Chapter two: The Thinking in Seeing

Chapter two of the book



Anders Fogh Jensen 


Mellem ting. Foucaults filosofi
Frederiksberg: Det lille forlag, 2005
2. edition.: København: THP, 2012

The Thinking in the Seeing

The Thinking in Seeing. 

Foucault and the French Epistemology




Foucault’s histories have their own history. In this chapter I argue, that Foucault is a straight forward continuation of the French Epistemological Tradition in the 19th and 20th century. One of the main concerns of Foucault was to ask which questions could be asked at a certain point in history – and which problems these questions gave rise to. In this chapter I argue, that the French Epistemological Tradition (called l’épistémologie) prepares the path for the foucauldian archaeology and genealogy in that it re-establishes historical thinking in an epistemological and ontological context. Foucault turns the gaze from nature to society and from the science of nature to the sciences of man but continues the French Epistemological Tradition in a traditional way.

The article presents the thinking and the main concepts of the French Epistemology. From its prehistory in G.W.F Hegel and Auguste Comte over Pierre Duhem and Émile Meyerson to Gaston Bachelard, Alexandre Koyré and Georges Canguilhem. In this movement the article presents how the main concepts are developed: historical rationalism, construction, sedimentation, realisation and reality, epistemological obstacles and ruptures. It argues 1) that this historical rationalism prepares the path for what is later thought of as constructivism, and 2) that the historical rationalism is still able to think metaphysically because it thinks the transformation of the thought in the seeing as a process of realisation. Finally the article shows 3) how Foucault continues this thinking.



The Given Is Not Given

The given is not given. This is one way of expressing the main thesis of historical epistemology.

Historical epistemology is a tradition in French philosophy, a school within the philosophy of science which is concerned with how it became possible for the given to be given. The answer it gives goes like this: By liberating imagination from the eye, imagination might re-envision a world which the eye is only then enabled to see. Consequently, thought gives shape to ways of viewing the world only by giving shape to itself first.

What can history teach us? It can teach us how we discovered that we were mistaken. It can teach us that just because we are now able to see that we were mistaken then, it is not given that we still fail to see how we are mistaken now. The lesson that may be learned is that we did not discover the old mistake by taking a closer look but by adjusting the world within thought itself in order that something new could give itself. In fact we may learn that it is precisely by looking that we go blind. In this way the history about history is also always about the present.


The Maternity Ward of Mistakes

Historical epistemology offers an answer to the question: what is a mistake? Its answer is: A mistake is an action that we subsequently recognise as a mistake. In this way history is introduced into epistemology inasmuch as a historical horizon is made relevant to the judgment of a proposition of truth.

To do justice to your own personal history is to understand yourself in those situations in which you made certain decisions at the expense of others. It is to bring the very premises of the situation to light rather than confining yourself to viewing the situation in the dazzling light of hindsight only. To bring the premises to light means to consider the information and motives at hand within the situation, but it also means to consider the kind of sense and the amount of experience you were able to draw upon at the moment in question. To be experienced means, firstly, that you have had experiences, but, secondly, it also means that you are able to see those experiences within their specific historical circumstances, as taking place on the basis of inexperience; it is to judge them without condemning them because they are products of inexperience. Thirdly, it means to acknowledge that those inexperienced actions are actually the conditions that have led to today’s increased experience. Finally, it is to see that if they appear as inexperienced, it may be due to the fact that they are viewed from a position of experience. The fact that I acted in an inexperienced way gave me experience, and that is exactly why my actions then may seem inexperienced today; considering the conditions of inexperience at the time, I was not necessarily to blame.

This is how epistemology thinks about science. To see the history of science as “a cemetery of errors, or a collection of monsters, which have justly been relegated to the junk room and are only suitable for demolition. A graveyard of forgotten theories or a chapter in Die Geschichte der menschlichen Dummheit” (Koyré 1954: 257, current translation) is not doing any justice to history. It is to meet history in its old age rather than in its innovative youth. With this, yet another theme has been introduced, apart from the historical sense of justice: the power of thought to engender cognition.

Within the tradition of epistemology history is understood as a cognitive process without any autonomous drive or final aim; rather, the discovery of a mistake leads to a new claim. To Hegel’s mind every negation is a definite negation; that is to say, it leads to a new position that is superimposed on the other (Hegel 1807: 67). Or rather: it is only by establishing a new position that the old position may be negated. When science is driven forward by the establishment of such positions, it is entirely a matter of metaphysics. The greatest leaps in the history of science are made not by the eye but by thought: It is when thought negates what the eye sees that the eye is enabled to see something new, causing science to move. That is why epistemologists refer to themselves as rationalists.


Historical Rationalism

Epistemology is historical rationalism. What does that mean?

Karl Popper is the originator of a falsificationist method of science which is also rationalist, but only partly historical. The falsificationist method demands that a claim, in order for it to be scientific, has to be put forward in such a way that it excludes the possibility of certain events to happen.

Why does a scientific claim have to exclude something? It has to do so because the inference of a general, positively formulated claim from singular positive observations is not a logically valid conclusion (the problem of induction), whereas it is alright to base the invalidity of a general positive claim on negative observations. In other words, it is not logically valid to infer that “everybody is” because “some are”, whereas it is valid to infer that “it is not the case that everybody is” from the observation that “some are not”. In order to bring the last conclusion into play a theory or scientific claim has to remain open to the possibility that it is mistaken if certain observations occur. Freud’s psychoanalysis, for example, and Karl Marx’ theory of history both share the problem that they are incapable of stating anything that is observable in principle which causes their theories to be refuted. Quite the contrary, Freud and Marx would be ready to see all possible events as but confirmations of oppression or the sex drive. In other words, it has to be possible in principle to carry out an experiment that shows that what a theory claims to be impossible is actually possible (Popper 1953: 37-39).

Who is to judge a scientific claim and its empirical conditions of falsification? Reason is. What reason? The reason. Although Popper’s philosophy of science is about the historical evolution of science, it places rational judgment outside historical processes.

In response to this, historical epistemology – and Foucault along with it – would claim that the judge himself is part of the process he is judging. What is seen as reasonable at a given point in time is part of the historically conditioned scope of thought that formulates the scientific claim. And, secondly, it claims that the imagination that conceives the claim also partakes in the same historical scope of thought.

Whereas Popper and other philosophers of science are fond of distinguishing between the context of discovery and the context of justification, in order to keep the genesis of a claim apart from its validity, historical epistemology claims that these two are indistinguishable. The reason that finds it reasonable to advocate a certain case is also the reason that judges the case, just as it is this reason that has the power to break the laws of tradition. To take an interest in the latter is to take an interest in the forces of cognition and realisation. To take an interest in the conspiracy of the former two is to take an interest in the actual history of thought in the making.

It is the very conspiracy between judge and advocate that caused the old mistakes, which now appear as such obvious mistakes to us, to appear as obvious truths at the time they were in force. You get the wrong idea of history if you think that it consists of a great darkness that is gradually illuminated. Rather, light replaces light; insights replace insights. Being an experienced historian of science and being experienced in life involves the same kind of respect for the light of the past; to recognise that it never was darkness.

Historical epistemology is practically anchored. Gaston Bachelard is a trained physicist, Jean Cavaillès is a mathematician and logician, Georges Canguilhem has devoted a lot of time to biology and Alexandre Koyré to the history of mathematics and astronomy. They are writing a history of science that reconstructs epistemological scopes, thus unfolding the epistemological scope within which Foucault develops his thoughts.

They are so closely in touch with science that they are able to sense the way it progresses, and they are so immersed in philosophy that they are able to see that reason in the world of science and reason in the world of philosophy are not worlds apart; on the contrary, it is the same reason engaging in a dialogue with itself – reason, imagination and experiments in dialogue with one another.


The Antecedents of Historical Epistemology 

Within the history of thought there are two tendencies in how the relation between reason and science is viewed.

On one hand there is a school reaching back to Aristotle’s claim in The Metaphysics that “knowledge is measured by the object of knowledge” (Aristotle The Metaphysics 305, 1057a). You have to adjust consciousness according to the object you wish to arrive at a cognition of, and since the sciences do not work with the same objects, their approaches differ from each other. If you mix up these approaches, you make the mistake Aristotle describes as “metabasis ais allo genos”, which is what we would refer to as a fallacy of category in modern philosophy.

In Regulae Descartes sketches another idea of science, claiming that “[…] the sciences taken all together are identical with human wisdom, which always remains one and the same, however applied to different subjects” (Descartes 1650: 360/3). Cognition remains unaltered by the object it apprehends, just as sunlight remains unaltered by the object on which it shines. In other words, Descartes breaks the law of the Aristotelian metabasis by establishing a mathesis universalis: Mathematics is envisaged as a universal form of cognition and the geometric method as a universal method capable of generalising the sciences as a universal science. Leibniz speaks of an arithmetica universalis. Kant not only speaks of scientific cognition, but strives to establish the universal human conditions of all cognition.

Descartes also claims something else. Not only does reason need something external to itself to be able to arrive at the cognition of something, like sight needs the sunlight (Plato) or the intellect depends on God (Augustine). Descartes also makes it clear that reason is complete within human reason already prior to the act of cognition. The process of cognition does not play any significant role until Hegel who claims that cognition is only cognition when an effort of cognition has been made. To understand how we understand is also to understand that we may not understand. In light of this it becomes relevant to Hegel, Kierkegaard and the historical epistemologists to go through the movements of cognition with a consciousness that has yet to understand (Bachelard 1938: 18).

One may wonder why the Cartesian-mathematical basis of the cognition of the universe is not feasible until Descartes and Galilee. The answer is to be found in the kind of circularity that takes place between pictures of the world and cognitions of the world. Before Galilee, all of nature was yet to be written in the letters of mathematics; the laws that applied to the divine celestial bodies were different from the laws that applied to the ones below the moon. Before the Renaissance the will to land on the moon was an altogether different matter than today; it would correspond to the idea of landing a vessel on God.


The Spirit of Epistemology

Comte: Scientific Spirit

Auguste Comte is preoccupied with the task of determining a ‘scientific spirit’, an “esprit positif”. Such preoccupation with the superposition of ‘spirits’ or ‘mentalities’ is itself part of a historico-philosophical spirit that characterises the 19th century. In ‘the spirit of this spirit’ Comte claims that reason is most advanced in the sciences that can be expressed by formula: mathematics, physics, chemistry, followed by biology and sociology. Comte thinks of science as an evolution of layers in human history, which is why a proper understanding of any science has to take in the historical layers that are deposited beneath contemporary science. For example, the curiosity of modern positivist science contains reminiscences of more primitive sciences.

The difference between old and new sciences lies in the kind of answers that are sought to satisfy this curiosity. In Comte’s philosophy of science three modes of 19th century thinking intersect. First there is the idea of a progressive movement. In Comte this takes the form of three phases: a theological, a metaphysical and a scientific phase. In the theological phase man entertains hypotheses of causes, e.g. about forces personified by gods. This phase is further divided into three sub-phases: a fetishist (things as animated by life), a polytheist and a monotheist phase. In the metaphysical phase man entertains hypotheses about essences and identities as the ontological foundation of phenomena, whereas in the positivist, scientific phase we relate only to the regularity of laws – with the purpose of being able to predict events (Comte 1844: 53). The theological and metaphysical phases are not done away with by a positive phase; they are embedded within the superpositions of reason. As in Hegel, the new position contains the negation it negates (Hegel 1807: 67/65). We still need religion in a scientific culture, but what matters is to be the master of nature. Historically, the theological phase is attributed to Antiquity, the Middle Ages and the Renaissance; the metaphysical phase is attributed to rationalism which passes into the Enlightenment which then gives rise to the scientific phase.

Secondly, this history of science includes a political and a social dimension: Scientific development is to serve humanity, and gradually, although with some delay, the development of reason will generate more sensible political and social conditions; that is why the philosophy of science is socially significant. This is an ambition that Auguste Comte has in common with the logical positivism of the 20th century, the Vienna school. Actually continental 20th century positivism has its source in an examination of meaning. What is examined is the scientific status and meaning of sentences. This does not develop into a question of the constitution of reality until the onset of American logical empiricism. According to Comte, positivism is as far from empiricism as it is from mysticism (Comte 1844: 24).

Thirdly, one of the great themes in the 19th century is that the evolution of the human race (phylogenesis) is mirrored in the evolution of individual organisms (ontogenesis). Comte makes a variation of such a biogenetic claim by proposing a form of logogenesis: that in terms of consciousness any individual human being evolves according to the history of human kind, from theologian to metaphysician to physician, or, if you like, from a superstitious being to an inquisitive child to an adult.

In the 20th century the philosophy of science continues in France in the form of reactions to Comte, for and against him. It all starts with Pierre Duhem and Émile Meyerson.


Duhem: Theory, Interpretation and Use

In the work of the philosopher of language, William v. O. Quine, one finds the claim that a language is so self-closed that it is impossible ever to know precisely how a word is to be translated to another language. If a people utters the same word every time they point at what we call an ant, we will never know whether they mean ants, antness, ant-essence, ants when the sun is shining or ants in cloudy weather, etc. (Quine 1960: 29-30). Although we may approximate the meaning of other language we will never be able to understand it completely (Fogh Jensen 2001: 115-136).

This is what is called the Quine-Duhem thesis, and the last name owes to an idea Pierre Duhem had about the experiment. In French an experiment is called ‘expérience’, which in English may be translated to both ‘experiment’ and ‘experience of’ in the sense of both the experience of an event and the accumulation of experience or knowledge about something. In Duhem’s philosophy of knowledge ‘expérience’ means ‘experiment’. An experiment is not only about observing a phenomenon, it also involves a theoretical interpretation of the phenomenon (Duhem 1906: 223). The theory that interprets the phenomenon does not express a claim about reality beyond theory, but a claim about how the laws within the experiment relate to one another. This differentiation recurs in Louis Althusser’s differentiation between l’objet théorique and l’objet réel, the theoretical object and the real object, and it propagates further in French philosophy of science (Althusser 1965 piece 10-11; 45-46). The idea of theory as a construction that makes it possible for us to see reality recurs in Foucault and his contemporaries. For instance, one can also find the claim in Bourdieu that sociological theory does not make any claims as to how social reality is constituted, but it is a model of reality which enables us to understand other sides to it than what common sense tells us (Bourdieu, Chamboderon and Passeron 1968 76-77 and 82). In other words: theory offers a grasp of reality, not a claim to reality. Yet, with this no claim is made, neither by Duhem, Althusser, Foucault nor by Bourdieu, that reality is not accessible. It is claimed that an experiment is not about demonstrating facts as they may actually occur in reality, but about judging the connection between theory and facts. To understand an experiment requires an insight into its theory which you can only understand if you understand the language of science. To understand a speedometer as an expression of velocity requires a theory of movement; the position of the pointer of the speedometer expresses velocity as a relation between movement as speed and the actual movement.

In this sense, a language is also a ‘spirit’, a spirit of science. Just as a particular law cannot be understood in isolation according to old hermeneutic principles, but has to be related to the spirit of the law within which it is written in order to be correctly interpreted, an experiment must be understood within the spirit of science. There are connections back to Comte’s positivist spirit and further back to Pascal’s concept of ‘ordres’, and here something is activated which is later to be formulated by Bachelard as a ‘scientific mind’ or ‘spirit’, an “esprit scientifique” (Pascal 1660 §793(308)) (Bachelard 1938 and Bachelard 1940: 2-17/5-16).



Moreover, another claim about technique becomes important in subsequent developments of epistemology: the claim that the use of a technical instrument is possible only on the basis of a theory, that is to say on the basis of a theoretical interpretation of phenomena (Duhem 1906: 248). You cannot use the speedometer as speedometer without a theory of velocity. This is expanded upon by Koyré and Bachelard who determine the instrument as materialised theory.

When instruments from the laboratories are introduced into everyday life, they work like Trojan horses, theories pouring out of them. Through persistent use of these instruments the metaphysics of their theories comes to be taken as unquestionable. Thus you may for example hear meteorologists claim that we feel a temperature despite the fact that temperature can only be read as figures on the instrument called a thermometer. You do not feel voltage; you get an electric shock. In other words, Comte’s idea about the superpositions of phases does not only apply historically to newer phases as superposed on older phases. It also applies materially: metaphysics is the basis of theory. Theory is the basis of a technical instrument. When technology moves from the laboratories into our kitchens and living rooms, metaphysics comes along. Thus the wrist watch is really Galilee’s metaphysics – time written in the letters of mathematics – wrapped around everyday life – or the other way round, everyday life wrapped around Galilee’s metaphysics. Bachelard’s ‘psychoanalysis’, Canguilhem’s and Althusser’s ideological criticism, Foucault’s archaeology are all attempts at digging in the unconscious of science and everyday life with an eye to unearthing hidden metaphysical and ideological layers.

The idea that thought sediments itself in materiality is developed by Foucault in a number of different spheres. First of all in the body:

The body is the surface of the inscription of events. (Foucault 1971: 143/375)

An event is a breaking up and a rearrangement of the rules of interpretation, e.g. in the ways in which man is read by the sciences. They take form in human bodies, e.g. in the ways in which people express their sexuality or the ways in which they eat. In Foucault’s grammar of the social this moulding is considered as a written expression.

Secondly, architecture is materialised theory: Hausmann’s boulevardisation of Paris, for example, is a materialisation of an idea of control, just as houses of detention are a materialisation of the idea of locked-in exclusion. Through materialisation theory sediments in social behaviour and existence: thirdly, institutions are materialised theory, the school doctor, for example, as the sedimentation of an idea of optimising the population. Fourthly, instruments are sedimentations of thought: the rack, for example, is a materialisation of a certain way of thinking in regard to punishment just as the surveillance camera is a materialisation of certain ways of thinking based on surveillance and punishment (p.#intratextual reference#). The social technologies are the materialisation of rationalities in various techniques, for example in questionnaires, tax returns or tachographs.


Meyerson: Metaphysics

A more radical ontology must be mentioned here because it becomes significant later on in the way in which the spirit of epistemology thinks metaphysics. In 1908 Émile Meyerson advances the thesis that “la science n’est pas positive” (Meyerson 1908: 438-439): the idea that science is not only a registration of the relation between a theory and phenomena, given outside a theoretical complex, but that science is a theoretical interpretation that establishes a reality. Here Meyerson makes a connection with the old thesis by Parmenides, Spinoza, Schelling and Hegel that a concord exists between thought and being which causes new thoughts to establish new ways of being (Meyerson 1907: 473).

Hence, it is an epistemological paradox that science, in refuting reality and establishing another idea of reality, turns into a producer of reality and, with that, becomes realistic. Normally we would refer to a realist as someone who relates to the real (the real as opposed to the speculative), but in epistemology realism is referred to as that which produces or shapes reality. A political realist normally claims to be a realist because his thoughts square with reality, as opposed to the idealist who uses the imagination. Historical epistemology, however, makes the same move in relation to science and philosophy as Che Guevara did in relation to political realism, that is, it asks for impossibilities: “Seámos realistas. Pedímos lo impossible.”

What is here taking shape within epistemology is, as Bachelard has it at a later point, that facts are constructed before they are ascertained (Bachelard 1940: 36/30).

 At the same time the philosophical movements that have been mentioned here, which all point to the philosophy of knowledge of the 20th century, involve a drawing closer to the sciences inasmuch as philosophers realise that the sciences are actually working and making progress. On the other hand, they constitute a movement away from an empirical interpretation of the sciences towards an examination of the role of thought in scientific development. In this respect, the movement within the tradition of epistemology, from Bachelard and onward, is returning to the laboratory inasmuch as scientific experiments and cerebrations are seen as decisive factors in the development of reason as imaginative and constructive: concepts are now tested in the laboratories with a view to further development. As for those who fail to carry out their work with concepts in the laboratories,

They are not philosophers at work, they are summarising philosophers. (Bachelard 1949: 8-9, present translation)


Historical Epistemology: The History of Philosophical Lawbreaking 

“There is no wave, there is only the ocean”. This was Claude Chabrol’s response to those cineastes who formed a group of Parisian film directors in the 1960s under the name of ‘La nouvelle vague’. The same problem of categorisation appears in French epistemology. Nevertheless, as was the case on the screen, a particular philosophical current seems so apparent that categorisation does not seem like such an unjust measure after all. Accordingly, in what follows we shall take the liberty of defining the epistemological tradition extensionally as including first and foremost Gaston Bachelard, Alexandre Koyré, Georges Canguilhem and Jean Cavaillès. A large group of students and offspring are connected with the ideas of these four thinkers, including François Dagognet, Louis Althusser, Jean-Claude Pariente, Gilles-Gaston Gragner, Pierre Bourdieu and Michel Foucault.


Bachelard, Koyré – the epistemology of mathematics, physics and chemistry

Cavaillès – the epistemology of mathematics and logics

Canguilhem, Dagognet – the epistemology of biology and medical science

Foucault – the epistemology of medical science and the human sciences

Althusser – the epistemology of history

Parientie, Gragner – the epistemology of linguistics

Bourdieu – the epistemology of sociology


Epistemology may be defined intentionally as a concern with the relation between science and philosophy where philosophy is understood as metaphysics, rationality and reason. The point of departure is that not all questions can be asked at all times, and, accordingly, the object is to study how it became possible to ask some questions. In particular the object of study is the obstacles thought has had to surmount in order to bring a given question to light. One of the weaknesses of science, as well as of the pragmatic uses of science that have become habituated to given ways of viewing the world, is that they take for granted that the given is given.

One does not become a geometer by measuring triangles (Hegel 1807: 31/29). A triangle is attributed with 180 degrees by a theoretical construction. Epistemology is concerned with the question how it became possible to construct the given in such a way that it passes as actual reality in our experience.

The object of epistemology is the reason of science. Epistemology is not a science because a science takes the world as its object. Science examines the world with theory. Epistemology examines scientific examinations of the world with philosophy; it is concerned with world pictures. When historians of science see a precursor of Nicolaus Copernicus in Aristarchus of Samos who lived 1800 years before because both of them claimed that the sun was the centre of the universe, then it is exactly the business of epistemology to call attention to the difference between the two world pictures: Aristarchus was thinking within a Greek world picture, Copernicus within a mathematised world picture, which makes the two claims widely different. Epistemology is about world pictures.

With this I have also indicated that the endeavour of epistemology is often the opposite of that of hermeneutics. Whereas hermeneutics is concerned with the question of how it is possible to understand the past on the basis of common features between people, culture or history that may bridge temporal distances, epistemology is concerned with the exposure of different worldviews: how far apart seemingly identical worldviews may be and how close together seemingly distinct worldviews may be.

Complaisance in seeking, finding and celebrating precursors is the clearest symptom of an inaptitude for critical epistemology. (Canguilhem 1966: 21/205)

Comte’s thesis about a positive spirit is elucidated by Bachelard’s thesis about a scientific spirit and is sharpened to a point by Foucault’s thesis about the episteme: Different scientific theories constructed within the same historical horizon look more like each other than they look like their own so-called precursors. The will to precursorism in history is a violation of the ruptures of history.



How then does history progress if not by connections and continuity? The answer is that it moves rhythmically: It alternates between the exploration of the possible in thought and the eye’s exploration of the real, that is to say, between the realisation of what is actually possible and the eye’s confirmation of the real. Scientific exploration encounters obstacles and problems and in order to overcome these, thought has to speed up and ram against such obstacles and problems to facilitate a leap up into a new construction of reality.

“On fonde en construisant”, one founds through construction, says Bachelard (Bachelard 1949: 21). What does he mean by that? He means that in order to be able to construct something scientifically one needs a metaphysical foundation. To take Galilee as an example:

Philosophy is written in this grand book, the universe, which stands continually open to our gaze. But the book cannot be understood unless one first learns to comprehend the language and read the letters in which it is composed. It is written in the language of mathematics, and its characters are triangles, circles, and other geometric figures […] (Galilee 1623: 600/237-8).

To say that the world is written in the language of mathematics, i.e. that it is fundamentally measurable, is not a scientific claim. It is a metaphysical claim. Galilee posits a metaphysics in order to practice mathematics on celestial bodies. This is a transition that Koyré refers to as the transition from a closed cosmos to an infinite universe (Koyré 1954: 260).

If the world is measurable in numbers and the sequence of numbers is infinite, it means that the world cannot be limited. The Greek cosmos with its qualitative differences is closed and a new book is opened. In the new book movement is no longer understood as an advance toward something, it is seen in terms of affected bodies that collide or do not collide. The matematisation of physics is a metaphysical movement.

In the same way Saussure’s claim that language is a system of differences is not a linguistic but a philosophical claim:

[t]he object is not given in advance of the viewpoint: far from it. Rather, one might say that it is the viewpoint adopted which creates the object. (Saussure 1916: 23/8)

Saussure practices metaphysics when he makes claims about what language is – a metaphysics that makes synchronic linguistics possible.

When Watson and Crick made the claim that genes are codes, they made it possible to study the body as a communication system (Crick and Watson 1953). That is not a biological claim; it is metaphysical.

Epistemology wages war on more than one front: It fights the self-image of linear philosophy and the heritage from Descartes and Kant that presupposes reason to be already complete in man and in need only of being put to use. In this respect the historical epistemologists appear to pull the weight of Hegel when he raises the objection against Kant that reason is not ready for use until it thinks, that it comes into being only as it thinks (Hegel 1807: 58/55). On the other front, epistemology fights against empiricism and realism, which are really two of a kind in the self-knowledge of positivism, “l’autointerprétation et l’autorestriction positivistes de la science” (Koyré 1954: 265).


Realism and Empiricism 

According to the prevalent philosophical tradition, realism is about the nature of the objects we arrive at a cognition of and empiricism is about the ways in which we arrive at a cognition of something. Empiricism claims about cognition that 1) the arrival of cognition is passive, and 2) knowledge is achieved by summing up observations, i.e. by demonstration without explanation.

However, Bachelard objects, the world is not speaking all by itself. It only speaks when prompted.

For a scientific mind, all knowledge is an answer to a question. If there has been no question, there can be no scientific knowledge. Nothing is self-evident. Nothing is given. Everything is constructed. (Bachelard 1938: 14/25)

Historical epistemology is concerned with the question: Which questions are possible to ask the world at a certain point in time? The frame of the question also indicates the frame of the answer. If you ask about the square root of time, you do not get answers like “red sunrise”.

Foucault inserts the problem into the question: The problem indicates what can be known, how this knowledge may be procured, administered, etc. The scope of epistemology is stretched out between the question and the problem.

Thus the objection of epistemology against the positivist self-knowledge of the sciences does not concern mathemathisation. On the contrary, mathemathisation is considered as a step forward. What epistemologists object to is the implication that the answers mathematics may give are supposed to exist prior to the questions formulated in the language of mathematics.

For example, in a Danish history of science it says:

A rational cosmology would not be possible if the universe was not a well-ordered whole. Since we actually do have a rational cosmology, achieved through 3000 years of theory and collection of data, we have every reason to conclude that the world as a whole is regular and ruled by laws and in that sense well-ordered. It is a cosmos, not a chaos. Of course it may be objected, as some philosophers have done, that the order the world appears to give proof of is the order established by astronomers and physicians and not the world’s order as such. That is an argument you have to be a philosopher to accept (Pedersen & Kragh 2000: 9, present translation)

My point exactly. Epistemology does not disagree. Historical epistemology is about how the well-ordered world was brought about through a metaphysical and not a scientific construction. It became possible to see the mathematical order of the world only after its mathematical construction.

Is the given given before it is constructed as given? No, as Koyré says, data does not give itself without a prior positive knowledge about what one expects to see:

You don’t look if you don’t know that there is something to see, and especially not if you know that there is nothing to see (Koyré 1948A: 351 note 2, present translation)

Argument one against the realism of the sciences: One always views the world through a theory about what it is that one sees. There is thinking in seeing.

Argument two against the realism of the sciences: One tends to cling to the theory one has developed, even if experience contradicts it every day. This is what is later to become Thomas Kuhn’s argument against positivism and falsificationism: If we had only been ready to reject theories on the basis of isolated observations rather than stubbornly sticking to them, science would never have moved on.

In Foucault’s philosophy it is a question of unearthing the precursory principles of order that instigate the empirical gaze that bears the stamp of a given period in time, the episteme. It is the silent order behind the sayable that determines which questions can be asked at any given point in time and what kind of answers these questions construct ahead of themselves.


The Realism of Epistemology: Realisation

Does that mean that epistemology is anti-realism? Not at all. Rather than realism Koyré and Bachelard speak about realisation: constructive work is a realisation of what can be given to experience. In this connection Bachelard speaks about a constructed realism, whereas Koyré speaks about a mathematical realism: It is through its construction that mathematics actually become real. It is through the definition of uniform acceleration as two times velocity divided by the square of time that uniform acceleration assumes a mathematic expression as a proposition about the world.

Thus realization takes precedence over reality. By so doing it demotes reality. A physicist only really knows reality when he has made it come real […] (Bachelard 1940: 36/30)

An epistemological rupture happens through several breaks and ruptures: I) with Copernicus, Earth is displaced from the centre of the universe. II) Galilee puts forward the law of falling bodies and the law of inertia, and to do so he has to presume a mathematised space. Galilee examines thrown bodies, projectae, in order to examine the planets, planetae, but he has to be quiet about it because of the Church – Galilee turns all space into the same space. III) Kepler’s discovery of the planetary orbits as elliptical deprives the heavenly bodies of the perfection of the circle. IV) Newton does away with the remaining bits of the classic ideas of cosmos by synthesising attraction and gravitation: they are one and the same force. The world has become matter in motion.

With Copernicus not a great deal happens in metaphysical terms. His explanation for the position of the sun at the centre of the universe is that God planned it that way in order that the sun may better light up the planets. The offensive part of his explanation is that 1) if the Earth revolves around the sun, it means that it is but one planet among others of the same kind, and 2) that it is not still, it moves – and movement is seen as imperfect. The latter idea has yet to be broken and rearranged to make Newton’s universe possible.

Order is in the world, but man can only discover it by constructing it. An experiment is a question asked of nature and in order to receive an answer one has to phrase it in “nature’s own language” – i.e. the language of mathematics.

To Bachelard it does not make much sense to speak of a world order as designed by nature itself. Man orders the world by means of his comprehension – whether comprised in terms of substances, relations or quarks.

Reality is not nature, pure and simple. It needs to be laboured to become an object of scientific experience. (Bachelard 1972 134, present translation)

The same claim is made by Heidegger when he speaks about how the world becomes the world through human understanding: Chaos is turned into an ordered world the moment human understanding assigns itself to it (Heidegger 1927 87/120).

And empirical science certainly occupies an important position in epistemology, but the eye only sees clearly when sharpened by thought:

Empirical thought is clear in retrospect, when the apparatus of reason has been developed. (Bachelard 1938: 13/24)

It takes a scientist of a decidedly empiricist cast of mind to deny that claim.

In Foucault the ‘realisation’ of a new order happens more anonymously, and Foucault prefers to speak with Nietzsche and Heidegger about events or occurrences. Prevalent lines of fracture in history are often rejected in favour of tiny events, as when the date of the formation of the prison system is fixed at “that glorious day […] when a child in Mettray [the prison] remarked as he lay dying: ‘What a pity I left the colony so soon’” (Foucault 1975: 300/293).

To Foucault realisation is the realisation of a problem. A problem seldom has one father only; it is an event, a change in rules of interpretation that puts man in a new situation:


[…] if interpretation is the violent or surreptitious appropriation of a system of rules, which in itself has no essential meaning, in order to impose a direction, to bend it to a new will, to force its participation in a different game, and to subject it to secondary rules, then the development of humanity is a series of interpretations. The role of genealogy is to record its history: the history of morals, ideals, and metaphysical concepts, the history of the concept of liberty or the ascetic life; as they stand for the emergence of different interpretations, they must be made to appear as events in the theatre of procedures. (Foucault 1971: 146/378-379).

Foucault’s so-called genealogy shows how history realises new problems.


In Dialogue with the Laboratory

One may now be caused to fear that epistemologists were busy claiming that philosophy, reason, is the real drive of science. But this is not the case. Science is propelled by a dialogue between reason and science. The philosophy of science is understood neither by the scientist nor the philosopher before they have understood this.

Scientists, indeed, consider metaphysical preparation unnecessary […] For them the hour of philosophy strikes only when the actual work is over; they therefore conceive of the philosophy of the sciences as the sum of general results of scientific thought, […] For the scientist, the philosophy of science still belongs to the realm of facts. (Bachelard 1940: 2/4)

However, the picture is just as superficial in the philosopher’s eyes, only the other way round:

 […] the philosopher simply asks science for examples to prove the harmonious activity of the mental functions, but he thinks he is empowered, without benefit of science and with precedence over science, to analyze this harmonious activity […] For the philosopher the philosophy of science never remains totally in the realm of facts. (Bachelard 1940: 3-4/5)

 They valorise the a priori and the posteriori respectively, and the rational and the empirical. But actually, Bachelard further maintains, both will gain by making room for the other:

Indeed the one triumphs by assenting to the other: empiricism needs to be understood; rationalism needs to be applied. (Bachelard 1940: 5/6)

However, science plays a greater role in its time than that of an interlocutor; it is the agent of change. Koyré: reason learns from the sciences. Bachelard: science instructs reason. How is that?

First of all this is so because science does not come after its time, it is its own age. Kepler is able to calculate the movements of celestial bodies only because a softening of the Aristotelian world picture is already in the pipeline. The philosophical idea that the same laws may apply to both the sublunary and the celestial sphere has been aired so much that the scientist may now air the rational thesis that mathematics also applies to the planets – that it also goes for the book of the planets that it “è scritto in lingua matematica” (Galiee 1623: 600/238).

Secondly, science may run against barriers from which it can move on only if it manages to change reason. For instance, Bohr’s wave-particle experiment, which demonstrates that in some experiments the electron behaves like a wave and in others like a particle, would not have been possible to sustain and develop without a modification of reason. According to Popperian falsificationism a dilemma like this demands that science choose sides and decide which of the two experiments is a falsification of the theory of the atom. But there is a third possibility: that the reason that judges makes itself flexible in order that it may contain both positions. In this manner science develops reason by forcing it to surpass itself if it wants to move on. Reason constitutes an epistemological impediment, and it is precisely by ramming against this impediment with the power of empiricism that reason is given a chance to overcome the impediment or, rather, to demolish its own barrier. “En somme la science instruit la raison”, “science informs reason” (Bachelard 1940: 144/122). And yet. Bachelard does not refer to his theory as an empirisme abstrait, but as a rationalisme appliqué, applied rationalism:

[…] one of these two metaphysical directions needs to be given greater stress than the other; this is the one that moves from rationalism to experience. (Bachelard 1940: 6/6)

Thinking comes before seeing.

‘Applied’ here means this: what is the concept capable of when put to use? Rather than engaging in deep philosophical discussions about the a priori value of induction, one has to consider what induction may yield when applied. Mathematical physics represents an approximation of the two diverse areas, which “rationalises the real and realises geometry” (Bachelard 1928: 190, current translation).

Bachelard’s steps towards application, away from ideals, are taken simultaneously by Wittgenstein in 1936-37 as he makes the initial moves in preparation of a philosophy of language whose main theme seems to be: man is not supposed to live up to any absolute logic. Rather, logic is to be formed by human logic.

According to Umberto Eco, art may be viewed as a “metafora epistemologica” (Eco 1962: 42), i.e. as a reflection of the way in which the scientific epoch views reality – which is an idea Foucault demonstrates in The Order of Things (we shall return to this in the next chapter). In continuation of this idea it is worth noticing that while Bachelard celebrates the transgression of the visible and that which can be imagined as visible in surrationalismsurrealism flourishes in France.


Obstacles and Ruptures

It is therefore this striving towards rationality and towards construction that must engage the attention of epistemologists. We can see here what distinguishes the epistemologist’s calling from that of the historian of science. Historians of science have to take ideas as facts. Epistemologists have to take facts as ideas and place them within a system of thought. A fact that a whole era has misunderstood remains a fact in the historians’ eyes. For epistemologists however, it is an obstacle, a counter-thought. (Bachelard 1938: 17/27)


The Rupture with the Unscientific

An obstacle is the kind of resistance a scientific advance may encounter. For the moment we will lean on Bachelard who understands obstacles (obstacles épistémologiques) mainly as conceptual obstacles. Like Comte, Bachelard claims that the spirit of science goes through three stages of development: a pre-scientific stage, a scientific stage and a new scientific stage (Bachelard 1938: 7). In Bachelard’s schema, Comte’s scientific ideal belongs to the second stage. At the first stage, which corresponds to the theological and metaphysical phase in Comte, obstacles are characterised by imprecise language use. At this stage language is charged with metaphors and imagery and concepts need weeding out in order for science to progress.

Bachelard’s own works are as ambiguous as his point of view is clear: poetry and science do better without each other. Bachelard characterises such obstacles as external to science: Science has to shape itself as science by distinguishing itself from religion, poetry and speculation. It has to develop its own concepts.


The Rupture with Opinion

The first epistemological rupture in science is a rupture with opinion, with common sense.

Opinion thinks badly; it does not think but instead translates needs into knowledge. By referring to objects in terms of their use, it prevents itself from knowing them. Nothing can be founded on opinion: we must start by destroying it. (Bachelard 1938: 14/25)

The philosophy of no is a no to meaning.

As Bourdieu points out, the rupture with common sense is particularly hard in sociology as everyone seems to have an opinion of their object of study (Bourdieu et al.1968 35). But in order to practice sociology scientifically one has to break with the categorisations of opinion and construct other models with which to understand society.

scientific experiment is therefore one that contradicts ordinaryeveryday experience. (Bachelard 1938: 10/22)


The Rupture with Ordinary, Everyday Experience: Rationalism and Scientific Spirit

When a science has established itself, obstacles take the form of internal obstacles, i.e. obstacles are created by the products of science itself. Science bumps against its own conceptual ceiling. Just as Newton’s mechanics prevented us from thinking in terms of Einstein’s theory of relativity, diachronic philology prevented us from thinking in terms of synchronic linguistics. The concepts that science produces constitute a foundation and an obstacle at one and the same time. “On fonde en construisant”, through construction one founds, says Bachelard (Bachelard 1949: 121), and, it should be added: through construction one creates obstacles. As Regnault has it: “when thinking something, something else becomes unthinkable” (Regnault 1968: 80).

Bachelard wants to practice what he refers to as a psychoanalysis of the sciences, not to analyse the scientists, but to analyse their concepts of ‘mass’ and ‘energy’ (Bachelard 1938: 9). He establishes five different profiles of the relation between the two concepts as they appear throughout the historical development of science. Not only do the profiles mirror the development of the relation between mass and energy in the history of science, they also reflect the relation between the two concepts in Bachelard’s personal development from a child to a post office official to a philosopher of science. Bachelard shows how a person grows from a naive realistic concept of mass (mass is size, e.g. the belief that the biggest piece of cake weighs the most) to a positivist concept of mass (mass is what something weighs, e.g. the heaviest piece of cake weighs the most) to a rationalist concept of mass. With the rationalist concept of mass the elastic thread to the immediate experience of mass snaps: force is mass multiplied by its acceleration (Newton’s second law). This concept of mass is established on the basis of an epistemological rupture: it presupposes a mathematisation of the world.

How did Galilee and Newton arrive at their theories? Stories are often told of things falling from the tower in Pisa, pendulums swinging in churches and apples dropping from garden trees. But in practice these popular experiments have played no role. The decisive thing is not that Galilee could not possibly arrive at the definition of the free fall (distance = one half of uniform acceleration multiplied by the square of time) with the water clock that was at his disposal. To put it in another way, you cannot observe your way to the definition uniform acceleration, it has to be arrived at through an act of construction. You can assume a uniform acceleration and then proceed to measure the fall.

Even so, no one has ever seen a straight line; at Galilee’s time no one had seen uniform accelerations or frictionless slopes. Indeed, every single day scientific thought argues against ordinary, everyday experiences, like the way we see the sun as rising, fire as rising into the air, stones as falling, etc.

Inertial motion, indeed, not only has never been encountered in the world, but it is even impossible that it ever should be […] The infinity of the universe cannot, of course, be asserted on the ground of experience. (Koyré 1954: 261/108)

In the 17th – 19th centuries physicians did not feel their way forward through empiricism but through theory. When Galilee wanted to demonstrate the curve of the uniform movement of a pendulum, he used mathematical arguments. And later when Huygens corrected him regarding the form of the pendulum movement this also happened in mathematics.

To epistemologists yet another point is to be added here: scientific experiments are instrumentalised and, as such, they break with simple sense perception. As Duhem puts it:

Between sensory knowledge and scientific knowledge there is a gap. Temperature is seen on a thermometer, one does not feel it. Without theory, one would never know whether what is seen and what one feels correspond to the same phenomenon. (quoted in Bachelard 1940: 10/9)


The Rupture with the Narrative of Science

Things are not made any easier by the fact that the translators of the Greek, Latin and Italian texts are so entrenched in the positivist self-image of science that they mistranslate the originals. Thus, Copernicus’ orbium celestium, celestial orbits, is translated as celestial bodies, turning Copernicus into a Newtonian avant la lettre. Galilee’s theory-based concept of comperio is translated as “discovered by experiment”, turning him into the father of the experiment, which is how posterity likes to see him. As Koyré notes, it is no wonder that this legend is sustained in America in view of the fact that historians always cite Galilee in the English translation (Koyré 1943: 272-274). Epistemology is not only a philosophy of historical ruptures, it is also a philosophy that breaks with the narrative of science, i.e. the narrative of the eye and the thing in which thought always comes last, systematising things.


The Rupture with Earlier Science: ‘-non’: The New Spirit of Science

The rational concept of mass is in itself an obstacle to further progress, or rather, in Bachelardian terms: Classical rationalism (Galilee-Newton) amounts to “un tissu d’erreurs”, a tissue of errors preventing us from moving on (Bachelard 1940: 8/8). Science can move on only by means of an epistemological rupture (coupure): the view of mass as independent of time and movement has to be done away with. And what is just as important to Bachelard is that reason is not left untouched in the process. It has matured even more, it has said no, it has become dialectic.

Science says no by encircling what has been rejected: What is rejected is turned into a special case by the new theory, which is expressed by a ‘non’: non-Euclidian geometry, non-Newtonian physics, non-Kantian philosophy, non-Maxwellian physics, non-Lavoisierian chemistry. The last two levels – the goodbye to the visible and that which can be imagined as visible and the encircling ‘non’ – are referred to by Bachelard as surrationalism, or ‘the new scientific spirit’ (Bachelard 1934 and Bachelard 1938: 7).


The Rupture with Continuity

When science bumps into itself, it has to break its own laws. This is not a break with metaphysics in every respect, which is the kind of break only positivists hope to make.

The mind may change its metaphysics, it cannot do without metaphysics. (Bachelard 1940: 13/11)

It is not a break with darkness either, but with light; a historical position is not darkness, but light that has been broken in a certain way.

Bachelard finds it hard to understand why his contemporaries do not understand that it was once impossible to understand what we understand today; that it was once extremely difficult to see what we perceive as perfectly clear today, and that the ease of today’s comprehension is but at the end of a long path cleared through hard struggles of cognition:

I have often been struck by the fact that science teachers, even more than other teachers if this is at all possible, cannot understand that their pupils may not understand. (Bachelard 1938: 18/28)

Bachelard is annoyed by three typical pedagogical ploys: that the complex is explained from the perspective of the simple, that the world is explained on the basis of things and that insight is explained as light that replaces darkness. In the latter case, people fail to understand that “ignorance is a tissue of positive, tenacious, interdependent errors”, “errors are not easily destroyed one by one. They are coordinated” (Bachelard 1940: 8/8). This may also be expressed the other way round, as Canguilhem does: One can only make a mistake from a position of being within the true (Foucault 1970: 36). Making a mistake involves a movement within the domain of the given that claims about the non-given that it is given, or vice versa. Any idea that creates a rupture with the continuity of thought ruptures with the domain of the given entirely. To the extent that a rupturing idea can be said to be an activity that transforms the epistemological domain, one may speak of it as a rupture with the prevailing system of thought.

Does history happen by ruptures? It may be interpreted that way (Foucault 1966: 14/xxiv-xxv), but it would be more correct to read it as rearrangements or a series of several small ruptures (which we will return to in more detail in the next chapter): incipient ruptures accumulate before a rupture which never happens without great pains or forerunners. Besides, ruptures may very well happen rather undramatically since new concepts may be hiding in the same words. Words and concepts are not the same, however. Insofar as a rupture (coupure) does not consist in an epoch-making idea but in gradual ruptures (ruptures), we may speak of ruptures as small re-ruptures or as rearrangements.

Foucault thinks of the rupture with immediate experience and opinion in a slightly different way. Firstly, this is due to the fact that he does not relate to science in any prescriptive manner in regard to the question of what proper science is supposed to be. He works descriptively, concerning himself with different natures of science across time. Secondly, because he understands the sciences as inextricably inscribed in subjacent principles of ordering, epistemes, which govern opinion, experience and the sciences. At the same time it must be noted that Foucault insists on breaking with all the common ideas of when the decisive events in history occurred. A Foucauldian approach typically starts by breaking down the commonly-held opinions about epochs by showing how consistent problems continue across different periods in time which are commonly seen as separate epochs. Subsequently, Foucault delves into these problems in order to distil various complexities within them, which then gives rise to new historical lines of rupture. This is the case for instance when Foucault criticises Nietzsche’s thesis about Antiquity as a pre-moral condition by calling attention to Greek morality and proceeding to show how we may establish different ruptures or rearrangements by tracing the development of this morality.

The rupture with immediate experience and the rupture with opinion is a negation. The rupture with the experience of the visible and the imagined and the subsequent rupture with opinion by means of a ‘non-’ are dialectical ruptures. Not in the Hegelian sense, Bachelard notes, as that is an a priori dialectics presupposing an autonomous driving force on behalf of history. Bachelard is in league with the Russian mathematician, Nikolaj Lobatjevskij, who in the 1830s, in concurrence with Jànos Bolyai but independently so, showed the possibility of constructing a geometry that dispenses with the need for the Euclidian parallel postulate – what was later to be known as hyperbolic or non-Euclidian geometry. Lobatjevskij refers to the transgression of Euclidian geometry as a dialectisisation and it is in this sense that Bachelard employs the concept.

Bachelard’s dissociation from Hegelian dialectics is connected with his dissociation from any necessity in history whatsoever. Bachelard sees no development as necessary. Development depends entirely on what the eye allows us to see and what thought allows itself to be corrected by. To agree with Bachelard that this is contrary to Hegel probably requires that one does not think Hegel with Hegel, as I see it.

Epistemology is Kantianism raised to the second power, the way it is also to be found in German neo-Kantianism, in e.g. Ernst Cassirer. It is a rationalist critique of rationalism which raises rationalism above itself. It does not reject rationalism, it transgresses it. Even mathematical truths presuppose thought as a construction (Cavailles 1941: 271-274).

Epistemology is surkantianism.


The Rupture with Agreement and Repetition

Epistemology is the philosophy of no: no to continuity. A dialectic no, but not a necessary no. The no of rationalism wherever empiricism erects its constructions. A definite no that builds on a new metaphysical foundation. It is both a descriptive and a prescriptive no. The no is progress:

If two men want to agree they really have to contradict one another first. Truth is the daughter of discussion, not of sympathy. (Bachelard 1940: 134/114)

Bachelard cuts out Popper’s subsequent template of knowledge, the idea that science has to try to disconfirm the known in the real. At the same time, he follows a line in German thinking on the problem of experience from Gadamer back to Hegel which claims that experiences are only experiences if they break with previous experiences.

The normative template of epistemology goes like this: for the sake of progress, you must cause ideas to rupture. In Foucault you do not find any normative instructions for the sciences as concerns notions of proper scientific character. To Foucault, any scientific progress, if possible, is about whether new sciences have managed to map nature, man or society in any greater detail than before, and what role recently acquired characters of judgment play in shaping the reality that they judge. To Foucault scientific progress is about the enabling of new forms of governance.


The Rupture with Images?

But what it is the no to images about? Bachelard’s oeuvre is consistently split in two. One half, from the end of the 1920s till the beginning of the 1950s, is about concepts; this is epistemology. The other half, from the publication of La Psychanalyse du feu in 1938 to La Flamme d’une chandelle from 1961, is about poetry and the creation of images. In science you are allowed to smash up images:

[T]hose who would be constructors of mathematical thought must first be iconoclasts. (Bachelard 1938: 77/85)

Nevertheless Bachelard thinks that the creation of images is just as important in human life as what psychologists refer to as a sense of reality:

A person deprived of the function of the unreal is just as neurotic as the person deprived of the reality function. (Bachelard 1943: 14/ 7)

Why strike such a hard blow at the images in science? Bachelard does not want to deny that images may have a creative effect in the fostering of thought, but they quickly coagulate as conceptual Trojan horses, the original context of the image pouring out from the abstract metaphor as a consequence. Concepts like ‘force’, ‘charge’, ‘work’, ‘mass’, ‘switching on and off’ all ooze of old-fashioned functionality, which confuses science. You do not switch electric power on and off. Metaphors smuggle realism into science. It is not enough to call them in for a strip search. As with concepts, they have to be expelled at the border of science.

Bachelard launches an attack on the idea that the simple is the foundation of the complex and, accordingly, that it is possible to understand the complex on the basis of a substantialist model. Science needs abstractions, but images are charged with realist and rational ways of thinking that seize on the concepts.

The poet and the scientist both run against verbal obstacles, words being weighed down by meaning which is neither poetic nor scientific but constituted in everyday realism. Realism may be demolished only by keeping poetry and science apart.

Bachelard is quite alone in that view, however. Koyré emphasises the creative element in the creation of images in science:

As far as I am concerned, I do not think that we should speak too ill of images. […] scientific imagination – or intuition – manages to produce such beautiful images and (we see this over and over every day) deeply penetrate into regions – such as for instance the atom and its nucleus – which it at first appeared to be completely cut off from reaching (Koyré 1954: 255, present translation)

Canguilhem engages with this perspective in a more concrete way inasmuch as he strives to determine the role played by the development of analogies and models in the development of biology (Canguilhem 1961).

In Antiquity philosophers worked with a reciprocal description of the body as mechanism and mechanism as body, which we still use, e.g. in expressions such as the valves of the heart and a toothed wheel. Plato speaks of vertebrae as pivots (Plato Timaeus 74a) and Aristotle speaks of the forearm as a catapult. The interesting thing is not only how deeply embedded these images are, but also how they engender a certain methodology: From Antiquity to Descartes bodies are thought of in terms of mechanism rather than function. The difference is that mechanism is incapable of reacting within a game.

With James Watt’s development of conductors and the development of the battery and the induction coil, a certain knowledge was constructed which was later to form a model for biology: the reflex. The reflex formed a model for electro-biology; nerves, for instance, were conceived of as conductors. The analogy of electricity and conductors had several advantages: Nerve fluid is not visible like blood which is why electricity was a useful model. In the beginning electricity was not used for the transport of energy, but for the transmission of messages, which also went well with biology.

After the development in mathematics of cybernetics James Watson & Francis

Crick succeeded in 1953 to translate communication theory into a model for biology (Crick & Watson 1953). The model was convenient as it made it possible to explain the regeneration of the brain, which could not be explained through the metaphors of mechanics or electricity.

The reason for the difference between Bachelard’s iconoclastic approach to metaphors and Canguilhem’s more development-oriented approach is not only to be sought in the fact that Bachelard compensates through an entire parallel work on the creation of images. The explanation is probably also to be sought in the difference between the very objects of the two epistemologies: So far biology has shown itself less mathematisable than physics and has followed the course of the latter, which is why the images in physics have been decisive in the development of biology.

In physics mathematics is employed to work out measurements. In biology mathematics do not only play a part as an instrument, but also as a model. The distance between biology and mathematics causes mathematics to remain a metaphor or an analogy in biology. The purpose of the mathematical model is to explain biology as mechanism, physics, chemistry, mathematics or communication. However, as it seems hard to conceive biological elements without a model, illustration (image) and representation (theory) start to mix little by little. The model hypostatises in biology. Yet, Canguilhem asks, may it not sometimes be productive for a science if its object (objet théorique) merges into a simplified image?

But has it not sometimes happened that the analogical models of the biologist have benefited from an unconscious validation having as its effect the reduction of the organic to its analogy mechanical, physical or chemical? Despite their great degree of mathematical complexity, it does not appear that cybernetic models are always safe from this accident. The magical aspect of simulation is strongly resistant to the exorcism of science. (Canguilhem 1961: 514-515)


The Rupture with Substance

However, because of their nature of being visible, images are sometimes hard to drive out from science; the connection between the visible and the power of persuasion is not to be underrated. Liberalism’s Newtonian image of the world as a quantity of independent bodies in motion, and, accordingly, of freedom as the freedom of movement, is a persuasive image because of its lucidity. Thinking based on substance appears as “a tissue of positive, tenacious, interdependent errors”, and that is what makes it hard for the social sciences to push through with insights based on relations rather than substances. Substance – in this case the individual  – is confirmed whenever consciousness slackens.

In Foucault this tissue of tenaciously interdependent truths, which are later referred to as mistakes, corresponds to the time-limited episteme, i.e. the silent order behind the sayable and the visible. When Linné saw the historical evolution of animal species, for example, he saw variations of the essence of the animal in question. Linné was able to think in terms of development, but not as a founding principle because he was situated within an episteme different from Darwin’s, which conceived of order in an ahistorical manner.

In Althusser’s version the tissue of tenacious, interdependent errors corresponds to ideology. The rupture is a Marxist break away from ideology and into science.

The idea that society is not the sum of individuals, but of relations was pointed out already by Marx. Bachelard entertains the same preferences for relations. His argument is that a phenomenon cannot be understood without its context.

[T]here is no such thing as a simple idea, for […] no idea can be understood until it has been incorporated into a complex system of thoughts and experiences. (Bachelard 1934: 152/ 148)

It is a typical substantialist error to pin one’s theory on pure substance which is then assumed to carry a complete whole of attributes. Observations lead phenomena back to a body, e.g. a photon is supposed to have been emitted by a small body called the electron. But we are always mistaken when we start with the simple, we have to start with the whole picture, because “true science will never be a science of objects, but a science of relations.” (Bachelard 1928: 182, current translation).

It is on this background, drawn by Marx and Bachelard, that Bourdieu establishes his sociology of relations in which any phenomenon must be understood at all times on the basis of the relations and struggles between relations as they come across in the field of study and in the social space as a whole. With a link to Marx, the relational positions in life become struggles for recognition. Sociology as a science about relations is haunted by the ghost of Hegel.

Foucault is struggling with exactly the same problematic when he introduces a concept of power which has no substance and does not belong to any substance, but is 1) the expression of a complex strategic situation in a society (Foucault 1976: 123/93) and 2) is always taking place in actions as ingrained in relations which are never simply power relations (Foucault 1981: 235). One has to break away from the idea of the social as a thing.


The Immersion of Epistemology in Culture

Practical and Epistemological Possibilities

With Bachelard and Koyré epistemology has yet to be immersed in culture. When asked to explain what an epistemological obstacle is, they will never leave it at a sociological or socio-economic explanation. For example, when Koyré is to explain why there were so few machines in Antiquity, he claims that explanations are not to be sought in the historical of Antiquity circumstances, but in the scientific concepts at that time: techne and episteme are separate entities (Koyré 1948B: 319-323). The development of technical knowledge is a modern thought. Engineers in the age of Antiquity were hardly engineers, it would be more correct to say that they were architects. They could make artificial birds and automatic door-openers, but not a rudder. The Romans took over Greek ingenuity, but without developing it. Facing a problem of water supply, they opted for the architectural solution: an aqueduct. In the Renaissance the same problem brought about the development of the water pump (Koyré 1948B 335/footnote 1). A precondition for any development is the will to development.

In short, one might say that the reason why the Antique world did not develop mechanisation and made no technical progress in general was that such things were felt to be of no significance whatsoever. And the reason why the modern world did so was that such things seemed, on the contrary, to be of the greatest importance. (Koyré 1948B: 323, present translation).

Does this mean that the fact that people think something is important is a decisive factor for development? Yes it does. The practical possibility was actually present all the time, but the possibility was not epistemically present: it was not thinkable the way it is to us. Or rather: the possibility was not realised.

We may refine Bachelard’s concept by introducing a distinction between two binary oppositions employed by Deleuze: the possible is that which is yet not real. Yet, the possible is that which is actualised (which is not virtual). Possibilities are real possibilities, virtualities are possible realities. Virtualities may be actualised as real possibilities. The movement Koyré refers to as mathematisation and Bachelard refers to as realisation is what Deleuze refers to as actualisation. It has nothing to do with an exploitation of the opportunities at hand, but has everything to do with the introduction of a new reality (Deleuze 1966: 99-100). The rudder was a virtuality to the Greeks of Antiquity, but not an actual possibility.

To the extent that epistemology is immersed in culture in Bachelard and Koyré, this only applies to a research community. To be part of a modern research community is to have broken away from everyday thinking and to have embedded oneself in a new scientific world picture. This is already an old world picture when the researcher enters it and he gets just as old himself:

Even when it first approaches scientific knowledge, the mind is never young. It is very old, in fact, as old as its prejudices. When we enter the realms of science, we grow younger in mind and spirit and we submit to a sudden mutation that must contradict the past. (Bachelard 1938: 14/24-25)


Ideology and Cultural Horizon

Dragging a cultural baggage along into the laboratory and the hospital ward is the starting point precisely for Canguilhem’s criticism of the concept of norms in science: medical science is ideological. Ideology is that which is already old: all that has been thought before, the given, prevailing opinions and ideas. The images that have to be smashed up are the images of ideology, the objects of ideology. The theories that allow themselves to be developed within an ideological practice are objects of ideology. In Althusser, the first break consists in arriving at a cognition of the fact that there is a difference between the object of a theory (objet théorique / objets de la connaisance) and the real object (objet réel). The break elucidates how the truth of theory does not originate directly in the thing itself. The second break consists in abandoning the images that ideology makes of things (Althusser 1965: 196 and Althusser & Balibar 1965: 46-50 and 71).

In Althusser the Bachelardian culprit, everyday realism, is converted to ideology, but he still operates with everyday experience the way Bachelard does. Science is a break away from everyday experience.

Foucault rejects the idea of such a break between everyday experience and science, among other things because he makes a bid for inscribing the history of experience into the history of problems. Yet he acknowledges the distinction between a real object and a theoretical object, and the reflexive exercises in L’archéologie du savoir all aim at illustrating how we may carry out the exposure of the rules of formation that generate the so-called discursive practice within which the theoretical object is produced. It goes for these rules of formation that:

As a group of rules for a discursive practice, the system formation is not a stranger to time. (Foucault 1969: 98/74)

To Foucault it is not crucial to distinguish between knowledge and science the way it is to

Althusser and Bachelard. As far as normativity is concerned, Foucault’s project is not about prescribing an ‘ought to’ for the sciences, it is about the influence of science on normativity.

Science is part of he total cultural knowledge and normativity it forms a part of:

 it assumes a specific form and function according to the field and the level to which it is assigned (Foucault 1969: 17/9)

 Foucault is interested in science, surely, but primarily because he wants to examine this “champ de savoir”, which he defines as an “episteme” in The Order of Things (Foucault 1966: 13/xxiii). According to Foucault, the rules of formation that generate the discourse of a science, and in that way cause enunciations (énoncés) to be given as something given (positif) and become knowledge (savoir), constitute the fundamental codes of a culture,

[…] those governing its language, its schemas of perception, its exchanges, its techniques, its values, the hierarchy of its practices – [they] establish for every man, from the very first, the empirical orders with which he will be dealing and within which he will be at home. (Foucault 1966: 11/xxii)

Things are ordered by these codes. To Foucault the thinking in seeing is the cultural thought within the cultural gaze. It is exactly these codes that he proposes to examine in The Order of Things.








1970: L’ordre du discours, Paris: Gallimard 1971.

1971: ”Nietzsche, la généalogie, l’histoire” in Dits et écrits 1954-1984, Paris: Gallimard, 1994; II, pp.136-156.

/ “Nietzsche, Genealogy, History” in James D. Faubion (ed.): Michel Foucault. Aesthetics. Essential Works of Foucault 1954-1984. Volume 2, London: Penguin, 2000

1966: Les mots et les choses, Paris: Gallimard 1966
The Order of Things, London and New York: Routledge Classics, 2002

1969: L’archéologie du savoir, Paris: Gallimard 1969.
The Archaeology of Knowledge, London: Routledge, 1992

1975: Surveiller et punir. Naissance de la prison, Paris: Gallimard 1975
/ Discipline and Punish. The Birth of the Prison, London: Penguin, 1991.

1976: Histoire de la sexualité 1, La volonté de savoir, Paris: Gallimard 1976
The Will to Knowledge. The History of Sexuality: Volume One, London: Penguin, 1998

1981 ”Le sujet et le pouvoir” in Dits et écrits 1954-84, Paris: Gallimard, IV pp.222-243.



Althusser, Louis 1965: Pour Marx, Paris: Maspero 1977

Althusser, Louis & Étienne Balibar 1965: Lire le Capital, Paris: Maspero 1975.

Aristoteles: Metaphysics, London: Heinemann 1935.

Bachelard, Gaston 1928: Essai sur la connaissance approchée, Paris: Vrin 1928

1934: Le nouvel esprit scientifique, Paris: Quadrige/Presses Universitaires de France 1999
/ Bachelard 1934 / The New Scientific Spirit, Boston: Beacon Press 1984.

Bachelard, Gaston 1938: La formation de l’esprit scientifique, Vrin: Paris 1993.
The Formation of the Scientific Mind, Manchester: Clinamen Press 2002.

Bachelard, Gaston 1940: La philosophie du non, Paris: Quadrige/Presses Universitaires de France 1994
The Philosophy of No, New York: The Orion Press, 1968

Bachelard, Gaston 1943: L’air et les songes, Paris: José Corti, 1943.

Bachelard, Gaston 1949: Le rationalisme appliqué, Paris: Presses Universitaires de France 1949.

Bachelard, Gaston 1972: L’engagement rationaliste, Paris: Presses Universitaires de France 1972.

Bourdieu, Pierre, J.-C. Chamboderon og J.-C.Passeron 1968: Le métier du sociologue, Berlin: Mouton 1968.

Canguilhem, Georges 1961 : 1961: “The role of analogies and models in biological discovery” in A.C.Crombie (red.): Scientific Change, Historical studies in the intellectual, social and technical conditions for scientific discovery and technical invention, from antiquity to the present. London: Heinemann 1963; pp.507-521.

Canguilhem,  Georges 1966: ”L’objet de l’Histoire des sciences” in Études d’histoire et de philosophie des sciences, Paris: Vrin 1968 ; pp.10-23
/ “The Object of the History of Sciences” in Gary Cutting (ed.): Continental Philosophy of Science, Oxford: Blackwell 2005; pp. 198-207

Cavaillès, Jean 1941: Philosophie mathématique, Paris: Hermann 1962.

Comte, Auguste 1844: Cours de philosophie positive, Paris: Hermann 1998.

Crick og Watson 1953 : “Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid” in Nature 1953 april 25; 171(4356). pp-737-738.

Descartes 1650: Regulæ ad directionem ingenii in Oeuvres de Descartes (Adam/Tannery), Paris: Vrin 1996 (1701)
/ Descartes 1650 / Descartes – Key Philosophical Writings, Hertfordshire: Wordsworth Editions Limited 1997

Deleuze, Gilles 1966: Le bergsonisme, Paris: Presses Universitaires de France 1998 (1966).

Duhem, Pierre 1906: La théorie physique, Paris: Chevalier & Rivière 1906.

Eco, Umberto 1962: Opera aperta, Milano: Bompiani 1962.

Fogh Jensen, Anders 2001: Metaforens magt. Fantasiens fostre og fornuftens fødsler, Modtryk: Århus 2001.

Galilei, Galileo 1623: Il saggiatore in Opere II, Milano: Rizzoli & C. Editori 1938.
/ Discoveries and Opinions of Galileo, New York: Doubleday Anchor Books, 1957.

Hegel, G.W.F. 1807: Phänomenologie des Geistes, Hamburg: Felix Meiner Verlag 1988

Koyré, Alexandre 1943 : ”Traddutore – traditore” Études d’histoire de la pensée scientifique, Paris: Gallimard 1973 (1966) ; pp.272-274.

Koyré, Alexandre 1948A: ”Du monde de l’ « à-peu-près » à l’univers de la précision” in Études d’histoire de la pensée philsophique, Paris, Gallimard 1971 (1961) ; pp.341-362.

Koyré, Alexandre 1948B: ”Les philosophes et la machine”in Études d’histoire de la pensée philosophiqueÉtudes d’histoire de la pensée philsophique, Paris: Gallimard 1971 (1961) / “Filosofferne og maskinen” in Tankens enhed, København: Hans Reitzels Forlag 1998; pp.96-130.

Koyre, Alexandre 1954 ”De l’influence des conceptions philosophiques sur l’évolution des théories scientifiques in Études d’histoire de la pensée philsophique, Paris, Gallimard 1971 (1961); pp.253-269
/ “Influence of Philosophic Trends on the Formulation of Scientific Theories” in The Scientific Monthly, Vol. 80, No. 2 (Feb., 1955), pp. 107-111.

Meyerson, Émile 1908: Identité et réalité, Paris: Vrin 2001.

Pascal, Blaise 1660: Pensées, Paris 1976.

Pedersen, Olaf og Helge Kragh: Fra kaos til kosmos. Verdensbillets historie gennem 3000 år, København: Gyldendalske boghandel, Nordisk Forlag 2000.

Popper 1953:”Science: Conjectures and Refutations” in Conjectures and refutations: the growth of scientific knowledge, 4. ed., London, Routledge and Kegan Paul 1976.

Regnault, François 1968: ”Hvad er et epistemologisk brud” in Olesen (red.): Epistemologi, København: Rhodos 1983; pp.78-84.

Quine, William van Orman 1960: Word and Object, Cambridge: MIT Press 1960.

Saussure, Ferdinand de 1916 : Cours de linguistique générale, Paris: Payot & Rivages 1995
/ Course in General Linguistics, Chicago: Open Court Publishing, 1986