[In the following essay, McAllester evaluates Bachelard's legacy as critic and philosopher.]

Bachelard died in October 1962, leaving us a rich and singular legacy: some ninety publications in all, and twenty-three books—twelve on the philosophy of modern science, two on time and consciousness, nine on poetic imagination—published between 1928 and 1961, and a tenth book on poetry left unfinished when he died. What has become of this legacy in the last twenty years? Many have expressed their debt to Bachelard's books on poetic images, and he is generally held to have inspired “la nouvelle critique” of 1965 and after. His epistemology has been equally seminal; Georges Canguilhem, for instance, in Idéologie et rationalité, refers to “la leçon de Gaston Bachelard” which has inspired and fortified his “jeunes collègues”—this is in 1977. Two years later, Vincent Descombes in Le Même et l'autre: quarante-cinq ans de philosophie française (1933-1978) mentions Bachelard only three times. Descombes admits that his definition of “contemporary French philosophy” is limited to “celle dont on aura le plus parlé” (p. 13). Bachelard is left out because apparently in 1978 he is no longer talked about. Yet Canguilhem was talking about him in 1977. And Bachelard was and still is widely read in France, an established part of the philosophy syllabus at school and university; the path to the agrégation de philosophie leads through Bachelard. Can it be that France's “apprentis philosophes” have lapsed into uncharacteristic silence? On the contrary, in the mid-seventies Bachelard became a controversial figure for “la jeune génération”. The Marxists Dominique Lecourt and Michel Vadée disagreed over him in books published in 1974 and 1975, and in 1977, Jean-Pierre Roy, wearing the colours of Althusser, Derrida, and semiotics, berated him for his “unscientific” approach to literature, for betraying “la scientificité moderne”. Three passionate, noisy books; Descombes's silence is all the more puzzling.¹

Is there some skeleton in the Bachelardian cupboard that is best kept hidden? Let us be bold and open the door, and have a good look for ourselves. We find, not a skeleton, but vigorous and very much alive, man, man who through the power of his reason and his imagination creates “une nouvelle nature”. In science and in poetry, Bachelard believes, “le monde est conditionné par la provocation de l'homme” (A.R.P.C. [L'Activité rationaliste de la physique contemporaine] p. 141)². This is the sort of statement that now runs Bachelard into trouble. He is a humanist, and in France humanism has become inadmissible. Structural linguistics in particular has challenged the long-held view of man the creator, constituting and controlling his language, and through language, his experience. The last twenty years in France have seen an intellectual revolution, the dethronement of the sovereign subject. Is Bachelard's humanism wrong-headed and outdated? If humanism necessarily implies idealism, if it argues that man is all-powerful, autonomous, the origin of all experience, then yes, his humanism is an embarrassing legacy. But does humanism have to rest on idealism? I wish to argue that Bachelard's humanism does not, that is rests on a conception of man decentred, transcended by something beyond his control, and yet, paradoxically, not denied by this transcendent other, not destroyed by it, but rather nourished and sustained by it. Bachelard's is a subversive humanism. In both his work on science and on poetry, he undermines traditional views of the subject every bit as effectively as recent thinkers in France; his adversaries and theirs are the same, yet the conclusion he draws is very different. He reinvents man, against idealism, beyond conventional notions of subject and object. Why did Bachelard refuse what we may term traditional humanism? What led him to subvert rather than simply reject this view of man?

The matrix of Bachelard's thought is twentieth-century science, “le nouvel esprit scientifique” which he dates from 1905, from Einstein's special theory of relativity. His twelve books on modern science examine its impact on philosophy, and show how science has undermined our familiar epistemologies, so that neither rationalism nor realism, idealism nor materialism will serve as philosophies adequate to twentieth-century science. 1905 saw the break not just with all previous science, but with all previous philosophy. Bachelard's notion of the “rupture épistémologique” is probably what is best known and most widely quoted from his work, and yet those who borrow it—Foucault, Derrida, Althusser—fail to see that this epistemological break brings humanism in its train, a humanism which, in its turn, breaks with traditional humanism. This is, I realize, a controversial assertion on my part, yet it seems to me an obvious conclusion to draw and it is therefore fundamental to my argument. My grounds for it are as follows. For Bachelard, what is new about twentieth-century science is that “le progrès scientifique manifeste toujours une rupture, de perpétuelles ruptures, entre connaissance commune et connaissance scientifique” (M.R. [Le Matérialisme rationnel], p. 207). It is not simply a question of scientific knowledge no longer having first and foremost an empirical basis, but far more important, the rational basis of modern science has broken with reason as we practise it not just in “la connaissance commune” but in science, when scientific reason is no different from everyday reason, when reason means deduction. Modern science is characterized by the primacy of reason, but this is not Cartesian or Kantian reason, it has broken with “le bon sens” and, as we shall see, with deductive logic. This brings us to Bachelard's humanism. Both Descartes and Kant understood the rational subject in terms of this notion of reason as a priori and deductive; if, as Bachelard argues, reason in science is not like this, then consequently the rational subject is not as Descartes and Kant conceived him, he is not the unchanging centre of all knowledge and experience. Science is the matrix of Bachelard's thought, and from it springs a new conception of man, of consciousness, and in due course of poetry.

What is this new reason at work in modern science? Why does Bachelard argue that in consequence of this “révolution de la raison” the twentieth century has seen what he calls “une révolution psychique”, “une révolution spirituelle” (Eng. rat, [L'Engagement rationaliste], p. 12)? Why in promoting this revolution does he subvert not only Descartes and Kant, but Bergson, Husserl, and Freud? And why explain modern mathematics by a quotation from Mallarmé? Why regard reading Valéry as an equally effective way of experiencing the “révolution psychique” that marks our age? Five questions which even before we answer them show something of Bachelard's humanism.

“Nous sommes bien, avec les êtres de la physique contemporaine, devant une nouvelle nature, devant une nature instrumentée et pensée humainement dans une histoire de l'homme” (A.R.P.C., p. 141). This is surely a dramatic humanism, when man creates not just science but nature itself? “On peut, sans hésitation, parler d'une création des phénomènes par l'homme” (F.E.S. [La Formation de l'esprit scientifique], p. 249). But how does man create these new phenomena? Through mathematics: “c'est l'effort mathématique qui forme l'axe de la découverte, c'est l'expression mathématique qui, seule, permet de penser le phénomène” (N.E.S. [Le Nouvel esprit scientifique], p. 58). And more than this, through an entirely new, non-deductive, non-Euclidean mathematics. It was, of course, non-Euclidean geometry, and Riemann's tensor calculus in particular, that made relativity theory possible, and so inaugurated “le nouvel esprit scientifique”. What leads Bachelard to speak of “la révolution non-euclidienne” (N.E.S., p. 28), “la révolution riemannienne” (E.C.A. [Essai sur la connaissance approchée], p. 28)? There are, I suggest, two characteristics that signal this “revolution in reason”, and that consequently shake the foundations of the rational subject.

Firstly, it is the end of reason as a closed, a priori, deductive system. Einstein's theories, for example, were not deduced from Newton's: “l'astronomie relativiste ne sort en aucune façon de l'astronomie newtonienne” (N.E.S., p. 45); they are entirely new. It is the end, therefore, of reason as a closed system of necessity. Bachelard refers to “la contingence radicale des mathématiques” (E.C.A., p. 179; N.E.S., p. 34), to “le choix non-euclidien” (N.E.S., p. 28): we are free to choose our mathematics, to choose whether or not parallel lines will meet. Secondly, it is the end of reason governed by what Bachelard calls “cet idéal d'identification” (Eng. rat., p. 12). “Le sens profond de la révolution riemannienne” is, in his view, the use of differential equations to define mathematical functions (E.C.A., p. 28, p. 282), and more generally, Bachelard notes the increasing importance of the differential calculus in mathematics, its “puissance de diversification” (Eng. rat., pp. 110-11). Put very simply, modern mathematics offers ways of coping with difference, of holding together complex variables like space and time, of maintaining and extending, rather than reducing, complexity. Non-Euclidean geometry and indeed all modern science are characterized by what Bachelard calls “un rationalisme de différenciation” (ibid. pp. 130-1). All this is indeed revolutionary, but why should it matter to us? It is important because it forces us to revise our conception firstly of the rational subject, and secondly of the relationship between reason and reality, subject and object.

The rational subject first. The new mathematics was constructed by Lobatchewsky, Bolyai, Riemann: it is men who make difference. And even more subversively, Bachelard insists that men are made different by difference, that the mathematician is changed by his mathematics: “Si l'on veut bien se placer systématiquement au point de vue psychologique, on ne peut manquer aussi de voir les réactions de l'outil mathématique sur l'artisan” (N.E.S., p. 59). Moreover, it is not only the mathematician who is psychologically changed; all of us, if we read about modern science, will derive what Bachelard calls a “profit psychologique” (ibid. p. 88), our minds will grow more agile, more “disponible”, dynamic, creative (ibid. pp. 55-8, p. 65, pp. 182-3), and this is what he terms the “révolution psychique” that accompanies the “révolution de la raison”. And this is an important aspect of Bachelard's books on science, this desire not just to instruct us but to change us: if, for example, reading Bachelard, we try to grasp the dialectics of matter and energy in microphysics, the dualism of waves and particles, we shall learn to maintain difference, to handle complexity, we are shaken out of the reductive, identity-ridden habits of ordinary life and thought. Plainly, Bachelard does not see the rational subject in Cartesian terms, and in the last chapter of Le Nouvel esprit scientifique, he sets out a “non-Cartesian epistemology”, an epistemology that is not against but rather beyond Descartes. He keeps the notion of the cogito, but he refuses its permanence (p. 172); we can no longer say “je pense donc je suis”, but rather “je pense la différence, donc je deviens différent, et étant différent, je pense d'autres différences”. Bachelard subverts Cartesian rationalism, and consequently he subverts Kantian idealism. If man's mind is changed by scientific knowledge of the world, then surely we can no longer argue with Kant that the laws of the world conform to the laws of man's mind? Bachelard, like Kant, argues that scientific knowledge is obtained independently of experience, through reason; but whereas reason for Kant is a priori, deductive, Euclidean, reason for Bachelard is non-Euclidean, dialectical, and open. Bachelard proposes what he calls “un non-kantisme”, “un kantisme ouvert”, idealism beyond Kant, an “idéalisme discursif” which he defines as “cette reconstruction claire du moi devant le non-moi … une suite de constructions essentiellement différentes” (Et. [Etudes], p. 92). The rational subject is no longer sovereign, no longer autonomous, identical and unchanging, but rather transcended, upheld, created and recreated by something other than itself, by the “non-moi”, by the discursive, dialectical, dynamic interrelationship between reason and reality.

All this raises the question of the origin of difference. Is difference a property of the mind, the product of intricate mathematical pattern-making? Or is it a property of matter, of reality? Mathematics is the instrument of modern science, and more than this, mathematics is a language. This is a constant theme in Bachelard's books on science, and given that it is a conception of language that has led recent thinkers to decentre the subject, it is surely of considerable interest to find Bachelard discussing the language of mathematics in terms of difference, and this ten pages into his first work, Essai sur la connaissance approchée, his doctoral thesis of 1927. An equation, he suggests, is a structure of difference, the equals sign in fact establishes difference between the known and unknown, and he goes on to quote an article published in 1914 by J.-H. Rosny: “Chaque mot exprime d'abord une différenciation, sinon il se confondrait avec tout autre mot”, adding his own interpretation, namely that “plus une pensée est complexe, plus nette est la différenciation de ses termes”³. The language of mathematics, like all language, is a structure of difference, its symbols and operators are arbitrary and autonomous, and the mathematician is limited by this language, whose rules are not of his choosing. However, Bachelard does not argue that the subject is decentred by language, that difference belongs only to language. The context of his thought is “cette floraison de langues nouvelles” (N.E.S., p. 11), the creation by men of new mathematical languages, the “révolution de la raison”, or, as he also puts it, “l'effort poétique des mathématiciens” (N.E.S., p. 35). The mathematician, like the poet, is in language, as we all are. But the mathematician and the poet, unlike most of us, open and renew language. As indeed Bachelard so clearly demonstrates, “le langage de la science est en état de révolution sémantique permanente” (M.R., p. 215). There is an interaction between language and user.

Bachelard is convinced that this interaction occurs because the user is saying something about reality. Mathematics creates and realizes the phenomena of modern science, but Bachelard argues that mathematics is a “stratagème” designed to capture reality (Eng. rat., p. 39), that “c'est sur une variable franchement phénoménale qu'on raisonnera désormais” (ibid., pp. 110-11). Mathematics creates difference in response to a rich reality (ibid., p. 116, pp. 118-19). Mathematics is nourished by “la matière expérimentale” (ibid., p. 115; N.E.S., p. 59). Writing of crystals, for example, Bachelard declares that “on peut trouver dans le phénomène physique des raisons d'enrichir et de continuer la pensée mathématique” (Eng. rat., p. 118). Mathematics and experiment, reason and reality, subject and object are therefore reciprocal and interdependent. Mathematics speaks the language of an endlessly rich, elusive reality, or better, of possibility: “En se souvenant de ces beaux symboles mathématiques où s'allient le possible et le réel ne peut-on évoquer les images mallarméennes?” (N.E.S., p. 60). And Bachelard quotes from Divagations; “il ne faut jamais négliger, en idée, aucune des possibilités qui volent autour d'une figure, elles appartiennent à l'original, même contre toute vraisemblance”, and adds that “de la même manière, les pures possibilités mathématiques appartiennent au phénomène réel, même contre les premières instructions d'une expérience immédiate”. What the mathematician thinks possible can always be realized by the physicist: “le possible est homogène à l'Être” (loc. cit.) This unexpected reference to Mallarmé helps us to understand that through language, through new language that has broken with ordinary language, the mathematician, like the poet, explores possibility.

Mathematician and poet each sustain possibility, and moreover, each is sustained by it. Bachelard develops as a consequence of his epistemology a conception of consciousness which is non-Bergsonian and non-Husserlian. Husserl's conception of intentionality is too static, and he seeks to adapt it, arguing in his second series of books on science (1949-53) that, given the interdependence of reason and reality in modern science, “la conscience rationaliste” is dynamic: I am conscious of something other than myself, and this “other” changes me. This surely confirms both Bachelard's humanism and his refusal of idealism. And the same is true of his long polemic with Bergson. Bachelard wrote two books on time and consciousness against Bergson, L'Intuition de l'Instant (1932) and La Dialectique de la durée (1936), ending the latter with what one might call his consciousness of a poem, with a description of how he reads Valéry: “A voix ‘muette’, laissant les images succéder aux images, vivant dans la superposition des diverses interprétations … la réalité s'habillait, s'étoffait en conditionnels. A l'association des idées venait se substituer la dissociation toujours possible des interprétations” (p. 150). Poetry, like mathematics, is a structure of difference. Reading a poem is consciousness of difference, differences we create as we read, difference created in us by our reading. Consciousness for Bachelard is consciousness of the imbrication of subject and object. He therefore refuses Bergson's distinction between “moi supérieur” and “moi profond”, he refuses the idea of duration as continuity. Instead, consciousness is of discontinuous, new instants of thought which are provoked by the world; it is consciousness of difference. Without the world, there is no consciousness, no human being, no human becoming. Bachelard adapts Bergson's “élan vital” to make of it what we may call an “élan intellectuel”. Science requires us to think in new ways, it has turned us into creatures capable of profound intellectual change: “Par les révolutions spirituelles que nécessite l'invention scientifique, l'homme devient une espèce mutante …”; and having acquired such characteristics, we have become “une espèce qui a besoin de muter, qui souffre de ne pas changer. Spirituellement, l'homme a des besoins de besoins” (F.E.S., p. 16). For Bachelard, difference is an ontological necessity, and since ordinary life is under the rule of identity, the only way we can experience and sustain this difference in ourselves is by thinking about science or by reading a poem: “Raison silencieuse et déclamation muette … premiers facteurs du devenir humain” (A.S. [L'Air et les songes], p. 278).

Bachelard places a high value on poetry well before the publication of his books on poetic images, and this value is closely bound to his conception of the relationship between man and the world that grows out of his work on science. As we turn from science to poetry, we must remember that these books were written when Bachelard held the Chair in the History and Philosophy of Science at the Sorbonne, that his work on science and on poetry is interwoven. Science and poetry are, of course, distinct and separate activities, yet for Bachelard, there is a common factor: equation and image both break with everyday experience, both break with homo faber, both make us, as he puts it, homo aleator, the explorers of possibility in the world and in ourselves (N.E.S., p. 119). Bachelard's work does not present a diptych of “scientificité” and “humanisme”, “matérialisme” and “idéologie”, “la science” and “la fable”, as is now commonly said. Curiously, many who admire his work on science have imposed on it their own materialist preoccupations, ignoring its real lesson, which is, quite explicitly, that we must rid ourselves of all preconceptions, rationalist or realist, which blind us to the facts, to science as it is now, beyond the simple divisions of reason and reality. Yet while Bachelard's books on science have a pedagogical aim—he indeed describes himself as “plus professeur que philosophe” (R.A. [Le Rationalisme appliqué], p. 12)—he refuses the role of teacher in his books on poetry, he has hard things to say about teachers of literature. “Quel mauvais professeur de littérature nous eussions fait”, he declares in La Flamme d'une chandelle (1961) (p. 105), congratulating himself! And along with this, he rejects the role of literary critic, for literary criticism turns literature into “une sempiternelle classe de rhétorique” (D.R. [Le Droit de rêver], p. 177). This comes as something of a shock, given Bachelard's reputation in modern French criticism. Paradoxically, his dislike of teachers and critics of literature stems from the same wariness with regard to theory that made him a teacher of science, from the same desire to attend to the facts. Our shock at Bachelard's anti-critical stance is, in a way, a measure of his success, for it is largely thanks to him that criticism is now understood differently, not as dogmatic, but as attending to the facts, to the text itself. Hence his lasting importance whenever literary theory threatens to overwhelm our own experience of reading a text.

Bachelard likes to describe himself very simply as a reader, not out of intellectual laziness nor false modesty, but because of what happens when he reads: “l'imagination du lecteur … ne se révèle-t-elle pas pure mobilité d'images?” (A.S., p. 288), and more strikingly, “il semble que le lecteur soit appelé à continuer les images de l'écrivain, il se sent en état d'imagination ouverte” (T.R. [La Terre et les rêveries du repos], p. 92). Reading poetic images brings us “l'expérience même de l'ouverture, l'expérience même de la nouveauté” (A.S., p. 7), new images, new language, new possibilities in the world and in ourselves. We recognize a familiar pattern of thought. Is Bachelard simply applying to poetry his own preconceived theories? No, what he brings to it is an attitude of mind, a willingness to accept and not reduce complexity, to take reading a poem seriously, as an aspect of our relationship with something other than ourselves.

What does Bachelard read? Images, not ideas. Images of fire, water, air, and earth in his first books, and later images of space—cellars and attics, shells, corners, the cosmos—and then in his last book, images of a candle-flame. Material and dynamic images which are not perceptual, not rational, which do not express lived experience; images which are written, which are in and through language. How does he read? To begin with, he was attracted by the psychoanalytical approach, because in 1938, when he wrote La Psychanalyse du feu, it was the only one sympathetic to the irrational material images he found in poetry. Yet Bachelard is always critical of psychoanalysis, he modifies and subverts Freud, and eventually, in his second series of books on poetry (1957-61), he rejects it, preferring phenomenology. Why? Above all, because psychoanalysis is reductive; it reduces images to the unconscious, the unconscious to lived experience, to infantile social experience in particular; it reduces the diversity of poetic images to a stricly limited number of meanings, and it pays no attention to the fact that these are literary, written images. Bachelard modifies Freud, for example, by making the source of poetic images not the unconscious, for this would be to divide the subject as Bergson had done, separating him from the object, but rather what he calls an “intermediate zone” on the threshold of consciousness and thought; Bachelard's material images, in which man and matter are conjoined, spring from “la zone des rêveries matérielles qui précèdent la contemplation” (E.R. [L'Eau et les rêves], p. 6). His approach is never the diagnostic approach of the psychoanalytical critic, he is not really interested in the poet but in what the poet does to him: “les images littéraires correctement dynamisées dynamisent le lecteur” (E.R., p. 248). The poem acts on the reader not through some kind of unconscious communication but through language. In 1957, Bachelard turned from psychoanalysis to phenomenology precisely because this offers a better account of reading: La Poétique de l'espace (1957) and La Poétique de la rêverie (1960) are concerned first and foremost with reading, with the reader's consciousness of new language, of what he calls “l'extase même de la nouveauté d'image” (P.E. [La Poétique de l'espace], p. 1). But he modifies Husserl as he did in his work on science, insisting on the dynamic relationship between subject and object, so that the reader's consciousness is changed by what he reads. The newness of the poet's images is linguistic: “le poète, en la nouveauté de ses images, est toujours origine de langage” (P.E., p. 4); it is this new language that changes the reader: “par sa nouveauté, une image poétique met en branle toute l'activité linguistique … Elle devient un être nouveau de notre langage, elle nous exprime en nous faisant ce qu'elle exprime … Ici, l'expression crée de l'être … Ainsi l'image poétique, événement du logos, nous est personnellement novatrice” (P.E., p. 7).

A poem is language: does this make Bachelard a cryptostructuralist? No, he never in fact reads a poem as a whole, as a structure of images, he prefers to remain “au niveau des images détachées” (P.E., p. 9). And this does seem to be a weakness of the Bachelardian approach. “Nous laissons de côté le problème de la composition du poème comme groupement des images multiples” (P.E., p. 8) because, as he explains, this involves complex cultural and historical factors, and their influence on the poet, which he as a philosopher of science and not, one must remember, a professional literary scholar, feels to be beyond his competence. Yet the limits of his approach are also its strengths: Bachelard does not present a poem as a cultural or linguistic phenomenon, but as a personal experience; a poem is not something that confirms a pre-existing body of knowledge, a theory or a hypothesis, it is “un explosif” (A.S., p. 285), a shattering and shaking of our foundations. When we read, we are in language, in language which is not our own: thus far, he is in sympathy with the structuralists; this language—and here he differs—this autonomous “other”, changes, renews, opens our own language.

“Les poèmes sont des réalités humaines” (P.E., p. 190): this, of course, sets Bachelard apart from the present generation of structuralists and post-structuralists. Yet he also insists that poems are written language, and long before Derrida and “grammatologie”, he makes this written language an experience not of “clôture” but of “ouverture”. For Bachelard, reading something that has been written is quite different from listening to someone speaking, for the simple reason that the spoken word imposes itself on us, it requires our submission and our presence, whereas in the written word, read and slowly reread, “les pensées et les rêves se répercutent” (A.S., p. 285). The written word plays between the poles of subject and object, it interweaves and holds together ideas and dreams, the world and the poet, the text and the reader. In Bachelard's view, the language of poetry expresses at one and the same time both subject and object, it abolishes the frontiers of the internal and external worlds, making them reciprocal and interdependent. “Les poèmes sont des réalités humaines” because for Bachelard they exemplify our relations with the world, the imbrication of subject and object. We realize that in his work on poetry just as much as in his work on science, Bachelard is a subversive humanist.

“L'être de l'homme est un être défixé. Toute expression le défixe” (P.E., p. 193). This is surely important. Man is unfixed by language, not decentred. It is not a question of language being outside us or inside us. Bachelard discusses these notions of inside and outside in the penultimate chapter of La Poétique de l'espace, again anticipating Derrida in De la grammatologie (1967), (pp. 46-95). Metaphysics, Bachelard declares, is bedevilled with this simple opposition, with this simple geometric intuition; it fails to see the complex human fact, and it is precisely this complex human fact that is the lesson of Bachelard's work. Mathematics and poetry, the two ways in which men use language most effectively, show how inadequate is this division between inside and outside. Language may well impose its rules on us, its differences are not of our choosing, but great men, mathematicians and poets, have shown how to shape it to new possibilities. Man is in language: how hard it is to escape this simple metaphor! Better perhaps, man is language: “toute expression le défixe”. “L'homme est l'être entr'ouvert” (P.E., p. 200), so that inside and outside flow together and are inseparable. But this is perhaps too simple, too static. Bachelard suggests another metaphor: “quelle spirale que l'être de l'homme!” (P.E., p. 193). Here, there is movement and, more important, no centre. It is surely the very image of what I have called Bachelard's subversive humanism.

Notes

1. Georges Canguilhem, Idéologie et rationalité dans l'histoire des sciences de la vie. Nouvelles études d'histoire et de philosophie des sciences (Paris, 1977), p. 9; Vincent Descombes, Le Même et l'autre: quarante-cinq ans de philosophie française (1933-1978) (Paris, 1978), p. 108, p. 111, p. 142; Dominique Lecourt, Bachelard ou le jour et la nuit (Un essai du matérialisme dialectique) (Paris, 1974); Michel Vadée, Gaston Bachelard ou le nouvel idéalisme épistémologique (Paris, 1975); Jean-Pierre Roy, Bachelard ou le concept contre l'image (Montreal, 1977).

2. References to works by Bachelard, given here in the order in which they occur in this article, will take the form of the following abbreviations: A.R.P.C.: L'Activité rationaliste de la physique contemporaine (Paris, 1951; edition used: Paris: 1977); M.R.: Le Matérialisme rationnel (Paris, 1953); Eng. rat.: L'Engagement rationaliste (Paris, 1972); F.E.S.: La Formation de l'esprit scientifique. Contribution à une psychanalyse de la connaissance objective (Paris, 1938); N.E.S.: Le Nouvel esprit scientifique (Paris, 1934); E.C.A.: Essai sur la connaissance approchée (Paris, 1928); Et.: Etudes (Paris, 1970); A.S.: L'Air et les songes. Essai sur l'imagination du mouvement (Paris, 1943); R.A.: Le Rationalisme appliqué (Paris, 1949); D.R.: Le Droit de rêver (Paris, 1970); T.R.: La Terre et les rêveries du repos (Paris, 1948); E.R.: L'Eau et les rêves. Essai sur l'imagination de la matière (Paris, 1942); P.E.: La Poétique de l'espace (Paris, 1957).

3. J.-H. Rosny, “La contingence et le déterminisme”, Revue du Mois, janvier 1914, p. 22; quoted by Bachelard, E.C.A., p. 21.