Ideology and Rationality in the History of the Life Sciences Georges Canguilhem translated by Arthur Goldhammer “The MIT Press (Cambridge, Massachusets, London, EnglandEnglish translation copyright © 1988 Massachusetts Institute of ‘Technology (Originally published under the ttle Idéologie et rationalité dans Vistoire des sciences de la vies Nowwelles ire et de philosophie des sciences, copyright © 1977 Lit losophique J. Vein, Pats, France. [All rights reserved. No part ofthis book may be reproduced in any form or by any electronic or mechanical means, including photocopying, ‘recording, or information storage and retrieval, without permission in ‘writing from the publisher. “This book was typeset by Graphic Composition Inc. and was printed and bound by Halliday Lithograph in the United States of America. Library of Congress Cataloging-in-Publication Data ‘Canguithem, Georges, 1904— Ideology and rationality in the history ofthe life sciences. “Translation of: Idéologie et rationalité dans Uhistoire des sciences de Iavie. Bibliography: p. Includes index. 1 Life sciences—History, 2 Life sciences—Philosophy. 1. Title QH305.C2613 1988 574.09 H-610 ISBN 0-262-03137-X Contents ‘Translator’s Preface vil Preface ix Introduction: The Role of Epistemology in Contemporary History of Science 1 Scientific and Medical Ideologies in the Nineteenth Century ‘What Is a Scientific Ideology? 7 John Brown’s System: An Example of Medical Ideology 4 Bacteriology and the End of Nineteenth-Century “Medical Theory” sz ‘Triumphs of Biological Rationality in the Nineteenth and Twentieth Centuries “The Development of the Concept of Biological Regulation in the Eighteenth and Nineteenth ‘Centuries 8r‘On the History of the Life Sciences since Darwin The Question of Normality in the History of Biological Thought Sources Index CONTENTS 103 m5 147 149 Translator’s Preface Georges Canguilhem was born in 1904. He studied and began to teach philosophy but while teaching decided to work toward a medical degree. His reasons are worth noting: Ie is not necessarily to learn more about mental illness that a professor of philosophy will take an interest in medicine. Nor is it necessarily to practice a scientific discipline. What expected from medicine was nothing other than an intro- duction to concrete human problems. Medicine seemed to me then, and still seems to me now, a technique or an art at the crossroads of several sciences more than a science in the strict sense of the word. Two problems—that of the relation between science and technology, and that of norms and normality—could, I thought, be more precisely for. mulated and more fully elucidated by someone with med- ical training. . .. The present work {his 1943 thesis, The Normal and the Pathological) is therefore an effort to in- tegrate some of the methods and results of medicine into philosophical speculation. Canguilhem with his life's work has admirably ful- led this statement of intention. Along with Gaston Bach- lard he has been one of the primary influences in the reorientation of French philosophy in recent years. It was Bachelard who introduced the concept of an “epistemolog- ical break,” a concept whose importance and usefulness Canguilhem has demonstrated in his own way. But Can- guilhem’s work also shows how philosophy can span theviii coupure, so to speak, in order to reestablish continuity at another level, For Canguilhem, error is the truth of the past transcended, and he is able to show in concrete detail why the history of science should be studied not as a steady march toward truth but as a process of formation and re- formation of concepts and models. His method is more easily grasped in action than through description, and there is perhaps no better introduction to his work than the essay (included here) entitled “Bacteriology and the End of Nineteenth-Century ‘Medical Theo! In 1955 Canguilhem succeeded Bachelard as director of the Institut d'Histoire des Sciences et des Techniques, a position that he held until his retirement a few years ago. Pethaps the most noted of the younger philosophers influ- enced by his thought was the late Michel Foucault, who wrote of his debt to Canguilhem’s pioneering work. Inter- ested readers may wish to consult Le Normal et le patho- logique (1966, containing the 1943 thesis and later essays and now available in English), La Formation du concept de réflexe (x95 5, reissued 1977), La Connaissance de la vie (1952, 1965), Etudes d’histoire et de philosophie des sci- ences (1968), and the volume from which the present translation was made, Idéologie et rationalité dans I'his- toire des sciences de la vie (1977). Canguilhem also pro- vided a preface to a recent edition of Claude Bernard's Lecons sur les phénomnes de la vie communs aux ani- maux et aux végétaux (1966). ‘TRANSLATOR'S PREFACE Preface To err is human, to persist in error is diabolical. It is not up to me to decide the degree of error embodied in the texts gathered here. I am surely too old to make public confession of my mistakes, to proclaim my allegiance to newly instituted epistemological authorities at the cost of renouncing methodological axioms that I borrowed some forty years ago and subsequently exploited in my own way and at my own tisk, not without emendation, revision, and reorientation. In 1967-68, under the influence of work of Michel Foucault and Louis Althusser, I introduced the concept of scientific ideology into my lectures. This was not simply a mark of my interest in and acceptance of the original con- butions of those two thinkers to the canons of scientific history. It was also a way of refurbishing without rejecting the lessons of a teacher whose books I read but whose lec- tures I was never able to attend. For whatever liberties my young colleagues may have taken with the teachings of Gaston Bachelard, their work was inspired by and built on his, 1 do not believe, therefore, that the reader of my first Etudes d’bistoire et de philosophie des sciences (Studies in the History and Philosophy of the Sciences) will find inthese essays signs of change or evolution in my thinking.* As for the question whether my indifference to the devel- ‘opment of a history that would substitute for the distinc- tion between science and philosophy (or, in other words, between science and literature) a notion of their mutual interpenetration should or should not earn me the distinc- tion of being a “conceptualist fossil” I must admit I do not much care. When one’s own insignificant research has led ‘one to recognize the existence of discontinuity in histo: would be inappropriate to refuse to recognize discontinui- ties in the history of history. To each his own discontinuity, his own revolutions in the world of scholarship. (On the other hand I should like very much to answer question that has been raised by no one but myself. The author of The Archaeology of Knowledge, whose analysis of scientific ideology I have found quite useful, has distin- guished several “thresholds of transformation” in the his- tory of knowledge: a threshold of positivity, a threshold of epistemologization, a threshold of scientificity, and a threshold of formalization.’ In my published work I am not sure that I have distinguished as carefully as Michel Foucault might wish among the various thresholds crossed by the disciplines I have studied. It seems to me in any case that, the claims of certain geneticists notwithstanding, none of those disciplines has yet crossed the threshold of formalization. Unlike Foucault, however, I do not believe that experimental medicine as practiced by Claude Bernard and microbiology as practiced by Louis Pasteur were adequate in their cont ence of clinical medicine. I readily admit that I failed to pay adequate attention to the question of thresholds of * (Etudes d'histoire et de philosophie des sciences was published by Librairie Vrin in 1968. The French edition of the present work bears the subtitle Nouvelles études d'histoire et de philosophie des sciences.—Trans,] PREFACE transformation. But nineteenth-century medicine and bi- ology lend themselves less readily than, say, nineteenth- century chemistry to dissection of the conditions that made “progress” possible. One can irgue, I think, that Ber- nard’s physiological medicine exhibits a case in which “ep- istemologization,” at the hands of a Bernard himself in love with philosophizing, raced far in “advance” even of posi- tive empirical results. By contrast, Pasteur, a chemist rather than a physician, was primarily interested in making a pos- itive contribution to research and not unduly concerned with developing a consistent epistemology. It may be, finally, that my analyses are not sufficiently subtle or rigorous. I leave it to the reader to decide whether this is a question of discretion, sloth, or incapacity. Notes See Michel Foucault, L’Archéologie du savoir, pp. 243-247. . Woodger, Axiomatic Method in Biology (Cambridge: 1937), and “Formalization in Biology,” Logique et analyse, new (August 1958). 3 Cf. B. Dagognet, Méthodes et doctrine dans V'oeuvre de Pasteur (Paris: Presses Universitaires de France, 1967), conclusion, Gc. June 1977 PREFACE,Ideology and Rationality in the History of the Life SciencesIntroduction: The Role of Epistemology in Contemporary History of Science ‘To anyone who would examine the relations between epis- temology and the history of science, one fact stands out above all others: namely, that we possess at present more manifestoes and programs of research than we do hard facts, Statements of intention are numerous, concrete te- sults meager. Compared with the history of science, a discipline with a history of its own, epistemology at first sight seems to find itself in a false position. Chronologically, the history of science owes nothing to the philosophical discipline that appears to have acquired the name epistemology in 1854, Montucla’s Histoire des mathématiques (1758), Bailly’s Histoire de lastronomie (1775-1782), and Kurt Sprengel Versuch einer pragmatischen Geschichte der Areneikunde (1792-1803) were all written without reference to any sys- tem of critical or normative concepts. No doubt all these works were informed, whether their authors were aware of it or not, by a period consciousness, impersonally for- mulated in the doctrine of infinite perfect ‘man spirit and based on an almost unbroken series of revolutions in cosmology, mathematics, and physi revolutions associated with the names Copernicus, G: Descartes, Harvey, Newton, Leibniz, and Lagrounds of continuity it was therefore legitimate to believe in further scientific progress to come. Although Sprengel (the date being 1792) explicitly alludes to critical philoso- phy in the introduction to his history of medicine, he men- tions it simply as a doctrine in which certain physicians happen to be well versed, just as certain of their predeces- sors were well versed in dogmatic, empirical, or skeptical philosophy, rather than as a new and effective instrument for judging the validity of scientific methods. Hence there is no point in reproaching cighteenth- and nineteenth- century historians of science for not having employed any of the epistemological concepts that today’s philosophers are attempting to enforce as rules for writing scientific history. ‘Among historians of science, those who dislike the scrutiny of their discipline by epistemologists have not been remiss in pointing out that epistemology, itself nour- ished by the history of science, cannot presume to give more than it has received; that is, it cannot pretend to re- form the principles of a discipline from which it in fact derived. The acrimony of the controversy is not unrelated, however vaguely or loosely, to the ancient view of the re- ines and the faculties of the soul, according to which history corresponds to Memory. It is hard to say whose ambition is more exorbitant, the histor- s’ or the epistemologists’. Which is more pretentious: to claim memory or judgment? Errors of judgment are acci- dental, but alteration is of the essence of memory. About reconstructions in the history of science one must make a point that has repeatedly been made about reconstructions in other fields of history—political, diplomatic, military, and so on: namely, that contrary to Leopold von Ranke’s dictum, the historian can never claim to represent things as they really were (wie es eigentlich gewesen jksterhuis’s comment that “the history of science forms not only the memory of science but also its episte- INTRODUCTION mological laboratory” has frequently been commented on2 Since elaboration is different from restitution, one may conclude that epistemology’s claim to give more than it has received is legitimate. Epistemology shifts the focus story of science to science as seen in the light of history. To take as one’s object of inquiry noth- ing other than sources, inventions, influences, priorities, simultaneities, and successions is at bottom to fail to distinguish between science and other aspects of culture. A history of science free of epistemological contamination ‘would inevitably reduce the state of a scientific discipline— plant physiology in the eighteenth century, say—to a sum- mary of chronological and logical connections among var- ious systems of propositions pertaining to various classes of problems or solutions, The quality of historical work would then be measured by breadth of erudition and shrewdness in analyzing the connections between the work of different scientists, by skill in ferreting out si and differences in their views. But the diverse quality of historical works cannot conceal the fundamentally identi- cal relation of the historian to the object whose history is being told. A pure history of eighteenth-century botany would consider “botanical” nothing but what botanists of the period took to be within their scope of inquiry. Pure historians are interested only in what scientists thought they were doing and how they went about doing it. But a fundamental question must be asked: Does this science of the past constitute a past for the science of today? “Taken in an absolute sense, the “past of a science” is a vulgar concept. The “past” is a catchall of retrospective inquiry, Whether the question is the shape of the earth, the “hominization” of man, the social division of labor, or the alcoholic delirium of a particular individual, one turns to the “past” as required by present needs in search of more ‘or less remote antecedents to some present state of affairs. The past, moreover, is conceived beforehand as a vessel of larities ‘THE ROLE OF EPISTEMOLOGYinfinite capacity. Consider again the example of plant phys- iology. In the broad sense just defined, its past woul clude everything that people called botanists, phy: chemists, horticulturists, agronomists, or econo1 might have written in regard to conjectures, observations, or experiments with a bearing on the relation between structure and function in objects variously termed herbs, plants, or vegetables. An idea of the abundance of such source material, even allowing for selection based on chronological and political criteria, can be had by consult- ing the very useful catalogue of the works of French bota- nists who belonged to the Academy of Sciences compiled by Lucien Plantefol to commemorate the group’s three hundredth anniversary.’ But a catalogue of prior works is, at the time it is compiled, a history of botany in the sense that botany is itself a history, by which I mean an ordered description, of plants. The history of a science is thus a summary of readings in a specialized library, a repository and conservatory of knowledge produced and expounded from the time of the tablet and papyrus, parchment and incunabula to that of the magnetic tape. This is, to be sure, an ideal library, a library of the mind, which by definition contains a record of everything ever said about the subject. The totality of the past is represented here as an unbroken in this expanse it is easy to locate places from which itis possible to trace a line of progress to what- ever one’s current object of interest happens to be. Some historians are quite bold in locating these antecedents, while others are more cautious. But it is simply boldness or prudence that distinguishes their work. One can argue, however, that the history of science is entitled to expect from epistemology a set of ethical criteria, by which I mean a set of criteria for judging which moves within the vast expanse of the past are legitimate and which are not. After much rigorous argument this is the conclusion reached by Suzanne Bachelard in her Epistemology and History of Sci- INTRODUCTION ence. In her words: “The fact that the historian’s work is retrospective establishes powers. The historian constructs his objects in an ideal space-time, It is up to him to make sure that this space- time is not imaginary.” To return to my example, the eighteenth-century bot- anists who undertook to do research in plant physiology looked to contemporary animal physiology for models. Some were physicist-physiologists like Stephen Hales, while others were chemist-physiologists like Jean Senebier and Jan Ingenhousz. Yet simply because contemporary plant physiology uses analytical methods from chemistry and experimental techniques from physics, it would be au- dacious to say the least to construct a history in which a of intention was allowed to conceal a radical discontinuity of object, for biochemistry and biophysics have made substantial innovations in the nature of plant physiology. Between the chemistry of oxidation and the biochemistry of enzymatic reductions, plant physiology lar physiology (and cellular theory of course met with tremendous opposition) and then had to tid itself of its early concepts of the in order to study metabolism at the molecular level remarkable History of Biochemistry Marcel rowing Gaston Bachelard’s concept of an “epistemological break,” shows how an enzymatic theory replaced a proto- lasmic theory as a result of Eduard Biichner's discovery in 1897) of noncellular fermentation, which for a long, Pasteurian biology. It should by now be clear why the past of a present- day science is not the same thing as that science in the past. In order to understand the sequence of research, experi- mentation, and conceptualization without which it would be impossible to comprehend the work of Gabriel Bertrand (1897) on the necessary presence of metals in the molecules ‘THE ROLE OF EPISTEMOLOGYof enzymes and on the role of what he called “co- enzymes,”’ there is no point in going all the way back to Théodore de Saussure (1765-1845) and his work on plant nutrition, By contrast, there is good reason to look at the work of Saussure’s contemporary Brisseau de Mirbel (1776~1854) and the origins of cellular theory in botany, which can shed light on the heuristic value of the localiza- tion within the cell of objects that figured in the early work ‘on enzymatic biochemistry. In other words, events situated at the same point in historical time may or may not have theoretical significance, What matters is the overall line of scientific discourse. A particular end point may be related to one or more conceptually homogeneous points of de- parture, Each such trajectory has its own characteristic nature. Well, then, the historian of science may object, why should the role claimed by the epistemologist not be filled by the scientist? Who but the scientist has the competence to say, based on his instinct about the direction of future developments, which end points are of scientific interest and therefore worthy of historical reconstruction? Such an appeal to a third party can only surprise or embarrass the epistemologist. He is well aware that there have been, and are, scientists who, as respite from their scientific labors, have turned to writing the history of science, and that there have been, and are, scientists who, with the aid of episte- mological concepts borrowed from philosophers, have written critical histories not without constructive influence on the subsequent course of scientific progress. Ernst Mach’s Die Mechanik in ibrer Entwicklung (The Devel- opment of Mechanics, 1883) is a celebrated case in point, whose influence on the work of Einstein is well known, ‘Marie-Antoinette Tonnelat has given us a historical and epistemological case study in her History of the Principle of Relativity What epistemologist would not subscribe to INTRODUCTION her introductory statement, which dismisses a certain ap- proach to the writing of history: At the risk of disappointing certain specialists I shall argue that there is no authentic and unsurpassable principle of relativity whose earliest development tific theory it is the job of the historian to describe. No imperfect but proximation lurks behind the veil of ig- ince and prejudice awaiting anointment. The very idea . « Born in the confusion of late Aristo- telianism, made over by contradictions inherent in the elu- sive concept of the ether, the idea of relativity in each case appears to have been associated more with what followed than with what preceded? An innovative vision, it ighted its own way and to a large extent even determined the meandering of its path and the plumbing of its depths.!° is one thing to recognize the existence and value of an epistemological history written by scientists." It is an- other, however, to argue that the epi therefore concede that he has no special tory of science on the grounds that a relation can be established between the scientist and the history of sci- ‘ence, to the great benefit of the latter. Or that the episte- mologist must remain an outs relation to history may appear similar to that of the scien- tist, his motivation is fundamentally different. Jean-Toussaint Desanti, having noticed the widening gap between science and philosophy, questions the rele- vance of the questions that philosophers—epistemolo- gists—have posed to scientists concerning the ways and means by which knowledge is produced.” Since philosoph- ical discourse does not produce knowledge, is philosophy disqualified from discussing the conditions of its produc- tion? “Must one resolve to say nothing about the sciences that the task of criticism, which is to counter internalis ‘THE ROLE OF EPISTEMOLOGYand mimetic accounts of the work of science, requires adopting a standpoint within scientific utterance. That this involves practicing science is a part, by no means the least important part, of Gaston Bachelard’s teaching. “One must either say nothing about science or speak from the inside, that is, by practicing it.” But there is practice and there is practice. If the word is used in the sense in which Descartes said that he put his method into practice in solving mathematical problems,'* it may seem that productive practice of this kind is not within the phi- losopher’s reach; if it were, he would be among those in the vanguard of scientific progress. For the epistemologist, practicing a science amounts to mimicking the practice of the scientist by attempting to reconstitute the means by which knowledge is produced through studious atten- tion to the papers in which the producer explains his behavior.'s Since a scientific investigator in his theoretical work must necessarily take an interest in the work of his or her immediate predecessors, and since every immediate prede- cessor also has an immediate predecessor, it must be con- ceded that science has a natural interest in its own history, even if that interest is not very widespread among scien- tists. But the scientist's historical interest is part of the heu- tistics of research; hence it does not extend to very remote antecedents, where “remoteness” is to be construed i conceptual rather than chronological terms. A particular nineteenth-century mathematician may have been more in- terested in, say, Archimedes’ work than in Descartes’, Everyone's time is limited, moreover, and in the mind of the scientist the importance of a theoretical advance natu- rally bulks larger than does historical investigation. Unlike the scientist, the epistemologist is free to in- dulge his historical interest if not full-time then at least a large part of the time. His is an interest of vocation rather than avocation, for his problem is to abstract from the his- INTRODUCTION tory of science in the manifest sense—that is, a more or less systematic series of pronouncements claiming to state the truth—in order to uncover the history of science in the latent sense—that is, the order of conceptual progress that is visible only after the fact and of which the present notion of scientific truth is the provisional point of culmination. Furthermore, because the epistemologist’s historical inter- est is primary rather than secondary, he can range more freely than the scientist. His breadth of knowledge can compensate for the relative inferiority of his mastery of the latest scientfic discoveries and analytic tools. For example, Sir Gavin de Beer, in the course of research that led to the publication of his Notebooks on Transmutation of Species (1960-1967), became interested in rereading Darwin. His historical interest was motivated and illuminated by his work as an embryologist, which enabled him to see Dar- winian and pre-Darwinian concepts of the relation of on- togeny to phylogeny in a new light." But when Camille Limoges, drawing upon previously unpublished work of Darwin collected, published, and commented on by Gavin de Beer, challenged an assertion that had been repeated for nearly a century to the effect that Darwin had derived from his reading of Malthus the conceptual framework he needed to make sense of his observations, he revealed a quite different view of the past.” In effect, Limoges was challenging that mainstay of traditional historiography, the concept of influence. Using Darwin as an exampk sought to illustrate a new way of reading the sources with- ‘out attaching any special privilege to those in which an author believed that he was explaining his own assump- tions. By revealing a sharp contrast between the concept of natural selection and its predecessor, natural economy, Limoges was able to locate the cleavage between the new and the old natural history: the key point of contention was the concept of adaptation, which was now understood in a probabilistic sense and related to observations of a ‘THE ROLE OF EPISTEMOLOGYbiogeographical (or, as we would now say, ecological) order."* ‘The interest of epistemology in the history of science is not new. It is, as I have noted, a matter of vocation. In a sense, epistemology has always been historical, When the theory of knowledge ceased to be grounded in an ontology incapable of accounting for the terms of reference em- ployed in the new cosmological systems, it became neces- sary to examine not the justifications but the methods of science itself. Kant, in the second preface (1787) to the Cri- tique of Pure Reason, used a brief history of the mathe- matical and physical sciences, condensed into just a few lines, to justify his intention to invert the existing relation- ip between the known and the knowing (le connu et le connaitre). Commentators on this preface have tradition- ally stressed the pscudoreversal effected by Copernicus and neglected, | think wrongly, the novel meaning of the terms in which Kant defines the driving force behind what he calls revolutions in the techniques of thinking (Denkart). The mathematician (first exemplified by, say, Thales or some other Greek thinker) first had to produce (hervor- bringen) the objects that figured in his proofs. Similarly, the physicist (modeled on Galileo or Toricelli) first had to pro- duce (hervorbringen) the objects of his experiments, and this production was a result of purely intellectual progress (Vorangehen). The very fact that Kant believed he could abstract from scientific work a definitive set of rules and norms governing the production of knowledge tells us a great deal about the culture of the period. When one thinks of categories of scientific thought. In establishing such a close connection between epis- temology and the history of science I am of course drawing on the inspirational teachings of Gaston Bachelard."” The fundamental concepts of Bachelard’s epistemology are INTRODUCTION 1 by now well known, so well known, perhaps, that they have been disseminated and discussed, especially outside France, in a vulgarized, not to say sanitized, form, devoid of the polemical force of the original. Among them are the notions of new scientific spirit, epistemological obstacle, epistemological break (rupture), and obsolete or “official” science. Italian, Spanish, German, and even English readers have come to know Bachelard’s work not firsthand but through translations of critical commentaries, particularly that of Dominique Lecourt. To my mind, the best summary of Bachelard’s research and teaching can be found in the concluding pages of his last epistemological work, Le Ma- térialisme rationnel" Here the notion of epistemological discontinuity in scientific progress is supported by argu- ments based on the history and teaching of science in the twentieth century. Bachelard concludes with this state- ment: “Contemporary science is based on the search for true (véritable) facts and the synthesis of truthful (véri- dique) laws.” By truthful Bachelard does not mean that scientific laws simply tell a truth permanently inscribed in objects or intellect. ‘Truth is simply what science speaks. How, then, do we recognize that a statement is scientific? By the fact that scientific truth never springs fully blown from the head of its creator. A science is a discourse gov- emed by critical correction. If this discourse has a history whose course the historian believes he can reconstruct, itis because it is a history whose meaning the epistemologist must reactivate, “Every historian of science is necessarily a historiographer of truth. The events of science are linked together in a steadily growing truth. ... At various mo- ments in the history of thought the past of thought and experience can be seen in a new light.”2! Guided by this new light, the historian should not make the error of think- ing that persistent use of a particular term indicates an in- variant underlying concept or that persistent allusion to similar experimental observations connotes affinities of ‘THE ROLE OF EPISTEMOLOGYaa method or approach. By observing these rules he will avoid the error of, for instance, seeing Maupertuis as a prema- ture transformist or geneticist? There is a clear difference between retrospective criti cal evaluation of the scientific past in the light of a present state of knowledge (certain, precisely because it is scien- tific, to be surpassed or rectified in the future) and system- atic, quasi-automatic application to the past of some standard model of scientific theory. The latter is more in the nature of an epistemological inquisition than a histor- ical inquiry. The von oben bis unten ot “top-down method,” as Father Joseph T. Clark calls it, requires the historian of science to take the analytic philosopher's word for it that science has now achieved maturity and that in the future new results will be produced according to the same logical model as now: Accordingly, the work of the historian equipped with a finished scientific theory is to ask why past theories failed to achieve logical maturity. But, as Ernest Nagel observed in the discussion, to use a contem- porary model as a universal touchstone results not in pro- jecting a powerful searchlight into the past but in inkering our vision. To imagine how, for example, Cop- ernicus might have overcome certain limitations of his theory by formalizing all his assumptions is to confound logical with historical possibility. In Nagel’s view, Clark's thesis betrays dogmatic confidence in analytic philosophy and its theory of knowledge. ‘Thus it is easy to distinguish between epistemological recursion (récurrence épistémologique) and the top-down method. It is no less easy to distinguish between what Bachelard calls “normality”? and what Thomas Kuhn calls “normal science.” The two epistemologies do share certain points in common: in particular, the observation that scientific textbooks overemphasize the continuity of scientific research. Both stress the discontinuous nature of progress. Nevertheless, while the fundamental concepts INTRODUCTION 3 share a family resemblance, they do not really belong to the same branch. This has been noted by Father Francois Russo, who, despite reservations about the claims of su- periority to which epistemological historians are some- times prone, argues that Kuhn is mistaken about the nature of scientific rationality as such.2? Though ostensibly con- cerned to preserve Karl Popper's emphasis on the necessity of theory and its priority over experiment, Kuhn is unable to shake off the legacy of logical positivism and join the rationalist camp, where his key concepts of “paradigm” and “normal science” would seem to place him. These con- cepts presuppose intentionality and regulation, and as such they imply the possibility of a break with est: and procedures. Kuhn would have them play this role without granting them the means to do so, for he regards them as simple cultural facts. For a him, a paradigm is the result of a choice by its users. Normal science is defined by the practice in a given period of a group of specialists in a research setting. Instead of concepts of philo- sophical critique, we are dealing with mere social psychol- ‘ogy. This accounts for the embarrassment evident in the appendix to the second edition of the Structure of Scientific Revolutions when it comes to answering the question of how the truth of a theory is to be understood. By contrast, when Bachelard speaks of a norm or itis because in thinking of his favorite science, math- I physics, he identifies theory with mathematics. His rationalism is buile on a framework of mathematism. In mathematics one speaks not of the normal but of the normed. In contrast to orthodox logical positivists, Bach lard holds that mathematics has epistemological content, whether actual or potential, and that progress in mathe- matics adds to that content. On this point he agrees with whose critique of logical positivism has lost vigor or rigor. Cavaillés refutes Carnap by ‘THE ROLE OF EPISTEMOLOGY14 showing that “mathematical reasoning is internally coher- ent in a way that cannot be rushed. It is by nature progres- sive.” As to the nature of this progress, he concludes, One of the fundamental problems with the doctrine of sci- ence is precisely that progress is in no way comparable to increasing a given volume by adding a small additional amount to what is already there, the old subsisting with the new. Rather, it is perpetual revision, in which some things are eliminated and others elaborated. What comes after is greater than what went before, not because the present contains or supersedes the past but because the one necessarily emerges from the other and in its content car- ries the mark of its superiority, which is in each case unique.” Nevertheless, the use of epistemological recursion as a historical method is not universally valid. It best fits the disciplines for the study of which it was originally devel- oped: mathematical physics and nuclear chemistry. Of course there is no reason why one cannot study a particu- larly advanced speciality and then abstract rules for the production of knowledge that may with caution be extrap- olated to other disciplines. In this sense the method can be not so much generalized as broadened, Yet it cannot be extended to other areas of the history of science without a good deal of reflection about the specific nature of the area to be studied. Consider, for example, eighteenth-century natural history. Before applying Bachelardian norms and procedures to the study of this subject, one must ask when a conceptual cleavage” occurred whose effects were as rev- olutionary as those of the introduction of relativity and quantum mechanics into physics. Such a cleavage is barely perceptible in the early post-Darwin years," and to the ex- tent that it is visible at all it is only as a result of subsequent cataclysms: the rise of genetics and molecular biology. Hence the recursive method must be used judiciously, and we must learn more about the nature of epistemolog. INTRODUCTION 15 ical breaks. Often the historian in search of a major water- shed is tempted to follow Kant in assuming that science begins with a flash of insight, a work of genius. Frequently the effects of that flash are said to be all-embracing, affect- ing the whole of a scientist’s work. But the reality is differ- cent. Even within one man’s work we often find a series of fundamental or partial insights rather than a single dra- matic break. A theory is woven of many strands, some of which may be quite new while others are borrowed from older fabrics, The Copernican and Galilean revolutions did not sweep away tradition in one fell swoop. Alexandre Koyré has located what he considers to be the decisive “mutation” in Galileo's work, the decisive change in think- ing that made him unable to accept medieval mechanics and astronomy.” For Koyré, the elevation of mathemat- ics—arithmetic and geometry—to the status of key to telligibility in physics indicated a rejection of Aristotle in favor of Plato. Koyré’s argument is sufficiently well known that I shall not discuss it in detail. But in painting a quite accurate picture of Galileo as an Archimedean as much as a Platonist, is not Koyré abusing the freedom of the recur- sive method? And is he not somewhat overstating the case in saying that the change in Galileo's thinking marked a total repudiation of Aristotelianism? Is not Ludovico Geymonat right to point out that Koyré’s interpretation neglects all that Galileo preserved from Aristotelian tradi- tion even as he was proposing that mathematics be used to bolster logic?* Thus Koyré is himself challenged on the very point on which he challenged Pierre Duhem when he wrote, “The apparent continuity in the development of physics from the Middle Ages to the present (a continuity that Caverni and Duhem have so assiduously stressed) is i ‘No matter how well the groundwork has been ion is still a revolution? ** Parenthetically, it is worth asking why in matters of ‘THE ROLE OF EPISTEMOLOGY