The Encylopaedia of Educational Philosophy and Theory

Back to Contents

Back to Top

Table of Contents

Georges Canguilhem

James Marshall

University of Auckland

Georges Canguilhem was born on 4 June, 1904, at Castelnaudary near Toulouse, in South Western France. He died on 11 September 1995 at the age of 91.

Success at Castelnaudary Lycée (where he was a boarder), and the award of a scholarship to study at the prestigious Lycée Henri IV in Paris, enabled him to gain entrance to the École Normale Supérieure in 1924. In his cohort were Jean-Paul Sartre, Raymond Aron, and Paul Nizan. Later there were Maurice Merleau-Ponty and Jean Cavaillès. He was influenced by his teacher Alain (Emile Chartier) at Henri IV, by Cavaillès whom he succeeded at the University of Strasbourg in 1941, and by Gaston Bachelard, whom he succeeded as Professor of History and Philosophy of Sciences at the Sorbonne, and as Director of the Institut d’Histoire des Sciences and Techniques, in 1955. In addition to his qualifications in philosophy he gained a Doctorate in Medicine at the University of Strasbourg in 1943. His doctoral thesis was to be republished several times, and translated into English as The Normal and the Pathological in 1978 (with an introduction by Michel Foucault). He also occupied the important administrative post of Inspecteur Général de Philosophie between 1948 and 1955, having initially refused it at the time of the Liberation. In this post he was responsible for the teaching of philosophy in lycées. His writing was austere and he was noted as an exacting if not intimidating examiner. Nor was he the typical French intellectual, pronouncing on almost anything and prepared to occupy a radical political platform. If he was on the left he was not on the radical or revolutionary left. For full bibliographical details and accompanying biographic comments see Camille Limoges (1994). For a fuller historic account in English of Canguilhem’s life see David Macey (1998).

His friend Jean Cavaillès, who was to be assassinated by the Nazis in January 1944, had encouraged him into the resistance (if encouragement were needed – given his ‘rebellious’ stances at ENS and his early opposition to events in Germany). After the Gestapo raid on the Faculté des Lettres at the University of Strasbourg (by then in Clermont-Ferrand) in 1943, and in which two professors were killed, and many students and professors deported to Germany (but which he managed to evade), he was forced underground where he took a major part as a doctor in the Auvergne Maquis (code name ‘Lafont’). He was awarded the Military Cross and the Médaille de la Résistance in 1944. Later, in 1976, he was to publish a study of his former student, colleague and co-member of the resistance Jean Cavaillès: Vie et Mort de Jean Cavaillès.

Unlike Bachelard, who took physics and chemistry as historical examples of scientific rationality, he took as his major sources, biology and medicine. It could be said that in selecting biology and medicine, and in rejecting great scientific events such as the Copernican Revolution, that he forged a change of course in French History of Science. Biology and medicine were not as rigorous as physics and chemistry and are inextricably intertwined with non-discursive practices. Foucault is to extend this displacement further to the human sciences.

If Canguilhelm was an historian of science rather than a philosopher of science then he was also an historian who was extremely sensitive to philosophical issues. According to Dominique Lecourt (1975: 165f.): “There is probably no better definition of the history of the sciences as it is conceived and practised by Georges Canguilhem himself…it seems completely justified to make him Bachelard’s heir”. But his history is also epistemological. For Canguilhem , “the history of science is the history of an object – discourse – that is a history and has a history, whereas science is the science of an object that is not a history, that has no history”( Canguilhem in Delaporte, 1994: 26). Thus (ibid.):

…the object of the history of science has nothing in common with the object of science. The scientific object, constituted by methodological discourse, is secondary to, although not derived from, the initial natural object…The history of science applies itself to these secondary, non-natural, cultural objects. It is a discursive project about scientific objects’. But it is also concerned with “the progress of the discursive project”, a progress which may “meet with accidents, be delayed or diverted by obstacles, or be interrupted by crises, that is moments of judgement and truth

The objects of the history of science are then very different from the objects of science. For Canguilhem science arrests time, construing its objects as non-temporal, and as not having a history. The full reality of the scientific object is in principle available to the scientist in the present. It is of course true that these objects exist in time and change through time. However the objects of history of science are regarded themselves as part of an historical development which has not yet finished. The objects of history of science are incomplete. Whereas the objects of geology can be treated as complete, as “givens” open to analysis, the objects of history of science cannot, as their value and meaning are determined first by an epistemological and normative judgement and, second, are always in principle open to re-evaluation as to their value and meaning in accordance with the progress of science.

Canguilhem's epistemological concern then is with the history of concepts. The philosopher’s aim is to identify “the order of conceptual progress that is visible only after the fact and of which the present notion of scientific truth is the provisional point of culmination” (Canguilhem, 1988: 9). But this history of concepts is not the history of ideas. Nor is it a history of terms, or of phenomena, or even of theories. Perhaps the elimination of these possibilities can make clearer his view of the history of science.

If there were such a thing as a history of terms then it might concern itself with exploring the use of a term from its historical antecedents to more modern usages. For example, the term ‘atom’ in current use in modern science was also employed by the pre-Socratics. However whilst the same term or word was used by both pre-Socratics and contemporary scientists, the referents of these terms have almost nothing in common. Such a history was not Canguilhem’s.

Nor is the history of concepts to be identified with a history of phenomena. For example someone might produce a history of volcanic eruptions and earthquakes in, say, New Zealand. What would be crucial here for Canguilhem is not what is observed - the phenomena - but the interpretation of those phenomena. It is not enough then for someone to observe and describe phenomena no matter how new, or unexpected (perhaps as the result of a Baconian experiment), for that person to have a scientific concept, or to be doing science, or to be writing the history of science. One cannot explain in observational terms (thus for Canguilhem Priestly did not have a concept of oxygen whereas Lavoisiier did – Gutting 1989).

But Canguilhem also insists upon the separation of concepts from any theories which may “use” those concepts. Concepts are not imbedded in theories, and they do not derive their meaning from associated theories. Instead concepts permit one to identify data in a scientifically meaningful and useful manner: theories explain the data and/or phenomena identified prior to explanation by concepts. Concepts permit scientific questions to be formulated and theory(ies) provide(s) scientific answers to those questions. Concepts are also claimed to be “theoretically polyvalent” (Canguilhem, 1988: 6). By this is meant that one and the same concept can occur in different theories. Thus Canguilhem was able to write the history of the reflex arc, a concept which occurs in several quite different theories. This is not to deny that a concept may become reformulated and transformed between theories, but if the concept retains an underlying fundamental scientific content, it is still the same concept.

Canguilhem believes that there is a close relationship between concepts and phenomena. If he is rejecting the distinction sometimes claimed between (neutral) facts and theories, he is not doing so in any simplistic fashion which claims that there are no observed facts apart from their theoretical interpretations. His position is more complex as he makes distinctions between terms, concepts and theories. Concepts which are theoretically polyvalent identify phenomena, not necessarily or merely theories. Instead theories explain those phenomena identified prior to explanation by concepts.

Clearly Canguilhem does not view the history of science as itself possessing scientific status. Such a view of the history of science would see it in positivistic manner as presenting already constituted objects from the past of science to be scrutinised by the historian of science, just like any other data in a laboratory. What is wrong with this approach is Canguilhem's insistence that history of science is normative. Here he would seem to be following Bachelard in believing that the historian's judgements of the past are informed by the present. This involves a form of epistemological analysis which furnishes to the historian the principles for informed judgement of the past. Clearly there are normative notions associated with the evaluation of science's achievements and progress.

On the Normal and the Pathological is the work for which Canguilhem is best known. Not only was it important in the area of medicine but it was important for other areas of the human sciences. As Foucault said in the Introduction to the English version this work was important “for those very people who were separated from, or challenged, the establishment”. It was the work of Cavaillès, Bachelard, and Canguilhem on “a philosophy of knowledge, of rationality and of concept” which was important in the crises of the Universities and the status and role of knowledge in the 1960s, rather than “a philosophy of experience, of sense, and of subject”, ie, of Sartre and Merleau-Ponty, according to Foucault (Canguilhem, 1978: ix-x.). If that is the case however, for Canguilhem, the events of 1968 were not easy or comfortable, as he had given his life to the academy and may have been seen as a mandarin (Bourdieu, 1998: 191).

This work was essentially his medical doctoral thesis. He attacked the notion that the concepts of the normal and the pathological, so essential to the thought and activity of medicine, could be interpreted in a straightforward positivistic and statistical manner. He attacked the fundamental notion that normal was a statistical mean, because that amounted to conceiving and treating a living system as if it were structured and therefore governed in a law like manner. If that were the case it would have been in some pre-established harmony with the environment. Instead, Canguilhem argues, the human organism is a living vital organism which is by no means in any pre-established harmony with its environment, for “The laws of Galilean or Cartesian mechanics cannot by themselves explain the origin of coordinated organ systems, and such coordinated systems are precisely what one means by ‘life’. In other words, mechanism is a theory that tells us how machines (living or not) work once they are built, but it tells us nothing about how to build them” (Canguilhem in Delaporte, 1994: 78).

Thus for Canguilhem the normal begins instead with the living organism and an order of specific properties, arguing that medical practice must be based upon the diversity of life which in turn provides the paths for its own conceptualisation and for the restoration of its normal state. “To say that ‘no doctor proposes to produce a new kind of man, with a new arrangement of eyes or limbs’, is to recognise that an organism’s norm of life is furnished by the organism itself, contained in its existence” (Canguilhem, 1978: 159). Therefore we must proceed from life to understanding and not from understanding to life, and thus define life as a meaning inscribed in matter. Lecourt (1975: 184) translates this position into the form of an equation: < life = code = information = concept of life = concept >.

Essentially for Canguihem then normality means the ability to adapt to changing circumstances, to environments which are both various and variant. It thus involves activity and flexibility so that the living being lives in shifting relationships with a continuously changing environment. Medical dictionaries define the normal as “that which conforms to the rule, regular”. Canguilhem extends this brief definition as follows: ”(1) normal is that which is such that it ought to be; (2) normal, in the most usual sense of the word, is that which is met in the majority of cases of a determined kind, or that which constitutes either the average or standard of a measurable characteristic” (Canguilhem, 1978: 69). For Canguilhem and his views on life and concept, there can be no such sense of a pathological normal for living organisms and hence there can be no purely objective pathology (Rabinow, in Delaporte, 1994: 16).

Canguilhem’s thesis on life is known as vitalism. Gordon (1998: 185) states the thesis thus:

  • life is an irreducible concept and one which is necessary to science;
  • its content is given through experience as living beings as well as our observation of living beings;
  • our conceptual activity in general is a continuation and extension of our existence as living beings.

Also, for Canguilhem (1952: 143), machines are seen as an extension of living organisms: “Un outil, une machine ce sont des organs, et des organs sont des outils ou des machines”. His point is not that tools and machines are organisms but that they are extensions of the body (Hacking, 1998: 207). His work here, directed as much against Cartesian dualism, has obvious implications for the philosophy of technology.

In education there are at least three broad parameters along which a Canguilhem inspired research programme might proceed. First there is the importance for the human sciences, including education, of Canguihem’s approach to epistemology, especially his emphasis that this must be an historical epistemology. If this is being traced in the general area of the social sciences in Anglo-American thought (see eg, Economy and Society, 27 [2&3], 1998) it is almost non-existent in education (though see Marshall, 1996: 47-53). What would be of concern here is the specific nature of rationality in education, especially those aspects of educational theory and research which laid claim to scientific status, and the role which critical historical thought might play in relation to ‘the’ vital form of life displayed in a living organism. Second there is a need for a deeper exploration of the notion of ‘normal’ in educational thought and theory. Canguilhem’s notion of the norm as not being statistical but, instead, to be associated with normativity, that is the ability of a living organism to adapt with activity and flexibility to changing circumstances would be more than helpful here. Finally his views on vitalism, normativity and the notion that tools and machines are extensions of living organisms have interesting possibilities for problematising the educational thrust towards technology.

For ‘extensions’ of the ideas of Canguihem to the human sciences see entries for Michel Foucault, particularly ‘Foucault on Science’. See also entries for ‘Bachelard, Canguilhem and Foucault on Science’, and ‘Norms in Education’.


Balibar, Étienne et al. (1993) Georges Canguilhem: philosophe, historien des sciences, Paris: Albin Michel.

Bourdieu, Pierre (1998) ‘Georges Canguilhem: an obituary notice’, Economy and Society, 27(2&3), 190-192.

Canguilhem, Georges (1952) ‘Machine et Organisme’, Connaisance de la Vie, Paris: Hachette.

Canguilhem, Georges (1976) Vie et Mort de Jean Cavaillès, Ambialet: Pierre Laleure.

Canguihem, Georges (1978) On the Normal and the Pathological, Dordrecht: Reidal, 1978. Originally published as Le Normal et le Pathologique, Paris: Presse Universitaires de France, 1966.

Canguilhem, Georges (1988) Ideology and Rationality in the History of the Life Sciences, Cambridge, Ma.: MIT Press.

Delaporte, François (1994) A Vital Rationalist: selected writings from Georges Canguilhem, New York: Zone Books. Introduction by Paul Rabinow, pp.11-22.

Gordon, Colin (1998) ‘Canguilhem: life, health and death’, Economy and Society, 27(2&3), 182-189.

Gutting, Gary (1989) Michel Foucault’s Archaeology of Scientific Reason, Cambridge: Cambridge University Press.

Hacking, Ian (1998) ‘Canguilhem amid the Cyborgs’, Economy and Society, 27(2&3), 202-216.

Lecourt, Dominique (1975) Marxist Epistemology: Bachelard, Canguilhem, Foucault, London: New Left Books. Part I published as The Historical Epistemology of Gaston Bachelard, Paris: Vrin, 1969, and Part II as For a Critique of Epistemology, Paris: Librairie François Maspero, 1972.

Limoges, Camille (1994), ‘Critical Bibliography’, in (ed.) François Delaporte (1994), A Vital Rationalist: selected writings from Georges Canguilhem, New York: Zone Books, pp.385-454.

Macey, David (1998) ‘The Honour of Georges Canguilhem’, Economy and Society, 27(2&3), 171-181.

Marshall, James D. (1996) Michel Foucault: personal autonomy and education, Dordrecht: Kluwer.

Encyclopedia of Philosophy of Education



EEPAT is published in association with the Philosophy of Education Society of Australasia and the journal Educational Philosophy and Theory.