AN ESSAY ON The Psychology of Invention in the Mathematical Field BY JACQUES HADAMARD DOVER PUBLICATIONS, INC.Copyright © 1945 by Princeton University Pres AIL rights rewrved “under Pan American and International Copyright Conventions Published in Canada by General Publishing Com- pany Led. 30 Lesmill Road, Don Mills, ‘Toronto, Omatio. Pablished in the United Kingdom by Constable and Company, Lid, 10 Orange Steet, London we ‘This Dover edition, frst published im 1954 fan unaltered and unabridged seprint of the en larged (1949) dition, Te is reprinted by special arrangement’ with Pringcton ‘University Pres. International Standard Book Number: 0486:20107-4 Library of Congress Catalog Card Number: 34-4731, Manufactured in the United States of America Doves Publications, Tne 180 Varick Strect New York, N.Y. 10014,Ala compagne de ma vie et de mon awore.CONTENTS nvrRopuction 1. GENERAL VIEWS AND INQUIRIES Il, DISCUSSIONS OX UNCONSCIOUSNESS II, THE UNCONSCIOUS AND DISCOVERY IV, THE PREPARATION STAGE. ‘LOGIC AND CHANCE THE LATER CONSCIOUS WORK Vi, DISCOVERY AS A SYNTHESIS. ‘THE HELP OF SIGNS IX, THE GENERAL DIRECTION OF RESEARCH APPENDIX I APPENDIX IL APPENDIX III an 29 43 36 64 100 ue 1a 133, 387 1s. 144INTRODUCTION Concernine the title of thi ful. We speak of invention: it would be more correct to speak of discovery. The distinction between these two words is well known: discovery concerns a phenomenon, a law, a being which already existed, but had not been per- ceived. Columbus discovered America: it existed before him; on the contrary, Franklin invented the lightning rod: before him there had never been any lightning rod. Such a distinction has proved less evident than appears at first glance. Toricelli has observed that when one inverts 1a closed tube on the mercury trough, the mercury ascends to a certain determinate height: this is a discovery; but, in doing this, he has invented the barometer; and there are plenty of examples of scientific results which are just a8 much discoveries as inventions. Franklin’s invention of the lightning rod is hardly different from his discovery of the electric nature of thunder. This is a reason why the afore- said distinction does not truly concern us; and, as a mat- ter of fact, psychological conditions are quite the same for both cases. On the other hand, our title is “Psychology of Invention in the Mathematical Field,” and not “Psychology of Mathematical Invention.” It may be useful to keep in mind that mathematical invention is but a case of invention in general, a process which can take place in several domains, whether it be in science, literature, in art or also tech- nology. ‘Modern philosophers even say more. They have per- study, two remarks are use-xi INTRODUCTION ceived that intelligence is perpetual and constant inven- tion, that life is perpetual invention. As Ribot says,! “Invention in Fine Arts or Sciences is but a special case. In practical life, in mechanical, military, industrial, com- mercial inventions, in religious, social, political institu- tions, the human mind has spent and used as much imagi tion as anywhere else”; and Bergson,* with a still higher ‘and more general intuition, states: “The inventive effort which is found in all domains of life by the ereation of new species has found in mankind alone the means of continuing itself by individuals on whom has been bestowed, along with intelligence, the fac- ulty of initiative, independence and liberty.” Such an audacious comparison has its analogue in Met- schnikoff, who observes, at the end of his book on phagocy- tosis, that, in the human species, the fight against microbes is the work not only of phagocytes, but also of the brain, by creating bacteriology. One cannot say thet various kinds of invention proceed exactly in the same way. As the psychologist Souriau hi noticed, there is, between the artistic domain and the scien- tiffe one, the difference that art enjoys a greater freedom, since the artist is governed only by his own fantasy, so that works of art are traly inventions, Becthoven’s sym- phonies and even Racine’s tragedies are inventions. The scientist behaves quite otherwise and his work properly concerns discoveries. As my master, Hermite, told me: “We are rather servants than masters in Mathematics.” Although the truth is not yet known to us, it preexists and 1 See Dela de Payehotogis, bid, p. 4 i, L'Invention et le Géale (In G. Dumas’ Nouveau Praité Vol. VI), p- 449.INTRODUCTION xii inescapably imposes on us the path we must follow under penalty of going astray. ‘This does not preclude many analogies between these two activities, as we shall have occasion to observe. These analogies appeared when, in 1987, at the Centre de Syn- thise in Paris, a series of lectures was delivered on inven- tion of various kinds, with the help of the great Genevese psychologist, Clapardde. A whole week was devoted to the various kinds of invention, with one session for mathemat- ies. Especially, invention in experimental sciences was treated by Louis de Broglie and Bauer, poetical invention by Paul Valéry. The comparison between the circum- stances of invention in these various fields may prove very fruitful. It is all the more useful, perhaps, to deal with a special case such as the mathematical one, which I shall discuss, since it is the one I know best. Results in one sphere (and ‘we shall see that important achievements have been reached in that field, thanks to a masterly lecture of Henri Poin- caré) may always be helpful in order to understand what happens in other ones.I. GENERAL VIEWS AND INQUIRIES ‘Tue svpszcr we are dealing with is far from unexplored ‘and though, of course, it still holds many mysteries for us, ‘we seem to possess fairly copious data, more copious and more coherent than might have been expected, consider- ing the difficulty of the problem. ‘That difficulty is not only an intrinsic one, but one which, in an increasing number of instances, hampers the progress of our knowledge: I mean the fact that the sub- Ject involves two disciplines, psychology and mathematics, ‘and would require, in order to be treated adequately, that one be both a psychologist and a mathematician, Owing to the lack of this composite equipment, the subject has been investigated by mathematicians on one side, by psycholo- gists on the other and even, as we shall see, by # neurolo- ist. e always in psychology, two kinds of methods are avail- able: the “subjective” and the “objective” methods.* Sub- 11 speak of objective or introspective methods. I se that the modern epective mm James and Titchener), as though these were two different tciences, difering as to their object, while It seems to me that both Kinds of tnethods of observation could be applied and even help each other for the study of the same psychological processes. I understand, that for the behaviorist, the object of Introspection, Le, thought cloasness, Is to be ignored. ‘Already, in older times, the prominent biologist Le Dantee eliminated ‘consciousness by qualifying it iphenomenon.” 1 have always considered that an unscientific attitude, because if consciousness were fan epiphenomenon, it would be the only epiphenomenon in nature, where ‘everything reacts on everything else. But, epiphenomenon or not, it exits ‘and can be observed. We are not unjustified in presenting tuch observa2 GENERAL VIEWS jective (or “introspective”) methods are those which ‘could be called “observing from the inside,” that is, those where information about the ways of thought is directly obtained by the thinker himself who, looking inwards, re- ports on his own mental process. The obvious disadvantage of such a procedure is that the observer may disturb the very phenomenon which he is investigating. Indeed, as both operations—to think and to observe one’s thought— are to take place at the same time, it may be supposed a priori that they are likely to hamper each other. We shall see, however, that this is less to be feared in the inventive process (at least, in some of its stages) than in other men- tal phenomena. In the present study, I shall use the results of introspection, the only ones I feel qualified to speak of. In our case, these results are clear enough to deserve, it would seem, a certain degree of confidence. In doing so, I face an objection for which I apologize in advance: that is, the writer is obliged to speak too much about himself. Objective methods—observing from the outside—are those in which the experimenter is other than the thinker. Observation and thought do not interfere with each other tions, made by ourselves or by others, as T shall do in the course of this instances considered by behavloriste (I found them in J. B. Watson's Hehavioriom) are very different from those hich may concern us being generally taken from thoughts having « irect relation with our bodily sensations and which are more easily Interpreted in terms of the doctrine than others. In such cases, cor- respondences between bodily phenomena and states of consciousness fare easly seen and are more of lest known things. They are more hid~ den for cases of abstract thought, such as those we are going to study; rho reason why they should not be discovered at some future may happen, for Instance, with the help of the electele ‘waves which accompany cerebral processes (a suggestion which T take from an article of Henri Laugier in the Revue Moderne, reproduced Im his book Service de France au Canada).