The Imaginary Institution of Society 
by Cornelius Castoriadis.
Polity, 418 pp., £14.95, May 1997, 0 7456 1950 9
Show More
Les Carrefours de Labyrinthe: Fait et a faire 
by Cornelius Castoriadis.
Seuil, 281 pp., frs 139, February 1997, 2 02 029909 7
Show More
The Castoriadis Reader 
edited by David Ames Curtis.
Blackwell, 470 pp., £50, May 1997, 1 55786 703 8
Show More
Show More

The first business of government, Confucius wrote in the Analects, is to ‘rectify names’. His point was that rulers should seek agreement on final ends. But reflection on the realities of power takes us from nomenclature to the nomenklatura: names, in the right, or wrong, hands are potent instruments of rule. ‘Words,’ Hobbes noted in Leviathan, ‘are wise men’s counters, they do but reckon by them; but they are the money of fools.’ Hobbes’s nominalism became the handmaiden of his realpolitik. Terms like ‘justice’ had no meaning apart from the facts of power, in a kind of dominant ideology thesis avant la lettre. Hence Hobbes’s comparison, at the end of Leviathan, between the Papacy and the kingdom of fairies – fictive edifices both, reared on the credulity of the downtrodden. Modern writers like Ernesto Laclau have used a similar idea to explain how the meanings of words are fixed by ‘hegemonic’ power relations.

Hobbes’s ulterior aim was, of course, to defend a form of absolutism, extending John Selden’s thought that, in a household, only one man’s there to buy the meat. But Hobbes’s views about meaning threaten to prove subversive. Counterfactuals – imagined possibilities – can be used to challenge the status quo. Invert hegemony and, handy-dandy, who’s the justice, who the thief? It’s tempting to go one step further, and conclude that the very idea of social rank based on merit is illusory. The recent erratic fortunes of the Ukanian banana monarchy make the point. Recall Alan Clark’s account of gaga party grandees wishfully turning blue the blood of their then leader, a provincial turnip-vendor’s daughter. Recall the September Days, and the tsunami of Diarrhoea which slopped over the nation during that dark time; for a while it seemed the body politic awaited the advent of a bloodied Martius to rescue it from democracy, though in the end it only got poor old Charlie Althorp. Amid this galère, such fantasms as ‘rank’, ‘deference’, ‘estate’ or, if it comes to that, ‘royalty’ itself, are apt to wither. The glories of our blood and state are shadows, not substantial things.

These themes occur throughout Cornelius Castoriadis’s work and especially in his best-known book, The Imaginary Institution of Society. During its composition this article has, regrettably, become an obituary. Castoriadis, who died on 26 December 1997 aged 75, was a wartime Greek Communist, an OECD economist, latterly a practising psychoanalyst, and long-time doyen of Socialisme ou barbarie (mangled in the Guardian as Socialism and Barbarism). He was born in Constantinople in 1922 and emigrated to France with the onset of the Metaxas dictatorship in Greece after the war of liberation. In Paris, Castoriadis joined the Fourth International and soon fell out with the Trotskyists in it, forming Socialisme ou barbarie with a congeries of radicals which eventually included C.L.R. James and the Sino-American, Grace Lee Boggs; free of the French Communist Party, he managed to avoid embroilment in the latter’s dizzying volte-fesses, chronicled by Sartre in Les Mains sales. He sided with the Algerian rebels against his adopted homeland and with the Hungarian uprising, having long since repudiated the Union of Soviet Socialist Republics (four words, four lies, as he later told Milan Kundera). His opposition to the USSR and his support of workers’ uprisings in Eastern Europe were bones of contention with Sartre, who later admitted that Castoriadis had been ‘right, but at the wrong time’: Castoriadis riposted, fairly enough, that Sartre had been wrong at the right time. An icon of the soixante-huitards, Castoriadis collaborated with the now Green Danny Cohn-Bendit on From Ecology to Autonomy. Five volumes of his writings were published under the umbrella title, Les Carrefours du labyrinthe, the last, Fait et à faire, last February. This was soon followed by The Castoriadis Reader, edited by David Ames Curtis, Castoriadis’s major expositor in the English-speaking world: a festschrift edited by Curtis appeared last May as an issue of the journal, Thesis Eleven, which has been a major platform for Castoriadis’s ideas.

In later life, Castoriadis wandered from Trotskyism into the ‘bourgeois deviationism’ of psychoanalytic theory and practice. The heteroclite ruminations collected in Fait et à faire take him by way of Aristotle, Leibniz, Kant and Merleau-Ponty, to Freud; the subjects dealt with include the subconscious, the collapse of the Soviet bloc, phenomenology, more on ‘the social instituting imaginary’ first broached in Imaginary Institution, and autonomy as a political project.

Human emancipation and autonomy are the themes that tie together the five-odd decades of Castoriadis’s writings, and they lie at the heart of Fait et à faire. His main charge against Merleau-Ponty’s phenomenology, for example, is its ontological ingenuousness. The gravamen of the charge is illustrated by his observation that, while Merleau-Ponty claims, in Le Visible et l’invisible, to have ‘dans la perception la chose même, et non pas une représentation’, he nonetheless helps himself to distinctions between given objects of perception and the imagination. According to Castoriadis, emancipation can be achieved by throwing off the ‘weight of the ontological inheritance’ with which not just received philosophy, but in a sense the whole of social reality, is encumbered. What Castoriadis has in mind, however, is not the process, familiar in the history of philosophy, of shedding one dodgy ontology only to be lumbered with another, as when (in Kant) ‘dogmatic metaphysics’ gives way to the metaphysics of unknowable things-as-they-are-in-themselves. Instead, he is offering a critique of the very idea of metaphysical commitment.

This is not an unambitious project: to lay out the possibility of recasting human thought in its entirety. To this end, Castoriadis applies the notion of the ‘radical imaginary’. This is what is held in doubt when we consider the basis for acts of imagining, where ‘imagining’ is used pretty broadly, in the sense of ‘conceiving’ or ‘making an object of thought’. Something superficially similar can be found in Vico’s New Science, where the faculty of ingegno drawing on a shared human sensibility, or sensus communis, creates culturally specific artefacts like the Twelve Tables. As Vico appreciated, there must be something capable of bracketing mental representations together as being of a certain sort. But the radical imaginary goes beyond this. Even mathematics is held to be ‘ensidic’ – a nonce-word coined by Castoriadis from the terms ‘ensemblist’ and ‘identitary’ and intended to signify the fact that thinking occurs relative to a conceptual scheme. From the perspective of the radical imaginary, the scheme might have been entirely different. Applied to mathematics, the radicalism of the radical imaginary becomes apparent: if mathematics is ‘ensidic’, then the laws of logic themselves are only true relative to a conceptual scheme. The underlying thought seems to be that individual imaginings occur relative to such a scheme, but different conceptual schemes are a priori imaginable; so there must be a more fundamental notion of imaginability (in virtue of which the different schemes are imaginable), determining the possibilities for individual imaginings. We then have to ask what the basis is for imagining (conceiving of) the radical imaginary itself.

It’s hard to answer this question, however, without abandoning the radical imaginary. A version of this problem can be seen in Hume. His project in the Treatise was, very broadly, to psychologise epistemology and ethics – to provide a philosophical account of the apparent objects of natural and moral knowledge which referred their explanation not to an ‘out there’ but to an ‘in here’. Causes, perceptions of moral qualities, natural necessity and so on, were all explained by Hume as the products of psychological affect – that was the sense in which Hume’s was a treatise of human nature. So, for example, the perception of virtue or vice in an act of incest or parricide doesn’t inhere in the act itself, but is ‘in here’: as the second Enquiry puts it, ‘gilding or staining all natural objects with the colours, borrowed from internal sentiment, raises in a manner a new creation.’ Morality, causation and the other phenomena which naive realism takes to be part of the world out there were to be explained by psychological processes. This gave Hume his error theory, explaining why we think these phenomena are part of the world out there, even though in fact they aren’t. Therein, however, lies the rub. If naive realism is explained by cognitive psychology, then scepticism about psychological mechanisms themselves has to be suspended. The problem comes to a head when, having rejected naive realism about causation, Hume relies on causal mechanisms to explain why we are led to believe in causation (through the ‘determination’ of the mind to move from impressions of one sort of event to ideas of another). Hume’s cognitive psychology had itself to assume that thinking of the sort which scepticism allegedly exposed as wrong could be true after all, as if a theory which set out to explain belief in Santa Claus were to end up positing a Santa-like entity. His explanatory project founders on his error theory.

Empiricism regarded the imagination as a combinatorial faculty; imagined content was collaged together from a store of simple images: ‘the one-eyed undertaker, he blows a flugelhorn’ or (Hume’s example) the New Jerusalem, ‘whose pavement is gold and walls are rubies’. This was, to be sure, not a very imaginative view of the imagination: for one thing, to imagine is not merely to visualise. What I imagine when I imagine the physical appearance of Adolf Hitler is not what I imagine when I imagine Alec Guinness made up to look just like Hitler, though the visualised content may be as similar as you like. To a first approximation, the difference is made by the conditions under which a proposition representing the imagined content would be true (or assertable, etc) in each case. I can not only imagine Tweedledum and Tweedledee, but can imagine mistaking either for the other.

Castoriadis is not concerned here with bog-standard acts of imagining, or even imagining in its wider sense of ‘conceiving’ or ‘entertaining’. The imagination can deceive, seeming to hold out possibilities which aren’t in fact there or to invite philosophical confusion. I sit here and imagine, after a fashion, Berkeley’s famous unseen tree. The Bishop’s aim was to show that this was incoherent – at least when put to the use which external realism demands of it. That use, after all, was to show that there are things which exist outside minds, and Berkeley’s move founders once it is acknowledged that imagining successfully is a matter of securing reference. That a tree which is imagined is unseen may be a parenthetical comment on the imagined content rather than an internal aspect of it. It may also appear that the freedom of the imagination entitles us to say just anything. When someone said he wished he were King of China, Leibniz replied that that man’s wish was, first, to be dead, and second, that there be a king in China. Either what is imaginable goes a lot further than what’s conceivable, and so can carry no weight of argument in determining what there is, or else purported imaginings have to be screened for coherence.

These considerations cast doubt on the ‘radical imaginary’. His reliance on psychological explanation poses a problem for Castoriadis, which parallels the old Humean aporia. Castoriadis has a good deal to say about ‘the magmas’, the basis of the radical imaginary, which in some respects play a similar role in his thought to Kant’s ‘things as they are in themselves’. In Imaginary Institution, a magma is what exists prior to ‘ensidic’ logic; it admits of ‘an indefinite number of ensemblist organisations’ but cannot be seen purely as the totality of these organisations. Castoriadis is well aware of the difficulty of saying what magmas are, being what necessarily cannot be said; the ‘best intuitive support’ is ‘all the significations of the English language’. Very loosely, then, the idea is that the world could be divided up into any of an indefinite number of different modes of classification, different job-lots of properties – something like a computerised cross-referencing system on a grand scale, with different search strings used to pull out different sorts of information from a database. The magmas are what make for the possibility of signification but they cannot themselves be signified (at which point, of course, the database analogy gives out). Thus the magma of ‘social imaginary signification’ is not itself such a signification, but is rather its transcendental ground. The magmas are not the ‘given’ (the unvarnished sensibilia or qualia of old empiricism) so much as the condition of anything’s seeming to be given in the first place. But then Hume’s difficulty appears. To say that there is a psychic mechanism which works on the magmas to arrive at particular social imaginary significations is to say that the mechanism itself lies beyond imagination, on pain of explanatory collapse. The project of psychoanalytical reduction in epistemology is a quest for a golden turd.

More fundamentally, it is not obvious that the magmas can elude the fate of the ‘given’. What in fact does the ‘given’ give us? Either we succeed in applying our concepts to bits of the world or we don’t. If we don’t, then anyone might as well say anything, about magmas or anything else – the concepts just spin away, in a frictionless void. But if we do succeed in applying them, then the world must in some way answer to the concepts, whether it be through the world-shapedness of the concepts, or the concept-shapedness of the world. There is no intelligible standpoint from which the answering of the world to the concepts could be a matter of luck, not least because we need concepts (examples: ‘luck’, ‘the world’) in order to express that very possibility, and we then have to say what in the world could answer to that.

This does not mean that all our concepts are instantiated. But once we have dismissed, as a case of metaphysical double vision, the thought that the answering of world to concept is a matter of serendipity, we have apparently left the magmas behind. If nothing at all is ‘given’ in the world of the radical imaginary – not even the conditions for the answerability of world to concepts – then there is good reason to think that the radical imaginary is a world beyond the world. In Imagination Dead Imagine, Beckett sets out to imagine what it would be for the imagination itself to be defunct. This was not the incoherent project of imagining what it would be to imagine in such a world, but how it would be for imagination to be impossible. What would not be coherent, though, would be to try to imagine that, while imagining that those conditions themselves were not met. This worry is dealt with by Castoriadis in ‘The Nature and Value of Equality’, published in the second volume of Les Carrefours du labyrinthe, by saying that the question ‘what, in that which is thought, comes from the one who thinks, and what comes from that which is thought? will for ever remain undecidable’ because the activity of philosophising ‘is this supremely paradoxical enterprise which consists in creating forms of thought in order to think that which is beyond thought’. In Frank Ramsey’s often-quoted gloss on the last line of Wittgenstein’s Tractatus: what we can’t say we can’t say, and we can’t whistle it either.

Thoughts along these lines are sometimes regarded as raising the possibility of conceptual relativism, in which we confront a community of language-users whose concepts are (in the fashionable philosophical idiom) ‘incommensurable’ with our own. There is a well-known argument by Donald Davidson to show that such a people couldn’t be known by us to be concept-users, since, by hypothesis, we would have no basis for translating what they said: we could only do so by imputing beliefs to them, and we could only do that by assuming that they used concepts cognate with ours. We might fail entirely to make sense of these people’s beliefs – or make sense of their having beliefs. For the reasons given by Davidson, though, we could not say: these people don’t think at all, don’t use concepts as ways of representing the world.

These conjectures may seem to land us in scepticism of the most vertiginous, Pyrrhonian kind. Hume’s response, not unnaturally, was to play backgammon and overeat, to the point of stratospheric fatness. But it may come to pass that the philosophical inquirer, weary of so much cold and stale speculation, comes to revel in the backgammoning, guzzling contingency of it all, and to celebrate the folkways he so lately debunked, turning native on his own patch. Latter-day examples include the self-styled ‘Post-Modern bourgeois liberal’ Richard Rorty.

The politics of this turn tend to the folksy – in an arch, snowstorm-paperweight-collecting kind of way – if not downright völkisch. And, as Clifford Geertz has noted, the wogs start long before Calais. Castoriadis, it should be said, never succumbed to this infarctus. He used to insist: ‘quoi qu’il arrive, je resterai d’abord et avant tout un révolutionnaire.’ His writings display a breadth of reading, intellectual vitality and interest foreign to most currently-practising Anglophone political philosophers. At the same time, they leave very few of Marxism’s ontic furnishings intact – not surprisingly, since ontology itself goes down the pan. So: no class, no forces of production, no surplus value, no architectonics of history. What remains, it seems, is a bare ideal of emancipation, here given a more or less explicitly psychoanalytic reading. The radical imaginary cuts away, with everything else, the categories on which psychotherapy depends.

Such a view of emancipation, moreover, begins to look like a symptom of a condition for which psychotherapy might be touted as a remedy. It’s at this point that the radical imaginary is in danger of collapsing into fantasy (using the term in its everyday sense, rather than its sense in Kleinian theory). Fantasy might be thought of as that act of imagination in which a certain kind of content, which otherwise might play a part in practical deliberation, is not imagined as such. Putting the radical imaginary in the service of ‘emancipation’ risks becoming not a, but the fantasy: that action could exist without any structure whatsoever, freed of the generic constraints to which all agency is heir. The fantasy is one of power. Thinking that it would be possible to act in conditions where the conditions for action aren’t met is an instance of this, common among contemporary political philosophers and activists. The idea is that it would be a good idea to get rid of power, by the use of power. To which one can only respond with the old Francoist war cry: Viva la muerte!

Behind this is a history of degeneration. What Kant called the ‘free play’ of the imagination became, in Romanticism, the cult of subjectivity, to which today’s po-mo sapiens is but a final throwback. Artistically, the apotheosis of will surfaced in excesses such as the Symphonie fantastique. Its major political outgrowths were romantic nationalism, and later Fascism, which combined an instrumental attitude to politics with wilful subjectivity. In a way, Nazism’s political aesthetic was of a piece with Marcel Duchamp and René Magritte: passing a urinal off as art. Or, to put it the other way about, the Belgian’s jeux d’esprit were as wilfully self-assertive as the political fabulism of the greater Reich or a Jew-free Europe. Twentieth-century politics has offered a richer fount of Surrealist invention than contemporary fiction, except of course that its practitioners may be able to suspend disbelief by fiat, in the puckish form of the NKVD or Interahamwe. We might think of the political dadaism of North Korea under Kim Jong-Il, the ‘Dear Leader’ whose title to autocracy was recently confirmed, in a po-faced official communiqué, by the sighting of an albino sea-cucumber.

‘When I use a word,’ says Humpty Dumpty in Alice in Wonderland, ‘it means exactly what I choose it to mean ... the question is who is to be master, that’s all.’ The masters have learnt Humpty’s lesson, in the imagined communities, the ‘New World Order’, the ‘State Law and Order Restoration Council’ of ‘Myanmar’, and the Biblical figment, ‘Eretz’ Israel – rather as if British foreign policy were to take its cue from Beowulf. This compulsive onomasia, the common coin of despot and demagogue, is indeed marked in the verbless rhetorical idiom of our own dear leader. New Labour. New Britain. Britain head and heart; Britain arse and elbow. But the thing defies parody. Res ipsa loquitur.

Send Letters To:

The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN

Please include name, address, and a telephone number.


Vol. 20 No. 5 · 5 March 1998

There were several errors in Glen Newey’s essay on Cornelius Castoriadis (LRB, 5 February). Castoriadis broke with the Greek Communists during, not after, World War Two. Newey says that he emigrated ‘to France with the onset of the Metaxas dictatorship in Greece after the war of liberation’, whereas the dictatorship had ended by the time of Metaxas’s death in 1941, four years before Castoriadis left Athens. Newey writes that ‘in later life Castoriadis wandered from Trotskyism’, but in forming the now legendary group and review Socialisme ou barbarie in 1948, Castoriadis had already instigated, at the age of 26, a definitive break with Trotskyism.

Even a cursory reading of my most recent collection of Castoriadis’s writings, World in Fragments, would have told Newey that what Castoriadis calls ‘social imaginary significations’ are not the product of any ‘psychic mechanism’. As Castoriadis says time and again, society is irreducible to the psychical – Newey irrelevantly criticises a ‘project of psychoanalytical reduction’. The association of Castoriadis’s internationalist left-libertarian ‘project of autonomy’ – which was ever resistant to fads as well as to authoritarian responses – with ‘the self-styled “Post-Modern bourgeois liberal"’ Richard Rorty, ‘romantic nationalism’, ‘Fascism’, ‘Nazism’, ‘the NKVD’, ‘Kim Jong-Il’ and Tony Blair seems wilfully ignorant. Pace Humpty Dumpty (whom Newey quotes), words don’t just mean what we want them to mean.

David Ames Curtis

send letters to

The Editor
London Review of Books
28 Little Russell Street
London, WC1A 2HN

Please include name, address and a telephone number

Read anywhere with the London Review of Books app, available now from the App Store for Apple devices, Google Play for Android devices and Amazon for your Kindle Fire.

Sign up to our newsletter

For highlights from the latest issue, our archive and the blog, as well as news, events and exclusive promotions.

Newsletter Preferences