SOURCE: Culler, Jonathan. “Bachelard's Images.” In Framing the Sign: Criticism and Its Institutions, pp. 96-106. Norman, Okla.: University of Oklahoma Press, 1988.

[In the following essay, Culler judges Bachelard's overall contribution to science and literary criticism, contending that “the diversity of his accomplishments makes Bachelard difficult to assess and contributes to the neglect of his ideas.”]

At the time of his death in 1962, Gaston Bachelard was France's leading historian and philosopher of science as well as one of its most original and influential literary critics; but to pupils and colleagues he was a moral and philosophic example, a patriarch, majestically bearded, who had mastered the art of happiness and practiced it among books and friends. Rejecting Existentialism and its cognates with the simple credo, ‘L'angoisse est factice’ [anguish is factitious] he taught the delights of poetic reverie and the difficult beauties of mathematical physics as the two principal strains of ‘une pensée heureuse’. He was, one colleague wrote, ‘L'être le plus humain que j'ai jamais connu’ [the most human being I have ever known].

The diversity of his accomplishments makes Bachelard difficult to assess and contributes to the neglect of his ideas. Such a range of competence and originality is almost unprecedented, and whenever one tries to engage with some part of it one senses, looming behind the work and exceeding whatever categories one might invent, the mythical figure of the man himself.

Bachelard's personal history gives him, indeed, a special aura. Born in Champagne in 1884, son of a shoemaker of Bar-sur-l'Aube, he spent nine years after leaving school as a postal clerk—first in the Vosges and then, after military service, in Paris. Studying part-time, he took a licence in mathematics and, after serving in the trenches in the First World War, became a teacher of natural sciences in his native village. ‘Toute ma vie est sous le signe du tardif’ [belatedness is my sign]: licencié in philosophy at thirty-six, agrégé at thirty-eight, he took his doctorate at forty-three with his Essai sur la connaissance approchée, an incisive attack on Bergsonism and Cartesian epistemology and a contribution to the problem of induction. Two years later he began his university career as professor of philosophy at Dijon and in 1940 was called to the Sorbonne.

The circuitous route to academic fame, a persistent love for his native countryside and the objects of rural life, a simplicity of manner and habit, and a playful bonhomie which delighted lecture audiences are all elements of the myth which now surrounds Bachelard, as are his own lyrical evocations of a life of reading. He is the true man of books (‘Lire beaucoup, lire encore, lire toujours’), who conceives no higher form of contentment: ‘là-haut, au ciel, le paradis n'est-il pas une immense bibliothèque?’ [Is not paradise an immense library?] A voracious writer as well as reader, he enriched the celestial library by thirteen scientific studies and ten of a literary kind.

To the history and philosophy of science Bachelard's contribution seems to have been five-fold. First of all, beginning in his 1928 thesis, he asserted firmly and repeatedly the importance of providing a philosophy adequate to a science in constant transformation, of which contemporary physics was only the most striking example. Instead of starting from a theory of the relationship between consciousness and the world, determining what can in general be known, and then asking whether science can attain objective knowledge, philosophers should reverse their perspective: ‘L'esprit doit se plier aux conditions du savoir’ [The mind must conform to the conditions of knowledge]. This is an elementary but radical inversion. Instead of deciding what can in theory be known, one should begin with what science in fact knows. Objective knowledge is nothing other than the interpersonal knowledge produced by science over a given period, and the philosopher's task is to provide the theory of scientific practice, the epistemology or epistemologies of actual knowledge. Since knowledge is something constantly in production, philosophy must be dynamic, dialectical, even plural or ‘dispersée’.

This attempt to account for scientific knowledge prompts Bachelard's second contribution: the attack on Cartesian epistemology. In erecting science on a foundation of clear and distinct ideas and in treating analysis as a reduction of the complex to its simple elements, Descartes asserted a continuity between immediate knowledge of a common-sense kind and the refinements of scientific knowledge. But there are two problems here. First, Bachelard argues, there are no simple phenomena; ‘every phenomenon is a fabric of relations’, and the simple is not an irreducible building block but a construction and reduction. ‘Simple ideas are not the ultimate basis of knowledge’. Second, the history of science demonstrates that science advances only by refuting empirical experience. Alchemy, for example, is psychologically very concrete. It takes as prime the elements fundamental to immediate experience. The task of science is to replace these objects with its own constructs: ‘Rien n'est donné; tout est construit’ [Nothing is given; everything is constructed]. Playing on the ambiguity of the French expérience, he declares, ‘une expérience scientifique est une expérience qui contredit l'expérience commune’ [a scientific experiment is an experiment/experience that contradicts common experience].

Common experience, in the form of intuitions, conscious and unconscious models, and psychological preferences, constitutes a series of ‘epistemological obstacles’ which Bachelard charts with verve and subtlety in La Formation de l'esprit scientifique and elsewhere (the most compact and elegant example is perhaps the discussion of changing intuitions and conceptions of mass in La Philosophie du non). Animist, substantialist and rationalist models or intuitions focus scientific discourse but also limit it. Each is eventually replaced by a new conception which becomes an epistemological obstacle in its turn. Descartes' sponge, a heuristic device, circumscribed possible insights, so that in Descartes ‘la métaphysique de l'espace est une métaphysique de l'éponge’ [the metaphysics of space is a metaphysics of the sponge].

If science works by refuting immediate experience and transcending its own theoretical frameworks, this has implications for the problem of induction. Like Karl Popper, whose Logic of Scientific Discovery appeared six years after Essai sur la connaissance approchée, Bachelard argued that nothing is decisively verified inductively and that knowledge is therefore primarily produced by negation, by falsification. Repeated positive results do not verify a hypothesis, while a negative result forces one to investigate failure and leads to more proximate knowledge (this is, Bachelard says, the philosophy of ‘why not?’). One must seek partial negation, ‘fine negation’. A truly scientific observation is polemical; the atom, for example, is nothing but the sum of falsifications to which its original image has been subjected. The production of knowledge through progressive falsification is what makes science a dynamic, dialectical enterprise to be ranged under the banner of ‘la philosophie du non’.

This dialectical rather than inductive conception of science leads to Bachelard's most influential contribution to contemporary French thought: the demonstration that science (and by implication other disciplines) is a discontinuous rather than a continuous process; not a gradual movement towards truth but a series of differential stages separated by ‘coupures’ or ‘ruptures épistémologiques’. Change occurs when epistemological obstacles are eliminated in favour of other models which provide new conditions of knowledge.

The notion of historical discontinuity has not only inspired notable French work in the history of science, by Georges Canguilhem and Michel Foucault among others, but is also central to Louis Althusser's Marxist revisionism and to the semiological study of modes of discourse which succeed one another in time. Bachelard has helped to promote the conception of history as a series of synchronic stages rather than a diachronic continuum. There are similarities, too, between the epistemological discontinuities described by Bachelard and the more sociological notion of successive ‘paradigms’ of scientific research proposed as a development of Popper's theory of science by Thomas Kuhn. But, while Kuhn seems inclined to make epistemological stasis (‘normal science’) the condition of scientific productivity, Bachelard, with his emphasis on the polemical nature of scientific statements and the need for a ‘poly-philosophisme’ to account for the open character of science, is in some ways nearer to the epistemological anarchism advocated by Paul Feyerabend as the condition of scientific progress.

Finally, Bachelard articulates what he called in Le Nouvel Esprit scientifique ‘a new way of looking at ambiguity’, which becomes a structure constitutive of science rather than what scientific analysis eliminates. Writing in the wake of the twentieth-century revolution in physics, he argues that everyone who learns science must make use ‘of not one but two metaphysical systems. Both are natural and cogent, implicit rather than explicit, and tenacious in their persistence. And one contradicts the other’. Scientists are by turn realists and rationalists, treating scientific truths as constructions and as discoveries. ‘It would not be difficult to show that in forming scientific judgments the most determined rationalist daily submits to the instruction of a reality whose ultimate structure eludes him, while the most uncompromising realist does not hesitate to make simplifying assumptions just as if he believed in the principles on which the rationalist position is based’.¹ One must conclude that ‘science is divided, in actuality as well as in principle, in all its aspects’, and that ultimately the dualisms, the incompatible perspectives not only structure scientific thought but that ‘the ambiguity is found to reside in the scientific phenomenon itself’, which is not something given with a Cartesian immediacy but constituted by the interplay of rationalism and realism, or logic and experimentation. The influence of this radical conception extends beyond the history and philosophy of science, for if scientific observations and science itself are shown to be constituted by the interplay of contradictory principles, then there is no reason to expect a simpler, more unitary configuration in other intellectual domains.

In France Bachelard's importance can scarcely be overestimated, but, despite the fact that his work contains, at least in embryo, the positions later developed by Popper, Kuhn and Feyerabend, it has had little influence elsewhere. The reason seems, alas, quite simple. Those who were exposed to the man and his teaching value his books. Philosophers outside France, exposed to the books alone, find their fundamental conceptions overlaid by a prose which, in its quest for the image or aphorism, belongs to another age of science: that leisurely and mondaine science of the eighteenth century which Bachelard so often analyses. Reading him is strenuous in a way wholly different from what one now expects of either scientific or philosophic writing.

The relationship between Bachelard's scientific and literary investigations is a question which always exercises commentators, who are professionally inclined to look for unity. Jean-Claude Margolin, for example, brings them together in a transcendental theory of the creative imagination, whose adventures Bachelard is said to place under the headings of animus and anima: the masculine principle of science and the feminine principle of poetic reverie. Not only is this an application of Bachelard's most insipid categories; it nullifies M. Margolin's more astute observation that imagination interests Bachelard precisely at the point where it ceases to be the prerogative of the individual subject: in poetic reverie, where the subject is wholly absorbed in an object whose universal mythic potential is released, and in science, where the ‘thought’ is implicit in a conceptual framework.² Mathematics, Bachelard asserts, is not the language of physics but the thought of physics; tensor calculus knows physics better than the physicist does.³

The important point, no doubt, is that Bachelard prefers both science and poetry at their purest and most extreme. He disdains a comfortable interdisciplinarity which would teach literary folk the second law of thermodynamics and produce or enjoy novels about scientists. He always maintained that his two enterprises were separate, though perhaps complementary. In fact, one is the converse of the other: the psychologically privileged images which Bachelard studies as obstacles to science become, with a change of perspective, manifestations of a basic affective and mythical universe explored by poetry.

Biographically this is indeed their relationship: Bachelard's study of epistemological obstacles, especially those of alchemy, provoked a series of books on the psychological and poetic qualities of the four elements. The earliest and best, La Psychanalyse du feu (1938), is fascinating in its hesitation between the scientific and the literary. Finally Bachelard chooses to revel in the imaginative possibilities of fire but draws on scientific writings for proof of the resilience of, for example, the sexual associations of fire and electricity. Though the psychoanalytic concepts he employs are personal and diffuse, he introduces psychoanalysis into literary studies as a way of analysing not authors but images, whose power is said to derive from their exploitation of a primordial and archetypal experience—not unlike that of a nineteenth-century village childhood.

The task of the poet, Bachelard writes, is to liberate in us ‘une matière qui veut rêver’ [matter that would dream], and for him, as for the Surrealists, the instrument of liberation is the image. The analyst tries to participate in the poetic reverie and to discover, beneath the manifest images, the primitive affective experience to which they refer us. By treating poetry in this way, Bachelard not only established ‘imagist’ criticism in France: he put it on a sounder footing than is usual in England or America, recognizing that, if images are made the privileged elements of a poem (‘the poem is a cluster of images’), critics should not simply study them as reflections of implicit or explicit themes but should develop a theory to explain why images are treated as the source of literary power.

Roland Barthes once observed that ‘present-day French criticism in its most flourishing aspect can be said to be Bachelardian in inspiration’.⁴ Though he was thinking primarily of Jean-Pierre Richard, Jean Starobinski and Georges Poulet who, like Bachelard, made criticism a participation in the affective universe of an author, Bachelard has also had immense dialectical influence: much recent French criticism is an attempt to transcend his limitations. His theory of poetry has the virtue of falsifiability. We dispute it by showing that the force and significance of images depend more on specific ideological or differential functions within a text than on universal associations: that images of earth are not always ‘stables et tranquilles’ nor walls and houses welcoming and protective.

Moreover Bachelard's hypothesis leads us to argue that much poetry does not simply evoke or invoke an immediate and ‘natural’ experience of the world but works much as Bachelard claims science does; breaking down immediate intuitions, deconstructing a universe of archetypical clichés, and reinventing the world by giving it an order which is discursive rather than immediately affective. Bachelard's work, by its explicitness, has provoked a more adequate account of poetry.

If it seems excessively paradoxical to claim that a thematic, atomistic and sentimental approach to poetry should stimulate a formalism centred on notions of languages, codes and structure, one might look at Bachelard's most accomplished piece of criticism, his Lautréamont of 1939. This is both a judicious appreciation of an author previously ignored, except by the Surrealists, and a series of methodological suggestions. Investigating the energy and aggression of Les Chants de Maldoror, Bachelard tries to define the ‘algebraic weight measuring the vital action of the diverse animals’ and constructs a system of valencies relating crab to louse, eagle and vulture.⁵ The notion of an underlying system of forces enables him to treat Kafka's bestiary (based on coagulation and petrification) as an inversion of Lautréamont's and to see Leconte de Lisle's as an ‘adjectival’ version of what is ‘substantivized’ in Maldoror. In this investigation he perceives that a psychologically or psychoanalytically astute literary criticism ‘would come to pose in different terms the problem of influences, of imitation. For this, it ought to replace reading by transference, in the psychoanalytic sense of the term’.⁶ Psychoanalytic criticism has indeed gone in this direction, thanks to Lacan rather than Bachelard, but Bachelard may have helped to establish a link between a linguistically-oriented psychoanalysis and poetic structures. But above all he seems to have grasped the formal mechanisms of modern poetry. Arguing that an elementary ‘poésie primitive’ is always a late conquest, a deformation of existing forms, just as the more fundamental projective geometry is founded on the systematic deformation of geometrical forms, he speaks of a ‘poésie projective’ based on the question: ‘What are the elements of a poetic form which can, with impunity, be deformed by a metaphor while allowing a poetic coherence to remain?’⁷ This is a question that seems to determine a good deal of twentieth-century literature.

Bachelard offers many observations which, when pursued, yield conclusions diametrically opposed to the conception of poetry he puts forward. He speaks, for example, of the need to understand works ‘in their own systematicity’, as one understands a non-Euclidian geometry in terms of its own axiomatic system, or of the need for a diagram that would indicate the direction and symmetry of a poet's metaphorical coordinations; but when Gérard Genette studies the network of imagery in baroque poetry in these terms, or Claude Lévi-Strauss formalizes the relations among images in myths, they produce not a Bachelardian profondeur, in which each image reaches downwards towards some primordial experience of being, but a system of elements whose functions are purely oppositional or differential. Far from having intrinsic archetypal meaning, ‘sun’ and ‘moon’ function as metaphorical operators which, so long as they are opposed to one another, can be used to express almost any thematic opposition. The universality Bachelard tried to discover in individual images now becomes a property of the system itself, operating through the displacement of one binary opposition by another. Bachelard is a seminal influence, but what is reaped is very different from what he sowed.

In his last works Bachelard's method changed. His interest in formal networks or systems wanes and instead of tracing images to their psychological sources so as to explain their power, he rejects psychoanalysis and seeks to experience phenomenologically the image's direct manifestation of being. His writing becomes somewhat insipid and sentimental, the poetic examples more banal, as he seeks in reverie ‘un confort mental’. Even here, however, the sustained pursuit of a poetic image or complex of images may yield surprising results, which display the insufficiencies of his explicit account of poetry better than polemic could. In La Poétique de l'espace, for example, while expounding a theory of the poetic image as ‘independent of causality’, without a past, ‘referable to a direct ontology’, in which it is the immediate manifestation ‘of the heart, soul and being of man, apprehended in his actuality’, he takes up Baudelaire's use of the term vaste.⁸Vaste designates, he argues, not a property of the external world but an inward intensity in which man becomes conscious of the immensity of his own being: ‘It transmits to our ears the echo of the secret recesses of our being’. Bachelard's examples do indeed show that vaste in Baudelaire seldom has a geometrical, objective meaning; it works to intensify while uniting inside and outside, thought and world. Bachelard calls it ‘a metaphysical argument by which the vast world and vast thoughts are united’; ‘always, in Baudelaire's poetics, the word vast evokes calm, peace, and serenity’. In the most famous example, from the poem ‘Correspondances’, ‘Vaste comme la nuit et comme la clarté’ [Vast like night and luminosity], it reconciles contraries: ‘under the banner of the word vast, the spirit finds its synthetic being’. Moreover, its sound is crucial: ‘in the word vast the vowel a retains all the virtues of an enlarging vocal agent. … Like some soft substance, it receives the balsamic power of infinite calm. With it, we take infinity into our lungs, and through it, we breathe cosmically, far from human anguish’.

The affective power of that syllable is indeed considerable, but ‘Correspondances’ identifies the synthesis Bachelard celebrates as blending which is also confusion:

Comme de longs échos qui de loin se confondent
Dans une ténébreuse et profonde unité,
Vaste comme la nuit et comme la clarté,
Les parfums, les couleurs, et les sons se répondent.
[Like long echoes which from far merge
In a shadowy and profound unity,
Vast like night and luminosity,
Smells, colors and sounds answer each other.]

The vastness of night and of brightness is both an unbounded extension and the effacement of distinctive contours—night and brightness share the property of making it difficult to distinguish, as in a night où tous les chats sont noirs (vaste itself here seems a syntactic operator that intensifies and unifies by effacing the difference between night and day). A vast unity is one whose immensity of scope effaces difference, and this unity, we are told, is produced by a distance which makes things seem similar.⁹ The model for this is the phenomenon which Bachelard's analysis emphasizes, sound; but where Bachelard sees an echo of the infinity of Being, ‘Correspondances’ identifies a confusion, an impression of unity produced as echoes heard from afar blend together. And Bachelard himself illustrates the functioning of echoes when at the end of his discussion, calling vaste ‘a vocable of breathing’, he is led by an echo in his poetic memory to append a note: ‘for Victor Hugo the wind is vast. The wind says: “Je suis ce grand passant, vaste, invincible, et vain”’ [I am this great passer-by, vast, invincible and vain]. Hugo's example locates the value of vaste in a play of echoes—the repetition of the v and the modulation of oral and nasal a sounds—which produces the sequence. While Bachelard's response relates vaste to a calm of being ‘far from human anguish’ and from human rhetoric as well, both the structure of the poem and the intertextual echo Bachelard hears from Hugo identify the rhetorical operations that produce effects of unity.

Bachelard remarks, ‘If I were a psychiatrist I should advise my patients who suffer from “anguish” to read this poem of Baudelaire's whenever an attack seems imminent. Very gently, they should pronounce Baudelaire's key word, vast. For it is a word that brings calm and unity’. Although critics will smile at the idea that this poem in particular, of which Paul de Man has given a most unsettling reading, should have therapeutic potency, Bachelard's claim highlights the problem that his late critical career poses. Before one rejects out of hand the divagations of La Poétique de l'espace or La Poétique de la réverie one ought to consider the implications: this late prose, like therapeutic use of ‘Correspondances’, is the direct manifestation of Bachelard's art de bonheur. When disturbed late at night by Paris traffic, he tells us, he induces reverie and transforms the traffic noise into the sound of a furious storm at sea, through which, safe in the tranquil harbor of his bed, he can sleep.¹⁰

This is scarcely reprehensible, nor is reverie an improper use of poetry. However, Bachelard's criticism, which explores and champions this use, provokes critical readers to insist, rather, on the disparity between pleasurable reverie and critical analysis—a disparity at which Bachelard's own examples sometimes hint. Bachelard thus reveals a division within literary studies and literary objects which may be analogous to the constitutive division or contradiction he earlier discovered in science. The literary work and literary study are perhaps constituted by the conflict between affective response and critical analysis, which the notion of the aesthetic presumes to bring together but which are in contradiction. His pursuit, in his late criticism, of the inherent signification of images, when set against his earlier interest in systems of relations, can provoke reflection on the conflict between affective response and analytical investigation and the extent to which literature and criticism are generated by this conflict, whose harmonious resolution may be a goal or an illusion rather than an effective possibility.

Bachelard's own writing evokes this problem in a peculiar way. His prose, which addresses the reader as a friend, works out its verbal fascination in hyperbole and exclamation, and constantly invokes, in a minor Proustian mode, memories of the muscle, hand and eye, poses the question of the relationship between the pleasurable use of literature and writing criticism. In Le Nouvel Esprit scientifique he speaks of intellectual happiness that comes in moments of synthesis or comprehension: ‘this intellectual happiness is the first sign of progress’. But in his reading of literature and writing about it happiness seems to come from reverie and a poetic evocation of the delights of reverie rather than from rigorous analysis. Bachelard read and wrote with greater pleasure and enthusiasm than most critics. Must we disqualify this attitude and make persistent angst the necessary critical temperament? Should we sneer at the man who spoke of the library as paradise and formulated as a daily prayer the prayer of the reader: ‘Give us this day our daily hunger’? These are not easy questions. They suggest that critical theory should perhaps devote more attention to investigating the conception of the aesthetic as the harmonious reconciliation of affect and knowledge and the literary disruption of that supposed harmony in poems and in criticism.

Doubtless the best way to approach Bachelard's remarkable oeuvre is to read and quarrel with La Psychanalyse du feu and Lautréamont, but one could also to make the late works a topos for reflection on the possible relations between writing and happiness, and on the way that a critical method may prove oblivious to the insights it makes possible.

Notes

1. Gaston Bachelard, The New Scientific Spirit (Boston: Beacon Press, 1984), pp. 1-2.

2. Jean-Claude Margolin, Bachelard (Paris: Seuil, 1974).

3. Bachelard, The New Scientific Spirit, p. 55.

4. Roland Barthes, ‘The Two Criticisms,’ Times Literary Supplement, 26 September 1963.

5. Bachelard, Lautréamont (Paris: Corti, 1939), p. 30.

6. Ibid., p. 145.

7. Ibid., p. 70.

8. Bachelard, The Poetics of Space (Boston: Beacon, 1969), pp. xii-xiv. The discussion of vaste is on pp. 190-8.

9. In ‘L'Exposition universelle de 1855’ Baudelaire describes how ‘tous les types, toutes les idées, toutes les sensations se confondraient dans une vaste unité, monotone et impersonnelle, immense comme l'ennui et comme le néant’ [all types, all ideas, all sensations, would blend in a vast unity, monotonous and impersonal, as immense as boredom and nothingness], Oeuvres complètes, ed. Claude Pichois (Paris: Gallimard, 1975), II, p. 578. Here the word vaste appears in a passage emphasizing that unity is produced by confusing the distinct. L'ennui suggestively echoes la nuit of ‘Correspondances’.

10. Bachelard, Poetics of Space, p. 28.