Exactitude in Science

[173]

Gaston Bachelard

1. The History of Science in the Epistemological Works of Gaston Bachelard

Georges Canguilhem

An extract from the Annales de l’Université de Paris, 1963, no. 1.

When in November 1940 Gaston Bachelard was called on to succeed the recently deceased Abel Rey, this included the teaching of history and philosophy of science^*^* The word ‘science’ is often rendered in the plural in French. This phrase was originally l’enseignement de l’histoire et de la philosophie des sciences, ‘the teaching of the history and the philosophy of the sciences’. I will use the singular of ‘science’ throughout as it is more natural in the English language. at the Sorbonne’s Faculté des Lettres, and at the same time, the direction of the Institut d’histoire des Sciences et des Techniques (Institute for the History of the Sciences and Technologies) which the Université de Paris had founded on the 28^th of January 1932.

Although in France the history of science is not given the same important place in the course of higher education and learning which it is given in several other foreign countries, the teaching of history of science has a tradition in France which associates it with the philosophy of science. Whatever the judgment brought against this tradition, at the very least it is not disputed that it owes its existence to the fact that the history of science – a literary genre born in the scientific academies of the eighteenth century – was introduced into the customs and institutions of French culture in the nineteenth century through the concerns of a philosophical school which claimed to found its authority and base its credit on the necessity of its own emergence [la nécessité de son propre avènement] in accordance with a law describing the historical development of the human mind. This was of course the school of Positivism. We will limit ourselves, here, to recalling that a chair in the general history of science [chaire d’histoire générale des sciences], which Auguste Comte was not able to obtain from Guizot in 1832, was created sixty years later in his memory at the Collège de France and occupied by [173–174] Pierre Laffitte, president of the Société positiviste; and that the succession of Laffitte was refused to Paul Tannery in favour of another positivist, Wyrouboff. Citing here Abel Rey:

At the time in which the likes of Paul Tannery and Pierre Duhem were alive in France, the chair of the history of science at the Collège de France was entrusted to men whose work, in what concerned the history of science, was non-existent; the same chair was re-established again after an interruption of some years for Pierre Boutroux whose work was brutally cut short by a premature death.¹1. Rey, L’Histoire des science dans l’enseignement (Publications of the Enseignement scientifique, no. 2), 1933, p. 13.

In the meantime, the Sorbonne’s Faculté des Lettres had created a chair in the history of philosophy as it related to the sciences, which was first occupied by Gaston Milhaud, then by Abel Rey. The title of this position then became: the History and Philosophy of Science [Histoire et Philosophie des Sciences].

***

Coming from Dijon, Gaston Bachelard arrived in Paris with a collection of impressive publications: Lautréamont (1939), La formation de l’esprit scientifique (1938), La psychanalyse du feu (1938), La dialectique de la durée (1936), Le nouvel esprit scientifique (1934), Les intuitions atomistiques (1933), L’intuition de l’instant (1932), La valeur inductive de la relativité (1929). And yet, it was without doubt the two doctoral theses of 1927 which had destined Gaston Bachelard to the brilliant illustration of the alliance between the history of science and the philosophy of science.

The principal thesis, Essai sur la connaissance approchée, was an epistemological study in which the author attempted to explain “how concepts of reality and of truth should receive a new meaning from a philosophy of the inexact.” The complementary thesis, Étude sur l’évolution d’un problème de physique : la propagation thermique dans les solides, was a study in the history of science, but in a sense that was truly new. In the first chapter (the formation of scientific concepts in the eighteenth century), Bachelard proposed to show that the historical succession of scientific problems is not ordered according to their growing complexity. The initial phenomenon of research is not a phenomenon that is originally simple. It is “the discovered solution [174–175] that reflects its clarity on the given” and which inclines us to misunderstand the fact that “the problem has for a long time been obscured by grave and tenacious errors.”²2. Bachelard, Étude sur l’évolution d’un problème de physique (1927), p. 7. The history of a physics problem knows two times: (1) the time when research takes hypotheses as its initial object, when it is believed that a phenomenon is explained as soon as experience obliges us to change course by replacing analogies one by the other: “the eighteenth century drew to a close without there having been any true attempt at a mathematical liaison of thermal phenomena.” (2) The time, which begins in the given case with the work of Jean-Baptiste Biot, when a physics problem becomes mathematised, when “calculus is adapted as near as possible to experience and leads imperceptibly to an experimental verification, a verification itself intimately mixed with calculus.”³3. Ibid., p. 31. We note straight away that, from his first work, Bachelard considered mathematical physics as the royal science. Without a doubt, he held Fourier as a pioneer in matters of mathematical thermology, but not without some restrictive nuance: “The instructive power of mathematics to which Fourier gave all his trust must however appeal to physical elements.”⁴4. Ibid., p. 54. But it is both the method and lessons of Gabriel Lamé that Bachelard celebrated above all others:

With Lamé, calculus must do everything. It must furnish the hypothesis, coordinate the domains, construct the phenomenon from scratch. Calculus does not study laws but discovers them. Never has so great a role been assigned to reasoning [raisonnement].⁵5. Ibid., p. 104.

The history of the problem reconstructed by Bachelard ends with a lesson concerning the relation between science and its history, and to a greater extent, indirectly concerning the means of composing the history of science itself:

Scientific development is not a simple historical development; a unique force runs through it and one could say that the order of productive thoughts is a matter of the natural order [matière d’ordre naturel].⁶6. Ibid., p. 159.

Natural, and not simply human. A science has its destiny and not only a chronology. The history of science, philosophically questioned – that is, in regards to the formation, reformation, and formalisation of concepts – gives rise to a philosophy of science. It would be [175–176] easy enough to say that the philosopher (Bachelard) rediscovers here the philosophy which he has brought along himself. But it was not Bachelard who was responsible for the successions that he studied, from Biot to Fourier, Poisson to Lamé. It was not Bachelard who was responsible for the fact that Lamé was led to read Fourier in a different way to how August Comte had read him. The fourth chapter of Bachelard’s study had the title: August Comte and Fourier. Here Bachelard was equitable and generous to Comte, he tried hard to understand the intention of philosophical attitudes which are ordinarily made fun of or criticised. But his conclusion was as little positivist as possible. The evolution of the problem of thermal propagation authorised a conception of physical theory in the work of Bachelard that was entirely non-positivist (positive taken here in the sense of Mach as well as Comte).

Prediction that is based more on doctrine than on facts could very well be accused of recklessness. But one is obliged to admit that prediction which is based on this mathematics succeeds physically, and that it enters into the intimacy of the phenomenon itself. It is not a question here of a generalisation, but, on the contrary, by preceding the facts the idea discovers the detail and gives rise to specifications. It is the idea which sees the particular in all its richness, beyond sensation which understands only the general.”⁷7. Ibid., p. 159.

***

The 1927 thesis illustrates a conception of the history of science. Bachelard’s relation to the philosophy of science, however, still lacked the concept of epistemological obstacle [obstacle épistémologique], the invention of which Bachelard revealed himself as a brilliant innovator in the history of science. No doubt, as we have seen, Bachelard expressed his disagreement concerning certain ways of writing the history of science, those histories with a perspective of progressive complication and which misunderstand the stubbornness of the errors which had for a long time obscured a problem. The root of these errors, the reason for their stubbornness, had not yet been given by Bachelard, although perhaps it had been suspected. But from the first chapter of La formation de l’esprit scientifique we learn that this root must be looked for in knowledge itself, not outside of it. That which the scientific mind [esprit scientifique] has to overcome forms an obstacle in that very mind. It is, to the letter, a [176–177] survival instinct of thought [instinct de conservation de la pensée],⁸8. Bachelard, La formation de l'esprit scientifique, p. 15. a preference given to answers rather than to questions. The existence of epistemological obstacles renders the tasks of the epistemologist different from the historian of science. The epistemologist must retrace the evolution of scientific thought, and for that they have to choose amongst the documents collected by the historian and judge them. “The historian of science must take ideas as facts. The epistemologist must take facts as ideas, inserting them into systems of thought.”⁹9. Ibid., p. 17. But, in turn, an awareness of epistemological obstacles will allow the history of science to be authentically a history of thought. It will keep the historian from the false objectivity that consists in drawing up an inventory of all the texts of a certain period – or even of different periods – in which the same word appears, or in which research projects seem to be expressed in substitutable terms. The same word is not the same concept. It is necessary to reconstitute the synthesis in which a concept finds itself inserted, that is, both the conceptual context and the guiding intention [intention directrice] of experiments and observations.¹⁰10. We are pleased to reproduce a beautiful passage of J.-B. Biot who expressed the same rule of historical critique: “I am not able to leave this memorable period without discussing here an allegation which has had a great impact on the history of chemical science, as much as it seems to me very far from meriting the importance which has been given to it. It is a question of nothing less than taking away from Lavoisier and the modern chemists the fundamental discovery of the ability of metals to combine with one of the elements of atmospheric air, transferring it to the first years of the seventeenth century and giving the honour of its discovery to a French doctor of that time called Jean Rey. When a new and considerable fact, fecund in its consequences, comes to be produced in the scientific world, accompanied by proofs which establish its certainty and by applications which uncover its significance, it is a natural habit of contemporary minds that they curiously seek out whether there are no former traces of it. If they find any, even if it is indecisive, they hold on to it, and revive it so to speak, with a facility of conviction full of indulgence. This line of critique is quite merited, when it is equitable. Because it is always concerned with giving justice to misunderstood inventors. But, by transferring the meaning one attaches to these inventions in one’s own time back to the perspective where these inventors had placed themselves, by ascribing to the expressions they used; by giving to their ideas the entire range that they had been able to have; it is necessary then to apply to these productions the immutable laws of scientific discussion. Therefore, a clear distinction will have to be made between assertions and proofs, between glimpses and established truths; because there is neither utility, equity, nor philosophy in accepting as demonstrated that which in a former author would be refused as a hypothesis by a contemporary. If one asses Jean Rey’s book in accordance with these rules, the account of it is facile.” – J.-B. Biot, Mélanges scientifiques et littéraires, vol. II, 1858, p. 187: in relation to “Recherches chimiques sur la respiration des animaux,” by Regnault and Reiset. So history is indeed history of science, [177–178] the history of an evolution that is valorised by its demands much better than by its brute results.

History, in its principle, is indeed hostile to all normative judgement. And yet, it is necessary to place oneself within a normative point of view if one wants to judge the effectiveness of a thought.¹¹11. La formation, p. 17.

One must here understand the originality of Bachelard’s position in relation to the history of science. In one sense, he never actually wrote any history of science. In another sense, he did so continuously. If the history of science consists in drawing up an inventory of the variations in successive editions of a text, Bachelard was not a historian of science. If the history science consists in making the difficult, frustrated, re-established and rectified edification of knowledge sensible and intelligible at the same time, then the epistemology of Bachelard was a history of science always in actuality [en acte]. Thus the interest which he had for errors, horrors,¹²12. Cf. Ibid., p. 21: “we will display our haphazard museum of horrors.” and disorders, all of which represented the fringe of historical history [histoire historique], a form history with little similarity to historical epistemology [épistémologie historique]. For example, the history of electricity makes its way to Giovanni Aldini (1762–1834), Galvani’s nephew, and to his experiments of electric discharge across different organic substances (milk, urine, wine, beer, etc.) in view of determining the variation of the properties of the electric fluid according to the bodies that it crosses (Essai théorique et expérimental sur le galvanisme, 1804). But, Bachelard remarked, the concept of resistance formed by Ohm in 1826 (cf. Die galvanische Kette mathematisch bearbeitet, 1827), expurgates [épure] Aldani’s quasi sensualist hypothesis through abstraction and mathematisation, forming a sort of knot of concepts [nœud de concepts].¹³13. Ibid., p. 105.

In other words, the historian and the epistemologist have in common (or at least should have in common) the scientific culture of today. But by situating it in different ways within their perspectives, they confer a different historical function to it. The historian proceeds from origins towards the present so that the science of today is to some degree always announced in the past. The epistemologist proceeds from the actual towards its beginnings so that only a part of what was given as a science yesterday is found to be grounded by the present. But at the same time that the science of today creates [fonde] – never of course forever, [178–179] but incessantly renewed – it also destroys, and destroys forever. Of the sensualist and substantialist history of electricity in the eighteenth century “there remains nothing, absolutely nothing, in the scientific culture of today duly watched over by the city of electricians [cité électricienne].”¹⁴14. Rationalisme appliqué, p. 14.

In brief, as long as philosophy has not furnished the key concept of the epistemological obstacle to the history of science, epistemology risks being the victim of a naïve history of science “that almost never reproduces the obscurities of thought”¹⁵15. Ibid., p. 9. and which makes “us take for the incandescent light of knowledge all the wan flickerings of the past.” In this case, the epistemologist would then be inclined to a static psychology of the scientific mind. Like Émile Meyerson, by researching the real and the identical, the epistemologist would characterise a scientific thought in a unitary fashion which would never ceases to discover reality at different levels thanks to techniques of detection and measurement which are always more powerful and more precise.

To believe that the state of mind of a pre-Lavoisien chemist like Pierre-Joseph Macquer is similar to the state of mind of a modern chemist, this is precisely to be restricted to a materialism without dialectic.¹⁶16. Ibid, p. 9.

Despite the fact that the rapprochement might have for some an air of the paradox or be scandalous, it is quite necessary to observe that, just like August Comte, Meyerson believed both in the fixed nature of the methods and procedures of reason, as well as in the unity of scientific thought and common sense. Of course, Comte, the intimate enemy of Meyerson, said ‘phenomenon’ and ‘law’ where his critic said ‘reality’ and ‘cause.’ But both thought that the progress of knowledge was made with an unchanged march that was on a definitive path. Bachelard referred Meyerson and Comte back-to-back when he refused a continuity between the intellectual styles of common sense and scientific rationality.

How can it be proposed to report our sensible intuitions on the things which escape our intuition? … Modern science has entirely freed itself from the prehistory of sensible givens. It thinks with its apparatuses [appareils], not with the organs of sense.¹⁷17. Bachelard, Activité rationaliste de la physique contemporaine, p. 84.

In the summary text of the course on general history of science (26 March 1892), Pierre Laffitte defined, among others, the intellectual advantages of the history of science. He wrote that,

The historical method [179–180] constitutes a veritable mental microscope; because what in the current exposition of science is presented as a rapid succession then appears to us to be separated by long intervals and with all the difficulties that the great minds had to overcome in order to discover and disseminate.

It is obvious here that Laffitte has transposed time into space, and the action of slowing down into magnification. The history of science slows down a development that then appears with all its idleness, frictions, and “difficulties.” But saying difficulty, does not mean obstacle. The mental microscope cannot make out the difference between difficulty and obstacle, between delay and restless wandering. For Bachelard, the history of science is a School. We pass judgements at a school, and we teach there to make them [On y porte des jugements et on y enseigne à en porter]. “The history of science is, at the very least, a fabric of implicit judgements concerning the value of thoughts and scientific discoveries.”¹⁸18. Bachelard, Actualité de l'histoire des sciences, p. 8. A microscope does not judge. A microscope can detect a movement, but it cannot reveal a dialectic.

***

From the thesis of 1927 – however discrete it then was – Gaston Bachelard made extensive use of the term and concept of dialectic. If the term appears for the first time in 1936 in the title of a work, La dialectique de la durée, then the exposition of the concept and its naturalisation in the world of epistemological concepts is in the work Nouvel esprit scientifique. Both in the Essai sur la connaissance approchée and the Nouvel esprit scientifique, the concept of the dialectical conquest of living thought over inert counter-thought [pensée vive sur la contre-pensée inerte] is quite close to the biological concept of mutation and the psychological concept of animation.

If we knew how to fold a culture of psychology into the culture of objectivity, if we knew how to entirely engross ourselves into scientific research with all the forces of life, we would feel the unexpected animation that the creative synthesis of mathematical physics gives to the soul.¹⁹19. Bachelard, Le nouvel esprit scientifique, p. 179.

But it was in the Philosophie du non – which was intended as a philosophy of the new scientific mind [nouvel esprit scientifique] – that the concept of dialectic appeared; not, admittedly, as a philosophical category, but as a norm of Bachelard’s epistemological thought.

We should always be weary [180–181] of a concept that we have not yet been able to dialectify [dialectiser]. The thing that prevents a concept’s dialectic is an overburdening [surcharger] of its contents. This overburdening prevents it from being delicately sensible of all the variations in the conditions from which it takes its rightful place. To such a concept we certainly give too much meaning, since we never think of it formally. Yet if we give such a concept too much meaning we are in danger of a situation in which two different minds will not give it the same meaning.²⁰20. Bachelard, La philosophie du non, p. 134.

We return therefore, as always, to the intimate and internal relationship of epistemology to history. History, much rather than being itself an objective dialectic, illustrates the dialectic of thought. “The philosophy of no [philosophie du non] has nothing to do with an a priori dialectic.”²¹21. Ibid., p. 135. The philosophy of no is not structured by the dialectic of general history. It is it, on the contrary, that which confers to the history of the sciences a dialectical structure:

We grasp at all the occasions to insist, page after page, on the innovative character of modern physics. Often this innovative character would be sufficiently put by the simple juxtaposition of two examples, one being taken from eighteenth- or nineteenth-century physics and the other from twentieth-century physics. In this way we would see in the detail of knowledge [le détail des connaissances], as in the general structure of knowledge [la structure générale du savoir], that modern physical science presents itself with an incontestable newness.²²22. Le nouvel esprit scientifique, pp. 17–18.

***

The final use of the three concepts – dialectic, new scientific spirit, and epistemological obstacle – led Bachelard to develop a precise and definite doctrine on the relationship between epistemology and the history of science that was capable of many applications. It was outlined, at the beginning of 1951, in the first chapter of Activité rationaliste de la physique contemporaine, and at the end of the same year in a conference of the Palais de la Découverte, L’actualité de l’histoire des sciences. It was based on a new concept, that of historical recurrence [récurrence historique]. It applied this concept to the historical development of the dialectic of the wave and the particle. Bachelard noted, firstly, that “the modern forms of mechanics – that is, relativistic mechanics, [181–182] quantum mechanics, and wave mechanics – are sciences with no ancestors [aïeux].”²³23. Bachelard, Activité rationaliste de la physique contemporaine, p. 23. There is therefore a “historical rupture in the evolution of the modern sciences,”²⁴ and yet these modern sciences synthesise Newtonian and Fresnelian thought;24. Ibid., p. 23. wave mechanics must be taken as a historical synthesis. According to Bachelard, this synthesis is an epistemological act [acte épistémologique]. “The notion of epistemological acts … corresponds to the lurchings [saccades] of scientific genius which bring the unexpected impulses in the course of scientific development.”²⁵25. Ibid., p. 25. The epistemological act divides the course of history by giving rise to an opposition between a positive and a negative. The positive is recognised in what continues to be successful in modern thought, that which constitutes a timely past [passé actuel].²⁶26. Ibid., p. 25.

It is necessary to continuously form and reform the dialectic of expired history [histoire périmée] and sanctioned history [histoire sanctionnée] through a science that is presently active [actuellement active].²⁷27. Ibid., p. 25.

It is this reference to presently active science that forbids the conception of recurrent history being confused with either a historical relativism in the sciences or an aesthetics of the facets of history. The “learned scepticism” of Pierre Duhem claims not to decide between two theories like those of the particle or wave theory of light. It admits the equivalence of hypotheses and does not believe in the existence of criteria of discrimination.²⁸28. Ibid., p. 47. Gœthe (it is we who quote from him and not Bachelard) thought that

From time to time it is necessary to rewrite history, not because we discover new facts, but because we apprehend different parts of it, because progress leads us to points of view which allow us to perceive and judge the past from new perspectives.

But how, in science, can we dissociate progress and the discovery of new facts, how can we oppose facts and different points of view? Indeed, by obstinately opposing himself to Newtonian optics, Gœthe showed that he would have made a bad historian of science, incapable of distinguishing the expired history from the sanctioned. Bachelard took the example of the theory of phlogiston: its history is an expired history. On the other hand, the theory of caloric inspired the works of Black who “made the determination [182–183] of specific heat emerge in positive experiments.”²⁹29. Activité rationaliste., p. 26. Since the notion of specific heat has been a scientific notion ever since,^* Black’s works enter as elements in a history of sanctioned physics.* I have modified the translation of this sentence slightly: “Puisque la notion de chaleur spécifique est une notion scientifique pour toujours …” Here, therefore, recurrent history, judged history, and valorised history, are defended and illustrated. “The history of science will then appear as the most irreversible of all the histories … The history of science is the history of the dismantling [défaites] of irrationalism.”³⁰30. Ibid., p. 27. Bachelard was well aware here that he risked offending the conscience of historians of science more attentive perhaps to the usual deontology of the historian (do not judge!) than to the specificity of the object they applied themselves to. This was why he insisted on the fact that “the history of science cannot be an empirical history,”³¹31. Ibid., p. 13. and that rational values had to order the history of science since they polarised scientific activity itself:

Historians of science are often hostile to this determination of values; but without confessing to it, they themselves deal with a human valorisation peculiar to scientific work. Indeed, they do not fail to describe to us the struggles of genius. These struggles of genius are often analysed within the simple dialectic of social misfortunes and intellectual serendipity … The man of genius runs aground socially and succeeds intellectually – the future making him right. He has posterity to himself. The value of a man of genius becomes the prerogative of the city of science. The retelling of this valorisation is found on all the pages of the history of science.³²32. Ibid., pp. 27–28.

Thus, whoever proposes to compose a complete recurrent history of optical science should leave “the physics of Descartes in its historical solitude,”³³33. Ibid., p. 35. whereas it should consider that Huygens’s construction of the refracted ray on the basis of a wave hypothesis “is a definitive acquisition for science.”³⁴34. Ibid., p. 36. In regards to Newton, the explanation of the phenomenon of optical rings by the ‘theory of access’ [théorie des accès] suffices to show that his optics “is in sum, corpuscular in its simple workings but pre-wave-like [préondulatoire] in its more involved theory” and that, even when he gave his preference to the corpuscular theory, “his doctrines [183–184] of light are of a real dialectical sensibility.”³⁵35. Ibid., pp. 38–39. It is of little importance then that Euler had believed he was able to refute Newton, even if he was able to do so only on the basis of a phenomenological analogy between light and sound. If Fresnel inaugurated the first optical physics on an indestructible base (“finally Fresnel came!”), it was insofar as his calculus gave rise to further applications, constructions, and phenomena without precedent nor example in shared common experience: interference patterns. “We are here before a living scientific past [passé scientifique vivant], which is always timely … The works of Fresnel are, in this regard, models of active science.”³⁶36. Ibid., pp. 45–46.

***

We can see why and how the philosophy of the new scientific mind finds one of its first applications in a new art of writing the history of the science. This history cannot be a collection of biographies, nor a tableau of doctrines, as it would be if it were a natural history. It must be a history of conceptual filiations [filiations conceptuelles]. This filiation has a status of discontinuity, just like Mandelian heredity. The history of science must be as demanding and as critical as the sciences themselves. The desire to obtain filiations without rupture is to mix together all values, dreams and programmes, all presentiments and anticipations; one would find the precursors of everything everywhere. The desire to found modern science not on the coherence of axioms without premises or the cohesion of techniques without antecedents, but rather on the depth of its taking-root in the past of human intelligence, is to recreate, well after Dutens wrote it, the Recherches sur l’origine des découvertes attribuées aux modernes (1766).

But, as Bachelard said, “it is useless to put a false problem at the origin of a true problem, absurd even to bring alchemy and nuclear physics together.”³⁷37. Bachelard, Matérialisme rationnel, p. 104. Contemporary scientists have not realised the dreams of alchemists. “Art and literature realise dreams, science does not.”³⁸38. Ibid., p. 103. Since scientific thought incessantly reforms its past, since it is essential for it to be a continuous revolution, Bachelard was able to [184–185] maintain that “Under these conditions, when it is a question of fresh dialectics, science gains nothing by proposing false continuities to it.”³⁹39. Ibid., p. 103.

In brief, the historian of science must not become a victim of the confusion between the continuity of historical discourse and the continuity of history.⁴⁰40. Ibid., p. 209. In fact, the longer the historian stays in the realm of origins, in the zone of rudiments, the more they are led to confuse the slowness of initial progress [premiers progrès] with the continuity of progress.

In sum, here is the axiom of epistemology posed by continuists: since beginnings are slow, progress is continuous. The philosopher will not go further than this. They believe it is pointless to live in the time of newness, precisely the period of time when progress erupts from every part, precisely the period of time that makes the eruption of traditional epistemology necessary.⁴¹41. Ibid., p. 210.

It seems that we have touched on the genesis of Bachelard’s thought. He was the first French epistemologist who thought, wrote, and published, in the twentieth century, at the chronological and conceptual height of the sciences in which he worked. And this was apparent from La valeur inductive de la Relativité, in 1929:

One of the most evident, outward characters of relativistic doctrines is their novelty. Such doctrines astonish the philosopher who, faced with a construction so extraordinary, suddenly becomes the champion of common sense and simplicity. This novelty is therefore an objection, it is a problem for the philosopher.

A homage to Bachelard does not have to mention the names of the philosophers who believed they had to make themselves the champions of common sense and simplicity, any more than name the philosophers who believed they had to make themselves champions of fashion, taking what the approval of physicists still enclosed with prudence a step too far. In regards to relativistic physics, from 1929 Bachelard had said that it was “a doctrine whose historical antecedents explain nothing,” and that it “has a relation to history only through the rhythm of a dialectic.”⁴²42. Bachelard, Valeur inductive de la relativité, p. 6. Bachelard had first the awareness of epistemological ruptures. It was then that he elaborated the philosophical concepts that were needed in order to give an account of these ruptures. This elaboration led him to propose a conception of the relation between science and the history of science which also constituted a [185–186] rupture: a non-positivist conception of the relation between science and the history of science. Positivism was founded on a law of three states, a law of progress which, according to A. Comte, was a law of continuous development, the end of which would already be there at the beginning. Gaston Bachelard’s philosophy was founded on a norm of rectification which was expressed by three laws of three states (cf. the introductory discussion of La formation de l’esprit scientifique). But for Bachelard, the third state did not close itself off from the first; and he did not ignore the fact that in science one never actually goes back on a negation when this negation has been manifested by a deformation of primordial concepts and maintained by a new mode of calculation.

A young epistemologist, Michel Serres, perfectly characterised the decisive role that Bachelard’s epistemology had on the history of science:

A science that has come into its maturity is a science that has thoroughly consummated [consommé] the cut between its archaic state and its present state [son état archaïque et son état actuel]. Thus the history of science can then be reduced – in regards to its genetic explanation – to the exploration of the interval which separates the modern sciences from their precise point of rupture and recurrence. This point of rupture can easily be assigned, however, at the exact moment when the language used in this interval makes any attempts to go further back incomprehensible. Beyond such a point, it is a question of archaeology.⁴³43. Serres, “Géométrie de la Folie” (in relation to the Histoire de la Folie, by Foucault), in Mercure de France, sept. 1962, p. 80, footnote.

By renewing so profoundly the meaning of the history of science, by uprooting the state of affairs which had until then remained berried, by promoting it to the rank of a premier philosophical discipline, Gaston Bachelard did more than clear a path, he set a task. A homage to his memory that is worthy of him, should not only consist in making sensible the void that follows upon his departure, it should consist above all in being able to give the assurance that the lesson of this man of genius will never be lost.

[186]