex gives place to a dull whiteness of another considerably greater in extent. Ordinarily, however, we prefer -- according to the Princip der Denkökonomie -- to use names to denote our sensation-complexes, as it saves us the time and trouble of describing them. The usual account that one would give of the above phenomenon, for instance, would be that when we heat water over a bunsen-burner after a time it begins to boil. Indeed, it would be very awkward for the sensationalist, if he often had to carry out Pascal's principle of substituting the definition for the thing defined.

§ 311. The physicist selects the above class of sensations, which are characterised by greater stability, greater regularity, and are common to mankind, as the data of his scientific researches while the psychologist treats of these in another way, and also of other sensations which are less stable, more subject to the control of the will, and, hence, often peculiar to the individual. But the standpoint of Mach is really psychological throughout. Both psychologist and physicist treat of the same class of objects from different points of view.

All that we can know of the world is necessarily reduced to sense-perception: and all that we can wish to know is given in the solution of a mathematical problem, in the knowledge of the functional dependence which exists between sense-elements. This exhausts the sources of the knowable.^{4}

Professor Mach has given up the apparently hopeless task of reducing things to indefinitely small and ultimate elements. Both he and Poincaré prefer to regard atoms and such like as hypotheses, as mere picturesque fictions of greater or less utility, but of no objective value; and for things Mach substitutes sensations. Scientifically, indeed, sensation is regarded as a form of energy, the differences of which are probably quantitative. But in course of time, says Mach, we shall discover that the sense of hunger is not so very different from the action of sulphuric acid on zinc, and that our will is not so very different from the pressure of stone on its support, and so we shall get nearer nature. Thus, for Energetics, everything is reducible to energy, alias sensation, and the final aim of physical science is to demonstrate the truth of this assertion.

§ 312. Both Mach and Poincaré speak of the sense-data of science as if they were uninfluenced by the subjective factor in cognition. They regard them as relatively stable, independent of the individual, and therefore objective. But even in the objects of scientific knowledge, philosophers such as M. Le Roy would admit an element of 'contingency.' Sensations are not given in isolation, but are grouped together in complexes and integrated into percepts and in the construction of our percepts there may enter an element of caprice. We are influenced by our point de vue choisi d'avance, practical utility in some cases, the exigencies of scientific theory in others. Hence we introduce into our percepts just what suits our convenience and leave out the rest. This follows logically from the philosophy of Pure Experience, a philosophy which is practically identical with the metaphysical standpoint of MM. Karl Pearson and Mach. For if, as M. Le Roy says, "nothing is put before the mind, except what is put by the mind;" if, in other words, we do not copy reality, but construct it, as Dr. Schiller affirms then all is due to "hypothesis and fabrication" either by the individual or by the race, i.e., we construct our percepts as well as our concepts. Again, racial development takes place by individual variation, and this is possible in the sphere of experience only if thought exercises purposive control over the data of sense, in which case even in this, the lowest stage of human knowledge, we must admit that there is an element of caprice. Hence,

all scientific laws are unverifiable, to put the matter rigorously, first because they are the instrument with which we make in the continuity of the primitive datum the indispensable parcelling out (morcelage) without which thought remains powerless and shut in, and again because they constitute the criterion itself with which we judge the apparatus and methods which it is necessary to use in order to subject them to an examination, the accuracy of which may be able to surpass all assignable limits.^{5}

§ 313. Contingency and choice in the sphere of experimental science is emphatically denied by Poincaré "All which the scientist creates in a fact," he says, "is the language in which he expresses it." We do not interfere with facts, except in so far as we select those which are relevant to our purpose. In experience relations are determined, not by experiment, but by inexorable laws which govern the succession of our sensation-complexes. 'We do not copy reality' -- that is true; but the laws which govern the sequences and combinations of sensations are fixed for us, and not by us. They are something which we experience as a datum, not something we arbitrarily construct; and these laws may be known by us at least à peu près.

This view, though doubtless the correct one, is hardly consistent with the doctrine that sensations and not material objects are the data of science. If it be the mind that groups sensations together and so forms sensation-complexes or objects, then, as M. Le Roy and Dr. Schiller affirm, such groupings may not always be precisely identical. Not only may modification and even mutilation of fact have occurred during the long process in which habits of perception have been built up and have become common to the race, but such modifications are still possible since habits are only relatively constant and only approximately common to the race. Moreover, the significance of M. Poincaré's assertion that all we create in a fact is the language, in which we express it, is considerably modified when we compare it with another statement to the effect that 'language is strewn with preconceived ideas;' for the latter, since their influence is unconscious, are far more dangerous than those which we deliberately formulate and make use of in hypotheses.

M. Le Roy's statement, therefore, that scientific laws are unverifiable because they are the instruments by means of which we parcel out the primitive datum of experience would seem to be valid in a pragmatic and evolutionary theory of knowledge. His second argument (granting the validity of his premises) is no less conclusive. When the correspondence-notion of truth is rejected, our only criteria of truth are utility and consistency, both of which are determined by the development and systematisation of science itself. Scientific laws, as M. Duhem has pointed out, mutually involve and imply each other's truth. Hence, if in no individual case we can eliminate the subjective element and so prove that a law has arisen from the manifestation of reality itself to our minds, we have no right to assume one law to prove another; all laws, whether empirical or not, will be equally unverifiable in the pragmatic and pseudo-scientific theory of knowledge.

The unrestricted jurisdiction of the Princip der Denkökonomie points to the same conclusion for, according to Professor Mach, this principle is not confined to the realm of physical theory, but is a general principle applicable to all forms of cognition alike. It governs the construction of the percepts and concepts of common-sense, just as it directs the scientist in the formulation of definitions and physical hypotheses. Efficiency depends upon economy, and efficiency, adaptation to environment, and practical utility for the control of sense-experience is the final aim, not only of physical theory, but of all human cognition.

§ 314. The Pragmatism and Sensationalism of Mach and Karl Pearson, which is really a philosophical theory of knowledge, must be carefully distinguished from the view that in Physical Theory definitions and laws are merely symbolic formulae, useful for the classification, co-ordination and systematisation of scientific fact: for this view is held by many who, except on this point, are in no sense pragmatists either in regard to science or philosophy. A pragmatic interpretation of physical theory is, in fact, quite compatible with metaphysical Realism.

For instance, M. Duhem is a realist in regard to the notions of common-sense, yet he tells us that the aim of physical theory is "to construct a symbolic representation of what our senses, aided by instruments, make us know, in order to render easier, more rapid, and more sure, reasoning about experimental knowledge." Concepts for him as for MM. Poincaré and Mach are means to this end. Their function is symbolic. As definitions they are arbitrary, and in no way represent reality or reveal its inner rational structure. 'Masses' are "coefficients which it is convenient to introduce into our calculations." 'Energy' must not be confused with the force exerted by a horse in drawing a cart; it is merely "the function of the state of a system whose total differential in every elementary modification is equal to the excess of work over heat set free."^{6} Concepts as definitions form the basis of scientific deduction, but they do not reveal the nature of objective facts. The most they can do is to indicate certain experiences, and so enable us to verify the phenomenal relations which we have deduced by means of mathematical reasoning in which these symbolic definitions function as terms.

Some have endeavoured to find a similarity between M. Duhem's theory of chemical combination and the scholastic doctrine of matter and form. This, however, as he informs us in his work entitled Le Mixte et la combinaison chimique, in which his views on that subject are developed, is merely an analogy, and nothing more. 'Forms,' as conceived by the chemist and the physicist, are quantitative, not qualitative; whereas quality is of the essence of things, the nature of which it is the business of the metaphysician and not of the scientist to determine. Nevertheless, in spite of this denial that 'forms' in chemistry and in physics are comparable with the metaphysical forms of Aristotle, M. Duhem's standpoint is quite compatible with Realism; and it is so precisely because he relegates all questions as to the nature of quality and essence to Metaphysics.

§ 315. The standpoint of M. Duhem differs essentially, therefore, from that of Karl Pearson and Mach; for, while carefully distinguishing physical theory from physical fact, M. Duhem does not identify the latter with sensation, but leaves it to the metaphysician to determine the ultimate nature of the data of experience. Again, it is only in Theory that postulation and symbolism are admitted by M. Duhem, and that we are allowed to construct and modify definitions at will. Mathematical Physics in the course of its development is independent of Experimental Physics, and uses a different method. In the latter we are bound down by empirical facts, whereas Mathematical Physics is free to disregard all facts till theory is complete, when it must be verified as a whole by comparing the conclusions which have been mathematically deduced with the complexus of experimental data. "In the course of its development a physical theory is free to choose whatever way it pleases, provided it avoids all logical contradiction; in particular, it is free to disregard the facts of experience."^{7}

On the other hand, for Professor Mach, and apparently for M. Poincaré also, symbolism, postulation and the principle of Thought-economy apply to theory and fact alike. The experimental differs from the mathematical stage only in this respect, that in the former we group under one name sensations which are actually present in consciousness, and our grouping is more or less spontaneous; whereas in the latter we arbitrarily combine symbols denoting sensation-complexes already grouped, and postulate that the new symbol shall denote actual groupings which have never as yet been given in consciousness.

The real difference, then, between Karl Pearson and Mach on the one hand and Duhem on the other is in regard to their philosophic standpoint. Both Karl Pearson and Mach, and, to some extent, Poincaré also, philosophise on the data of experience and on the development of knowledge in general; and their philosophy is pragmatic. M. Duhem declines to philosophise, and, if a pragmatist at all. is a pragmatist only in regard to the methodology of physical theory, an attitude which is quite consistent with philosophic Realism.

§ 316. There is also a further difference between the views of M. Poincaré and M. Duhem in regard to the relation of Mathematical to Experimental Physics. M. Poincaré admits "truths founded on experience and verified almost exactly so far as concerns systems which are practically isolated; " and these truths, he says, when generalised beyond the limits within which experience verifies them, become "postulates, applicable to the whole universe and regarded as rigorously true."^{8} "Mais, le principe désormais crystallisé pour ainsi dire, n'est plus soumis au contrôle de l'expérience. Il n'est pas vrai ou faux, it est commode." Thus such principles as Newton's Laws of Inertia and of the Equality of Action and Reaction, Lavoisier's Conservation of Mass, Mayer's Conservation of Energy, and Carnot's Degradation of Energy are axiomatic, though not a priori. They are suggested by facts, but are unverifiable, because in their absolute form they are mere conventions; and our right to postulate them lies precisely in this, that experience can never contradict them.

Mathematical Physics, on the other hand, for M. Duhem is entirely independent of experience throughout the whole process of its development. No hypothesis whatever can be verified till the theory of Physics is complete in every detail, for every physical law is "a symbolic relation the application of which to concrete reality supposes that one accepts quite a system of other laws."^{9} No individual, physical law is, properly speaking, either true or false, but only approximate, and on that account provisional. Sufficiently approximate to-day, the time will come when it will no longer satisfy our demand for accuracy.^{10} Principles, therefore, which MM. Milhaud, Le Roy and Poincaré alike place beyond the control of experience are, says M. Duhem, either not physical laws at all (since every physical law must retain its meaning when we insert the words à peu près, which these do not) or else, when their consequences have been fully deduced, they must be rigorously subjected to the test of experience in the theory to which they belong, and with that theory stand or fall.^{11} In other words, Poincaré, admitting the existence of relatively isolated systems of experimental facts, thinks that it is possible to apply the process of verification to a physical theory in the course of its development; while Duhem, convinced that all physical laws are intimately connected, prefers to formulate a complete and self-consistent system of hypotheses before attempting to compare the consequences of any one of these hypotheses with experimental fact. A similar difference is manifest in regard to the method of teaching Physics. Poincaré prefers the inductive and experimental method. Duhem holds that physical theory should be presented to those who are capable of receiving it, in toto, and that experiments should serve merely as illustrations of different stages in its development.

§ 317. This difference between two of our most eminent physicists, though great at first sight, can, I think, to some extent, be explained. M. Duhem insists that all hypotheses must be verifiable à peu près if they are to have physical significance: consequently, there can be no laws in physical theory, when complete, which are not at least approximately true. On the other hand, the use of purely conventional hypotheses in the construction of a theory is allowable, provided they are ultimately verifiable in their systematic completeness. M. Poincaré points out that such conventional hypotheses are often experimental laws generalised beyond the limits within which they are verifiable, and so worded that they cannot, as such, be contradicted by experience. That in their most general form, as applicable to the whole universe, universal postulates of this kind cannot be verified directly, is obvious; for not only are they universal, but they are expressed in symbolic terms, such as energy and inertia, terms which it is almost impossible to translate into their corresponding sensations (if such there be). Yet, inasmuch as such postulates lead to particular conclusions about less abstract realities of which we can have immediate experience, inasmuch as their function is to guide us in the construction of hypotheses which are verifiable à peu près, and so have a physical sense, it may be said that even the most abstract laws and the most general principles can be verified indirectly through their consequences.

§ 318. Thus the divergent views taken by M. Poincaré and M. Duhem respectively in regard to physical 'axioms' are apparently reconciled. Yet the divergence has, I am inclined to think, a deeper root. M. Duhem's reason for affirming that physical laws are not as yet either true or false, but simply useful, is that they have not been and cannot as yet be completely verified, and hence are only provisional. But when they have been completely verified in the physical theory to which they belong they will become, together with the principles upon which that theory is founded, not only useful but true in the realistic sense of that term. M. Poincaré, on the other hand, can easily find room for 'axioms' which are already 'regarded as true,' because truth for him, as for the pragmatist, is apparently one with utility, though, as we shall see, it is not merely utility. Truth and objectivity, in the sense in which they were understood by Aristotle, by the Scholastics, and in which they are still understood by common-sense, do not exist for those scientists who are prone pragmatically to philosophise. Objects being identified with sensations,^{12} objectivity acquires a new sense; for the objective value of science is, according to M. Poincaré, the same as our belief in external objects. Granted that nothing is put before the mind except what is put by the mind, granted that "all that is not thought is pure nothing since we can think only thought, and the words we use in speaking of things can only be expressions of our thoughts," granted that "to say that there is anything besides thought is a statement that can have no meaning,"^{13} it would seem to follow that physics is merely a branch of psychology, and that in studying things we are really studying the work our own minds.

§ 319. What sense, then, can be given to the term objective truth? How can there be anything really objective or really true, if sensations and the laws which determine their relations proceed from the nature of the mind itself and refer to nothing beyond? Pragmatists attempt to solve the difficulty by distinguishing between what is peculiar to the individual and what is common to the race. Doubtless the relations between sensations "could not be conceived outside a mind which conceives them or which feels them, but they are, nevertheless, 'objective,' because they are, will become, or will remain common to all thinking beings."^{1} Some sensations "nous apparaissent comme unies entres elles par je ne sais quel ciment indestructible et non par un hasard d'un jour."^{2} These are the sensations which are common to all thinking beings, and which in this sense may be said to be objective. From these premises the nature of truth and the criterion of truth may easily be deduced. That will be true which the individual holds in common with his fellow-men, that false which he holds in isolation and the function of truth-criteria will be to distinguish what is held collectively from what is peculiar to the individual. The Philosophie de la Contingence, realising this, has defined truth as that which is 'normal' enough to be accepted by anyone of sound mind.

§ 320. The connection between the conception of truth as that which is useful, fruitful, commode, and the conception of it as 'normal objectivity' is clear. In the first place, the relation of intellect to sense is in this view regarded as merely 'functional.' Intellect symbolises and relates sense-data solely with a view to controlling them, manipulating them, and adapting them to the needs of human nature. In so far as the hypotheses which we formulate fulfil this end, they perform their function well, and are useful. The needs of humanity, however, and consequently the reactions which will satisfy those needs are more or less constant; hence there will be a certain degree of harmony in our judgments of what does satisfy our needs, i.e., we shall have a kind of 'normal objectivity' based on common consent as to the usefulness of certain hypotheses. Secondly, science, like all other knowledge, is in process of evolution. Mind interacts on mind through social intercourse. The thought-habits of the individual get modified and become more like those of his neighbour. Thus needs-for-knowledge and the processes by which we strive to acquire that knowledge have, in course of time, become more or less identical throughout the race. The Newtonian laws, says Mach, are the result of a long process of observation and experiment, of scientific evolution, in the race. Ideas, not found in sense-data, and whose working is often unconscious, guide the development of science. Now the psychological factor in cognition has obtained a fuller recognition, and the naive Realism of Mechanical Physics is being displaced by Energetics. We find now that what we took to be true, in the copy sense of the word, merely useful; useful, not because it satisfies the needs of the individual, but because it satisfies the needs of humanity at large, because it conforms to our preformed habits of thought, and enables us to go on adapting ourselves to our environment. It is however, only a pragmatic truth, this utility; a truth based on common needs, but always incomplete, always changing, never absolute; always provisional and approximate, subject to continual modification as human needs and human environments advance together on the wave of evolution.

§ 321. The doctrines of Poincaré, Mach and Le Roy, since they deal with truth, knowledge and objectivity in general, not merely with the scientific use of these terms, do not belong to physics, or, at any rate, not to physics alone, but to the theory of knowledge. They are philosophical doctrines which bear clearly the impress of Pragmatism; and will be discussed together with other pragmatic theories in the chapters entitled "Pragmatic Truth." It would be an error, however, to attribute the Pragmatism of Poincaré, Mach and Le Roy to the influence of Professor James and Dr. Schiller, though it can hardly be doubted that the world-wide reputation and the vigorous polemics of the latter have had their effect. The Pragmatism of the philosophising scientist has arisen from quite another source. It is, in fact, but a development of the doctrine that the definitions and laws which belong to Physical Theory are only symbolic formulae, figured hypotheses, postulates which are useful, but not true. And this doctrine can only be called 'pragmatic' retrospectively, for it existed before Pragmatism proper was invented, and by it -- at least in Dr. Schiller's case -- the general epistemological theory of Pragmatism appears to have been suggested.

§ 322. The origin of scientific Pragmatism -- if such a term is permissible -- arises apparently from the feeling of hopelessness aroused in the scientific mind by the multiform variety of physical theories at present existing, theories which, if taken as they stand, are mutually incompatible. Contradiction is so violently repugnant to the human mind that we are ready to take almost any means to escape it. And to many scientists, anxious to solve the mystery presented by a daily-increasing multitude of conflicting and seemingly irreconcilable hypotheses, only two alternatives seemed to offer themselves, either Scepticism or a kind of provisional Pragmatism. Believing themselves confronted with a choice such as this, many have adopted the latter alternative, i.e., rather than give up truth altogether, they prefer to regard it in Physical Science as a 'value.' Provisionally, at any rate, truth is not something which corresponds with reality or reveals its inner nature, but something which enables us the better to manipulate and control our environment. The reason of this choice is, I think, not the intrinsic reasonableness of the pragmatic position, but rather the reluctance which is felt towards the sceptical alternative. Scepticism would not satisfy our practical needs. It would involve the surrender of laws in practice extremely useful, and theoretically, at any rate possible, and sometimes probable. It would deny even approximate truth, and would destroy the possibility of certitude; and so would take away all motive for further research, rendering science both futile and meaningless. In the face of such an alternative, the scientist prefers to take up a position which admits of at least some degree of certitude, and gives, at any rate, some meaning to the notions of truth, validity and objectivity.

§ 323. But the adoption of a pragmatic attitude in regard to the function of physical theory, whether it be merely provisional or not, by no means implies the adoption of Pragmatism in general, with its idealistic interpretation of the universe, its theory of universal postulation, and its man-made truths which at bottom are no truths at all, but merely 'values.' Scientific Pragmatism is, as I have said, quite consistent with philosophic Realism, and the two are not infrequently found together, and were so found before philosophical Pragmatism had been conceived. One may be permitted to ask, therefore -- and the question is both interesting and instructive from the epistemological point of view -- whether some form of Realism is not still compatible with Physical Science itself; whether, for instance, the physicist in his aim should not be a realist after all nay, more, whether some physical principles and some branches of physical theory do not already admit the predicate 'true' in the correspondence sense of that term. It would be strange, indeed, if in one of the most extensive, most important, most fruitful spheres of human research, to which for centuries many of the keenest human intellects have devoted themselves, we should have, as yet, obtained no knowledge whatsoever of objective reality. One can scarcely credit the assertions of those who maintain that the whole complexus of physical hypotheses and laws is merely a working instrument, a useful calculating machine. Fortunately, however, Realism in Physical Science is not yet extinct. Not only were scientists of ancient and world-wide fame, such as Newton, Huyghens, Bernoulli, Lagrange and Laplace, realists, but realists survive to the present day. Nor is their survival an anomaly. They do not exist in spite of their failure to adapt themselves to an idealistic environment. On the contrary, the Realism of English science has proved one of the chief obstacles to the advance of Anglo-German Absolutism. Most scientists who do not philosophise adopt a realistic attitude in regard to facts; and many who do, find it extremely difficult, if not impossible, to eliminate realistic expressions and ideas. And if Realism is compatible with Physical Science, it certainly gives to it a fuller, a more consistent and a more satisfactory meaning than can be got from Pragmatism. It is worth while considering, therefore, whether a moderate form of Realism may not yet be upheld in regard to Physics, in spite of the manifold diversity of its hypotheses; and, accordingly, to this question I propose to devote the following chapter.

{1} Cf. La Valeur de la Science, chap. xi., especially p. 266.

{2} La Science et L'Hypothèse, Introduction, p. 2, and cf. pp. 180, 181.

{3} La Mécanique, pp. 450 et seq. (The examples here given are my own).

{4} Mach, op. cit., p. 287.

{5} La Roy, Revue Métaph. et Morale, 1901, p. 140.

{6} Le Mixte et la Combinaison chimique, pp. 202-205.

{7} Duhem, La Théorie Physique, vi., § 7.

{8} Revue Métaph. et Morale, 1902. (Italics mine).

{9} op. cit., ix., 1.

{10} Ibid., §§ 2,3.

{11} Ibid., vi., §§ 8, 9.

{12} M. Poincaré is not always consistent on this point. Sometimes he seems to speak as if he were a realist, as in chap. x. of La Valeur de la Science, where he attacks M. Le Roy; somenes as a sensationalist (chap. xi., and cf. infra).

{13} La Valeur de la Science. p. 276.

{14} Ibid., p. 272.

{15} Ibid., p. 270.