The meanings of autonomy

Autonomy as project

When Castoriadis calls autonomy a ‘project’ and speaks of the operation of this project throughout history, one might imagine this signifies a yet-to-be-realized aim. Insofar as an autonomous society is yet to exist except in partial and imperfect forms there may be justification for this interpretation. What it misses, though, is the deeper significance of the term ‘project’. For Castoriadis, autonomy is a project by its very nature, and remains a project whatever the degree of its realization. This is not because any realization of autonomy will be imperfect. Rather, autonomy is essentially a project: to be autonomous is to be engaged in a project.

Castoriadis explores the concept of project briefly in the first part of The Imaginary Institution of Society (1987: 71–9). There he explains, first, that the project of autonomy – or what he refers to here as the ‘revolutionary project’ – must be understood as a project in the sense that it is a doing, and a praxis. The significance of the term ‘doing’ relates to a distinction between the work of such a project and works of technical production or making. The doing of the revolutionary project has a very different relation to knowledge than obtains in technical production. Though full and complete knowledge is always impossible, technique presupposes knowledge sufficient to guarantee the achievement of the desired end. The doing of the revolutionary project draws on knowledge, but this knowledge is incomplete not because of the impossibility of total knowledge, but because the nature of the project means what will need to be known can exist only as an ongoing creation of the project itself.

Technical making presupposes an end or final product as its aim. In contrast, a project is a doing that presupposes as its aim a state of being, which, as Castoriadis observes, is not an end at all, but a beginning (1987: 75). It is a beginning in the sense that it opens up an indefinite multiplicity of possibilities of doing and being, none of which may be defined beforehand since their ongoing definition is precisely the work and prerogative of autonomy. It is also a beginning in the sense that autonomy is desired not only for itself, but for what it permits us to do and make of ourselves; in other words, it is valued as the beginning of the possibility of making of ourselves that which we wish to be (1991a: 170–4). It nonetheless remains true that autonomy is itself an aim and goal. This is the meaning of the term praxis, as Castoriadis employs it. Praxis is ‘that doing in which the other or others are intended as autonomous beings considered as essential agents of the development of their own autonomy’ (1987: 75). So defined, praxis is the element of all true politics as well as all true pedagogy and medicine. This is a ‘beginning’ in that the aim is not to produce a subject with specific characteristics, except insofar as this subject is free and capable of making him/herself into the subject s/he determines. The desired end is not determined by the practitioner but by the autonomous subject who is thereby created, and who creates him/herself.

The autonomous subject is nevertheless a subject of a specific type, an historical creation coinciding with specific instituted social-historical worlds. In this sense, the autonomous subject is a social individual like any other, and when we aim at an autonomous subject we are indeed aiming at a social individual of a specific type (1991a). There can never be any absolute justification for the choice of this type of individual, because any justification will have its basis in values that are themselves historically created and conditioned. According to Castoriadis, we choose such a subject not because we can claim for it any absolute superiority, but because it embodies values we find within ourselves and our history, values we have endorsed (though not without critical examination), and because such an aim is consistent with what we understand about our true nature as human beings and social individuals. (This means we recognize that all the laws and norms that have hitherto been regarded as springing from external and unimpeachable sources are in fact of human origin. If one believes this, continuing to accede to a heteronomous existence takes on the appearance of absurdity or brute compulsion, and the maximization of human dignity begins to seem intimately connected to the maximization of human autonomy.)

Although the autonomous subject is a specific type of subject, and autonomous society is a specific type of society, we cannot determine in advance the detailed characteristics of these types. We cannot say what any autonomous society must look like or what specific institutions it must incorporate. We can specify some minimal features, and identify some things that must always be alien and antipathetic to autonomy. But anything beyond this is not a logical extrapolation of the concept of autonomy but merely our own contribution to the project. We must be wary of confusing such contributions with prescriptions about what must be. From the standpoint of autonomy, such ‘musts’ are always to be viewed with suspicion.

Finally, what is true of action aimed at the autonomy of another is just as true of action aimed at one’s own autonomy. Whether as an individual or a collectivity, if our aim is our own autonomy, we are engaged in a praxis in which the subject we are to become is intended as its own beginning.

Autonomy as self-limitation

Autonomy cannot be equated with self-limitation. This concept is too one-sided and negative. Firstly, limitation is by no means peculiar to autonomy. Society always offers individuals roles, beliefs, values, modes of living and so forth, and it usually does more than simply offer them, it imposes them. The provision of such models represents a form of limitation, since one can only ever be provided with a limited variety of them, and in being given only these and not others, one is being limited. But this is humanity’s universal and ineradicable condition, an unavoidable consequence of being a finite being and therefore having to determine oneself (or be determined) to some degree. One cannot be everything, nor can one be free to be anything at all. If it is justifiable to describe the provision of models as limitation, we must nonetheless recognize that this provision is not entirely negative; it does not merely tell individuals what not to do, it also tells them what to do, what can be done to have a meaningful human life.

It is true, of course, that alongside this positive modelling there is prohibition. However, this negative dimension is no more peculiar to autonomy than the positive. In advocating autonomy, one is not advocating limitation (negative or positive) where until then there had been none. Advocating autonomy is merely saying that one would prefer it if these limitations were self-imposed rather than imposed by others. (Without getting into the intricacies of this issue, we should acknowledge here that the ‘other’ who imposes the laws under heteronomy is ultimately – to some degree and in some sense – still ‘ourselves’. This ‘other’ is, according to Castoriadis, society, especially in its anonymous instituting capacity. It is the fact that this ‘other’ is ourselves – or at least is connected to and reachable by ourselves – that opens up the possibility of wresting control as deliberate and lucid agents of self-determination.) ¹

What is peculiar to autonomy is self-limitation. Those who may be sceptical about the possibility of autonomy because they doubt our capacity to engage in such self-limitation cannot reasonably do so on the basis that humans cannot be limited – we always are, always have been and always will be. What such sceptics must show is that humans cannot be self-limiting, that they cannot be the conscious origin of their own limitations.

The fact that humans have always been limited does not mean that they always keep within these limits. They almost always do sufficiently to permit the continued life of the society in which they live, but never completely. It is always possible for humans to break laws and to deviate from accepted models of behaviour and thought, and they always do so to some extent. This can never be eliminated (even if we wished to do so), for two reasons that are difficult to disentangle. First, there is a difference and potential disharmony between the social individual and the psyche, so that one’s being as a singular human is never exhaustively determined by one’s social role. This means that individuals may attempt for entirely psychological reasons to break or subvert prevailing codes of behaviour. Second, there is a capacity for creativity within the human person that is not entirely personal since it participates in the collective instituting capacity – what Castoriadis calls the ‘radical imaginary’ – and this can always well up and make itself felt as the emergence of alternative modes of being and thinking. It is not possible in advance to identify what will become the seed of social transformation and what mere personal deviation. Ultimately, this reflects the interconnection between the merely personal psychical ‘radical imagination’ and the collective ‘radical imaginary’ – an issue we cannot pursue here.

The fact that limitations will be exceeded or subverted is not peculiar to autonomy. What sceptics must show is that because limitations are self-imposed rather than imposed by others these limitations will be so ineffective the society will be unworkable.

Consider the difference between autonomy and its antithesis, heteronomy. Under heteronomy, the laws and institutions that govern our lives are understood to come from a source outside ourselves as conscious and deliberate agents and are therefore beyond challenge and deliberate alteration, whereas under autonomy the laws are made by us and are therefore always open to challenge, reform and replacement. If someone believes that social laws really do come from an outside source, then that person might suppose that autonomy must fail either because humans cannot make their own laws autonomously, or because they should not make laws for themselves.

We could respond to this by endeavouring to show, as does Castoriadis (1987), that social-historical reality cannot be understood except as societal self-creation, and that society’s laws are not heteronomous in the sense that this is understood in most societies. We might also need to show that humans are capable, or can become capable, of being the agents of their autonomous self-determination. We could point to historical examples of social autonomy, short-lived and imperfect though they may have been, but we would not expect such examples to convince really determined believers in heteronomy.

There are also sceptics who are already convinced of the falsehood of heteronomous world-views. These sceptics, while they accept that social laws are not bequeathed to us by a superior source, natural or divine, nevertheless doubt the ability of humans to become the conscious agents of their own self-determination. For these sceptics, the heteronomous world-view may be false, but becoming aware of its falsity is disastrous for humanity. Once we recognize that there is no objective and unalterable basis for our laws, no law can carry sufficient force to operate effectively. The result will be lawlessness and disorder.

Castoriadis acknowledges that the self-limitation necessary within autonomy is a perpetual problem. He analyses this in relation to the Greek concept of hubris. Recognition of the danger of hubris accompanied the growth and flourishing of the Athenian democracy and was, according to Castoriadis, explored most profoundly in Greek tragedy (1991b, 2007).

Hubris emerges as a danger within democracy because democracy is based on a rejection of the notion of fixed and objective standards for laws, thought and action. This makes anything possible. If this limitlessness in relation to the possible makes some sceptics of autonomy quake, according to Castoriadis they are not wrong to be wary, only to turn away in terror. If autonomy means we have rejected the idea that there can be a norm of norms, that the laws we create can be grounded in any substantial principle that is not itself open to redefinition and replacement, this does not mean that we are bound to run amok at every opportunity.

Castoriadis cites the famous instance of the Athenians’ genocide of the Melians, and the response to this in Euripides’ Trojan Women (2007: 122–4; 1991b: 118). This shows that alongside the potential for monstrous acts, democracy opens up an equal potential for self-evaluation and self-criticism. But the tragedy Castoriadis finds most profound and reflective of the dilemmas inherent in democracy is Antigone. Contrary to those who interpret the play as centring around the conflict between divine and secular law, or between the personal and the political, Castoriadis sees the central issue as a conflict between two interpretations of right and justice, mutually incompatible but each with reasonable claims to validity. Each is represented in the play by an intransigent advocate, and what the playwright through the person of the chorus extols at the play’s conclusion is phronein, ‘prudence/wisdom/balance’, warning against the ‘big words’ and ‘excessive pride’ that have characterized the play’s central combatants (1991b: 118–20).

Castoriadis says two things about this notion of phronein in the context of democratic politics. First, Castoriadis points out that what Creon above all is advised to avoid is monos phronein, ‘being wise alone’. To be wise is the most needful virtue, but to be wise alone is a failing of wisdom. Being wise alone signifies two things: first, being wise by oneself, insisting on one’s own judgement in isolation from the judgements of others – the antithesis of the democratic spirit. It also means being wise only, clinging to the rational defence of one’s own judgement without recognizing there is more to justice than logic and the rightness of reason. This arrogance of reason and the blindness to arguments the validity of which is based on other than logic is also antipathetic to the democratic spirit. Secondly, Castoriadis argues that the hubris this warning against monos phronein combats involves not the breaking of moral boundaries but the inflexible and intolerant imposition of them. As Castoriadis puts it, this ‘shows that hubris has nothing to do with the transgression of definite norms, that it can take the form of the adamant will to apply the norms, disguise itself behind noble and worthy motivations, be they rational or pious’ (1991b: 120). What Castoriadis, after Sophocles, suggests is that we cannot prevent such behaviour once and for all, only be constantly vigilant against it.

This notion of phronein also places the emphasis back on the actor, on the self, which is only proper from the perspective of autonomy. If there are dangers to self-limitation, these cannot be solved by reference to the limitation, but only by reference to the self who acts to limit. One cannot provide oneself with permanent and irreversible limits without thereby destroying one’s own autonomy (which is possible, of course). If one wishes to remain autonomous, one can only guard against hubris by one’s own effort and commitment, vigilance and prudence (1997). This may involve concrete measures, but these cannot be specified once and for all, because as autonomous creations they are at all times subject to the will of the autonomous agents, can be removed if the body politic wills it, and will fail to operate effectively if the political will does not support them. There are no guarantees. Bad things can occur, and sometimes will; mistakes are inevitable. But the failure of autonomy is not the unavoidable consequence of such shortcomings.

On the other hand, there is something to be said for the sort of limitation to be expected under autonomy precisely because of those characteristics which distinguish it from heteronomy and which sceptics see as weaknesses. Autonomous laws only limit the activity of the self insofar as the self has made and consented to those limitations. As long as this is genuinely the case, we might expect the transgression of those limits to be less likely than where they are imposed by an external force. In the latter case, we have guarding against transgression the (erroneous) belief in the infallible truth and necessity of the laws as well as the mechanisms of power erected to enforce them; in the former we can have similar mechanisms of power along with the consent of the subjects as authors and authorizers of these laws. In autonomy, we also have the possibility of reforming the law rather than transgressing it, a possibility which, as long as it is real and not simply formal, ought to lessen the risk of breaking laws. If sceptics believe that heteronomy is more likely to minimize lawlessness (or maximize lawfulness) than autonomy, that reflects assumptions being made about human behaviour. I will return to this question later.

Some sceptics of autonomy are conservatives in the literal sense that they oppose both change and the freedom to initiate change. Such people are likely to favour orderliness over social effervescence and diversity. There is no denying that autonomy cannot guarantee the former, and if the Athenian experience is any guide, the latter is more likely. ² A relatively stable autonomy is by no means impossible, however. And compared to the radical instability capitalist modernity has given us for most of its history, autonomy might well represent the more stable alternative.

Finally in this connection, there is the question of the relationship between the collectivity and the individual under autonomy. Sceptics of autonomy doubt that social cohesiveness under autonomy will be sufficient to make such a society viable. In the absence of objective and inviolable principles, such a variety of opposing views and lifestyles is likely to develop that no accommodation between these will be possible, nor will any effective co-ordinated action by the collectivity. Alternatively, cohesion will be maintained by force rather than consent, with the rights and freedoms of the individual being trampled on by the enforced will of the majority. It ought to be acknowledged that just as instability is not peculiarly related to autonomy, so too social fracturing and diversification is a condition we all recognize as characteristic of modernity, and it is hard to imagine autonomy producing more of it than we have already experienced. Sceptics suppose that where capitalism, and to some extent nationalism, have proven capable of serving as anchors for diverse cultures and subcultures, the aim and institutional structures of autonomy must fail. It must be admitted that an autonomous society can be based only upon agreement about one thing in particular: the value of autonomy. This is unlikely to be the only point of agreement in such a society, but it is the most crucial one. From it much else follows: above all, willingness to work towards realizing this value, as well as recognition that this means working together as a collectivity as well as working to maximize and preserve the autonomy of the individual. In relation to this point, Castoriadis states repeatedly that the rights and freedoms of the individual are essential to autonomy. His disagreement with some others over the issue of individual rights and freedoms turns on the question of how these can be guaranteed. Castoriadis rejects any notion that one can protect such rights from the collective will, arguing that ultimately they can only be an expression of that will (1997: 405–13; 2007: 122–3).

Ultimately, these criticisms of autonomy amount to little more than saying that autonomy is in danger if a majority of the population does not believe in autonomy. If this is seen as a weakness of autonomy, then it is a weakness it shares with every other society that has ever existed. No society can survive without the adherence of a majority of its members to certain core values and aims, to that society’s central imaginary significations.

Autonomy as democracy

There are those for whom the idea of individual autonomy presents no insurmountable difficulty but who nevertheless doubt the capacity of individuals to act collectively to determine their collective life. Such people may value individual freedom but remain sceptical of the viability and desirability of democracy. This scepticism increases with the effective reality of democracy, so that while a limited form of democracy is acceptable, extending and deepening democracy to make it ‘participatory’ is anathema. For Castoriadis, the current system in the so-called democratic countries is not democracy at all but ‘liberal oligarchy’. Choosing one’s rulers is not democracy. Democracy is ruling oneself, not being ruled by another, however that other may be chosen (2007). Putting aside the critique of contemporary ‘democracies’, what can we say in principle about the viability of democracy as a form of governance?

The sceptics charge that participatory democracy is too unwieldy to be workable, and that this is particularly relevant for populations as large as we find today. I will address this criticism later. Secondly, it is said that democratic processes inevitably, or at least too often, lead to poor decisions. This is the old truism about decisions by committee, and often turns on the view that many or even most people are idiots, so that the more people one involves in the decision-making process, the more idiots will wield power. (Castoriadis plays on the Greek etymology of the term ‘idiot’. The Greek idioteuein originally implied one who is a mere private individual. In other words, what made one an ‘idiot’ in ancient Greece was precisely the inability to participate in public affairs; 1993: 318.)

The idea that most people are unfit to participate in decision-making reflects both a view of politics and an anthropology. To say that the common run of individuals cannot make political decisions is to suppose that such decisions require skills and expertise that most people do not have and cannot acquire. It is to suppose, as Castoriadis has argued, that politics is a variety of technical activity, that it is dependent on theoretical knowledge, epistēmē, when it is in fact a matter of opinion, doxa (2007: 127; 1991b: 104–5). It is not that all opinions are equal: some are wise and far-sighted, others foolish and ill-informed. But there is no correct political knowledge to oppose to opinion. This is so because the object of politics, society as a whole, since its mode of being is essentially indeterminate, escapes and exceeds theoretical knowledge. It cannot be known exhaustively and one can predict its development only to a very limited extent. Politics is doxa rather than epistēmē because its outcomes cannot be determined in advance on the basis of its object’s functions and intrinsic ends. On the contrary, politics is the activity of determining those ends.

The sceptic may respond that there are things we must do, functional requirements of any society, and our particular kind of society, that must be met, and this constrains our freedom, determining our political actions. Without denying the existence of some functional necessities, the essentially indeterminate and self-creating nature of society suggests that the realm of necessity is extremely limited. In any case, the aim of autonomy must be to acknowledge necessity only when it is real, and much that is represented as necessary in fact is not. ³

Non-democratic political systems of all sorts, including liberal oligarchies of the Western kind, are not themselves immune from idiotic rulers and bad decision-making. Proponents of democracy merely contend that they ought to have as much right as others to rule and make decisions – good or bad – about their own life. There is a more serious aspect to this criticism, however. It is true that the average citizen of a Western country today is woefully unprepared to make wise political decisions. Most people are, in effect, political idiots. Not only do most people not know how to engage in collective decision-making, they have very little grasp of how their society and its institutions operate. We are ill-informed and unskilled. We are unfit for democracy.

Having admitted this, the question becomes ‘Why?’ Are people unfit for democracy because that is just the way people are – too ignorant and unskilled to govern – or is it because this is what our society has made us? Sceptics believe there is a fixed limit to the capacities of the average human being, and wise political decision-making lies beyond that limit. Those who agree with Castoriadis’s analysis will, on the contrary, insist that individuals are created by their society, that the characteristics and potentials of the individual are not universally but socially determined. There are limits to what a human can do, but we have no reason to suppose that this excludes political decision-making, when even sceptics admit that some humans possess this capacity, and when historical experience shows us that in some circumstances these skills and propensities have been extended to large populations.

How does one get from a population of individuals radically unfit for self-rule to one capable of it? Creating democracy is not a matter of social engineering. The creation of democracy is a praxis and a revolutionary project of a profound and radical kind. What is necessary in order to bring about autonomy and true democracy is a radical self-transformation of society. Like any transformation, it necessarily begins with a status quo ante. In a self-transformation, and one aimed at autonomy, the mechanism of transformation cannot be anything other than the autonomous activity of the subjects themselves. The subjects in question transform themselves by acting autonomously to make of themselves what they have decided they wish to become: subjects capable of autonomy. This may be circular, but as Castoriadis repeatedly points out, this circle – the circle of creation – is unavoidable (1991b). What this self-transformation must involve in concrete terms is a process of self-education aimed at giving all citizens the knowledge and skills needed to be competent citizens of a democracy. Especially at first, this self-education will involve a good deal of on-the-job training.

One thing that autonomous subjects do not give themselves is, ironically, the very thing that is most essential to autonomy. It may seem a conundrum, but what autonomous subjects do not give themselves is the project of autonomy itself. This project is a social-historical creation, and this creation is the beginning of autonomy rather than its product. If there were no project of autonomy nothing could justify us in precluding the possibility of its creation (though of course we would not pose the question if the project had not already been created). But the fact is we live historically downstream from this creation, we are the inheritors of it, almost all of us to some degree – even if we do not endorse it, we are the beneficiaries (and victims) of its various effects. The project of autonomy in this way differs from works of autonomy in that we do not make it because we will it. Rather, we find it within ourselves and our society, and then either endorse it and make it our own value and aim, or reject and ignore it (1991a: 120–4).

Unless people adopt this project as their own, no autonomous self-transformation of society is possible. Castoriadis is sanguine on prospects for this. ‘No great collective political movement can be created by the act of will of a few individuals’ (2007: 150). ‘The rebirth of the project of autonomy requires tremendous changes, a real earthquake, not in terms of physical violence but in terms of people’s beliefs and behaviour’ (2007: 149). Although nothing can justify us in discounting the possibility of autonomy, it will not happen unless a significant proportion of people have made the project of autonomy a guiding value for themselves, have begun to make themselves into autonomous subjects and are prepared to change both themselves and their society in the ways necessary to make social autonomy a reality. We can point, as Castoriadis does, to a number of real crises and latent attitudes that would seem to give us hope that this might be a possibility, but the reality is that it has not happened, and we are fully aware of how great a change it would require (1987: 79–95). As long as this ‘collective hypnosis continues’, writes Castoriadis, ‘there is a provisional ethical and political stance for those of us who have the weighty privilege of being able to speak up, namely: unmask, criticize, denounce the existing state of affairs. And for everyone: try to be exemplary in one’s behaviour and acts wherever one finds oneself’ (2007: 150).

What of the sceptics’ view that democracy is inevitably too unwieldy to be viable, especially in the long term? This criticism may have its basis in the erroneous assumption that true or participatory democracy means that absolutely every decision must be the subject of exhaustive deliberation by all citizens. This is not a necessary requirement for democracy, any more than Rousseau’s idea – ridiculed by Castoriadis – that democracy requires the identity of the sovereign and the prince, or the legislative and executive arms of government. Castoriadis points out that no such system has ever existed nor could it, since it would require all members of the governing body to be involved in not only making but carrying out every decision (2007: 118–19). Democracy requires that the decisions affecting the institutions and laws of the polity, along with decisions on policies that affect the lives of the citizens, be given over to the citizens themselves. Not only does this not preclude delegation of tasks and responsibilities, there is no doubt that it requires it, even in small populations and all the more so when population sizes increase.

How are the processes of democratic decision-making to be organized? There are historical examples of democratic organization, beginning with Athens and continuing throughout world history, that we can take as inspiration. Arguments might be raised against each example. Leaving aside the question of the merits of these models, we must admit that if anyone expects to be able to buy their democratic institutions ready-made, they are liable to be disappointed. Castoriadis says often that he does not intend to hold Athens up as a model to be slavishly copied by future democracies (2007: 119).

If democracy is to be viable in the present society, we must create the institutions that will make it viable. To say that no such institutions are possible because they have not yet appeared in history is to declare that the inventory of all possible institutions is already complete, ignoring the power of social-historical creativity. If democracy is to become real, this requires the invention of new solutions to practical problems and the creation of new institutions which make such solutions possible. If anyone underestimates the enormity of this task they are deceiving themselves. A profound revolution is required, and this revolution affects the institutional structures of society as well as the attitudes and ideas of individuals. When Castoriadis decries the lack of serious questioning of the concept of representation that underpins the legitimacy of representative democracy, he is lamenting the lack of imagination and creativity being devoted to problems of how to make democracy real (2007: 128–9). Our culture today is characterized by an acceptance of the status quo, either because it is viewed as inevitable or because we have no confidence in our capacity to alter it. Until this changes, we are not likely to find new institutional solutions to the problem of democratic organization. Creative solutions do not fall from the clouds; someone has to be searching for them.

One reason people dismiss the possibility of genuine democratic reform is that they take too much of the current institutional structure as given. They do not entertain the possibility of truly radical changes, whereas those who understand the extent of the task and support it acknowledge the necessity for truly radical change. Everything must be open to re-evaluation and reform: not just the administrative apparatus of the state but the nation-state itself, and the international system, etc. One area that not only should be opened to reform but must be reformed if true democracy is to be possible is the economy. We will turn to this momentarily.

One thing cannot be denied: democracy demands far more time and energy than the pathetically small contribution most so-called democratic citizens commit to political matters today. Sceptics may reply that democracy demands too much time and effort. Anyone with experience of democratic processes knows they can be exhausting. Part of the reason for this is because we today are unschooled in such processes. We have so little experience or training in such matters – it was no part of our education as citizens – that we fumble about like amateurs. All this reflects our experience and socialization, not the inevitable limitations of the human animal. Overcoming such ineptitude would be a necessary task for anyone who makes social autonomy their aim. In an autonomous society, those who begrudge the time needed for true democracy will be a minority – just as today’s political animals are extraordinary compared to the apathy of the bulk of the population.

Autonomy and socialism

In a 1979 paper, ‘Socialism and Autonomous Society’, Castoriadis explains his views on the term ‘socialism’. He thinks it ought to be abandoned: first, because it has become inextricably associated with actually existing socialist and communist regimes which brought ‘exploitation, oppression, totalitarian terror, and cultural cretinization to new heights in human history’ (1993: 314). But it is not only an accidental historical usurpation of the term that Castoriadis deplores. He argues that the term is bad because it is either tautologous or ambiguous. It is tautologous when it is interpreted as signifying an affirmation of society and the social. ‘What does it mean to be “socialist”, or even “communist”? To advocate society, sociality (or community) – and to be against what?’ (1993: 315). Ultimately, says Castoriadis, every society is ‘socialist’ in that every society is arranged with a view to its maintenance as a society and as this society. Society is the ultimate object of every society; this is the nature of society, and this is as true of an individualistic society as it is of a communalistic one. If we move from this tautologous interpretation, we find ambiguity. The term ‘socialism’ ‘seems to posit a material, substantive, ‘value-laden’ primacy for society over the individual – as if there could be ‘choices’, ‘options’, for society and against the individual’ (1993: 315). Rather than opposing society to the individual or choosing the first over the second, what is at issue, for Castoriadis, is the aim of creating a particular kind of society. What he proposes, therefore, is to substitute the term ‘autonomy’ for ‘socialism’. ‘What was intended by the term “socialist society” we henceforth call “autonomous society”’ (1993: 319).

The abandonment of the term ‘socialism’ does not imply the abandonment of all that it had come to signify. It is, rather, part of an attempt to preserve what is essential, and to extricate this from what is deemed to be not only inessential but detrimental and corrupting. What remains of the socialist tradition concerns particularly the economic dimension of autonomous society, and the relationship between the economy and power. It reflects the need to transform the role the economy has come to play in capitalist society.

For Castoriadis, autonomous society must put the economy at the service of the autonomous collectivity, transforming it from an end in itself and the ruling institutional sphere of the society, which it is under capitalism, to one sphere of human activity amongst others and a means to the achievement of autonomously determined social goals. This overall aim may be divided into three parts. First, autonomy within the economic sphere: this sphere cannot be an area of human activity from which autonomy is excluded. Individuals must have the same power to determine their economic activities as other activities and spheres of life. Thus workers’ self-management of economic enterprises becomes for Castoriadis an essential part of his vision of autonomy (1988a, 1988b, 1993, 2007).

Secondly, reform of the economy is necessary in order for political equality to become an effective reality rather than merely a legal status. Citizens must have equal opportunity to participate in political activities and to influence political decisions, and this is simply impossible with the disparities of wealth and economic power produced under capitalism. Thus Castoriadis calls for wage equalization, for the prohibition of unearned income, and other measures designed to produce such effective political equality (1988a, 1988b, 2007: 137–45).

Thirdly, social autonomy requires the effective de-autonomization of the economy. The institution of the economy must not be permitted to dominate all others in the way it does under capitalism (1987: 361–4, 1988a, 1988b). This requires reform of ideas, beliefs and values, certainly, but also institutional reform, so that the economy is subjected to the decisions of the autonomous polity. Castoriadis recommends that decisions about socially desirable rates of profit, distribution and re-investment be made by the whole collectivity democratically (2007: 144–5). In addition, he advocates measures to ensure the effective operation of economic markets, to free markets to a degree that, he argues, they are never free under capitalism. This too is conceived as a measure to maximize autonomy, since autonomy is not only a matter of collective decision-making but also of private and individual choices, and because markets – subject to democratically-determined limits and operating in a manner undistorted by inequalities of power and the capitalist drive to maximize profit at the expense of all else – can indeed be an arena for a form of individual freedom (2007: 137–45).

The two issues that emerge most strongly out of Castoriadis’s specific proposals for economic reform concern, on the one hand, the relationship between freedom and equality and, on the other, the overcoming of subjugation to the economy. On the second of these, Castoriadis argues that the manner in which the socialist intellectual tradition came to accord the economy a predominant role covertly perpetuated the capitalist view of the economy and the capitalist imaginary tout court. Castoriadis places the blame for this squarely on the shoulders of Marx and Marxism, which in this respect he sees as having exercised a corrupting influence on the workers movement. Driven by his desire to discover an explanation of society and history (a desire with its roots in the drive for rational mastery that is an essential component of the capitalist imaginary), Marx hit upon the determining role of production, and was thereby led ‘to narrow down greatly the movement’s preoccupations and aims; to concentrate completely on questions of production, economy and “classes” (defined on the basis of production and economy); and, quite naturally, to ignore or to play down all the rest, saying or implying that the solution to all other problems would arrive as part of the bargain when the capitalists were expropriated’ (1993: 323). This same myopia led Marxists to demote all social movements that are not class-based and focused on the economic sphere – movements such as feminism and youth movements – to secondary, minor or derivative status. Everything is seen as subordinated to the relations of production, and the movement that is ostensibly opposed to this subordination in fact insists upon it and thereby perpetuates it. The idea that this subordination would be magically overcome by the victory of this movement is, for Castoriadis, mere fantasy.

The other conceptual issue concerns the relation between freedom and equality. If reform of the economy is important from the perspective of autonomy this is primarily because of the recognition that the economy is the actual and potential source of an inequality that is destructive of freedom, defined as the capacity to become and be autonomous. In social terms, this means the capacity of each person to have as much say over laws and social institutions as anyone else. Large inequalities of economic power preclude this. Freedom requires an equality that renders impossible the political domination of some by others (1993: 316–21).

Castoriadis’s proposals for the reform of the economy should not be understood as prescriptions, but as contributions to the collective project of autonomous self-determination. How the economy actually operates under autonomy is a matter for the autonomous collectivity. In view of the preceding considerations, however, we can say that any arrangements which mean the economy is not subject to the control of the autonomous collectivity but vice versa, or that leave open the possibility of inequalities of power that preclude or diminish the freedom to participate equally in collective decision-making, must signal a diminution if not the inevitable demise of social autonomy.

To return, finally, to the question raised by the term ‘socialism’ of the relative importance and value of society and the individual: understood in the way Castoriadis advocates, the individual and society are not exterior to one another, and society can never be adequately understood as opposed to the individual. Nevertheless, socialism has historically signified the idea of a society in which the needs and desires of the group outweigh the needs and desires of the individual. This communitarian ideal is not the whole of the historical meaning of the term ‘socialism’, but it is an important part of it. Even admitting that individuals stand on both sides of this division (whatever else a group may be, it is also a collection of individuals), there is a real issue here, and it is twofold.

First is the question of how to divide the social space into realms that are subject to collective decision-making and those that remain the province of individual autonomy. Castoriadis’s main contribution to this question is his analysis of the difference between the private and the public. He identifies three distinct spheres: the private sphere (oikos), the private/public sphere (agora) and the public/public sphere (ekklesia) (1997: 405–13). His analysis centres on the issue of divisibility: those goods that are shared but cannot be divided and apportioned are public; the rest is private (1991c). These spheres are not trans-historical; what they shall and should be is an open question from the viewpoint of autonomy. Neither are they entirely separate from each other, so that the way one conducts oneself in the private sphere must always be in some respects and to some extent subject to laws of the broader public sphere. (As Castoriadis puts it, no one ‘is proposing that we remain indifferent to spouses murdering each other, or to parents raping their children’; 1997: 410). However, an autonomous society must guarantee the greatest possible mutual independence of such spheres (1997: 409). This obviously excludes the kind of collectivism in which the activities and choices of individuals are determined in minute detail by the collectivity. On the contrary, it envisages a regime in which the greatest possible freedom is given to individuals in their private activities consistent with the health – above all, the autonomy – of the collectivity. The reason for this stance is simple: an autonomy of the collectivity that is not at the same time an autonomy of the individual is not autonomy at all. It is, instead, what opponents of democracy have always accused it of essentially being: a tyranny of the majority. True social autonomy is the autonomy of the whole society, which means the autonomy of all.

Might one nevertheless say that autonomous society requires the instauration of a culture that recognizes the importance and value of the collective good? Yes, certainly. And more than that, one that recognizes our mutual interdependence and connection to the collective dimension from which we all emerge and outside of which we could not exist. And must this collective dimension be paramount? Must it in the last instance take precedence over the individual and private? The answer to this question is not so certain, nor is it essential to autonomous society. Such moral questions are the very ones that an autonomous society must grapple with, and to which it may never provide final and definitive answers. Certainly if the individual is accorded primacy over the collectivity in circumstances that would mean the demise of the collectivity, this would lead to the contradictory and self-defeating result of the demise of the individual as well, since individuals cannot exist outside of society. Ultimately, all historical societies have decided such issues in favour of the collectivity. I see no reason to doubt that this harsh reality would remain in an autonomous society, though in such a society we might hope that the collective good is really that and not just the good of the particular segment of the society that is advantaged by current social institutions.

Democratic decisions are rarely unanimous, and however one may organize processes to give everyone the opportunity to express their views, make their needs known and understood, and present their arguments, sometimes these will be contradicted by the collective will. This means not just that what an individual wishes does not occur, it also means that individuals will sometimes be required to comply with laws with which they do not agree. Some argue that this means there is an ineradicable element of heteronomy even within the most autonomous society, but I think this obscures the crucial difference between decisions which are made without any input by those who are affected by them and those in which all affected have the effective opportunity to participate. The term ‘heteronomy’ is best reserved for the former. Nevertheless, autonomy inevitably means that sometimes individuals are forced to obey laws they would not have chosen for themselves. Can this still be autonomy?

One answer is to recall Plato’s description of Socrates’ relationship to the laws of Athens (1969). For Socrates, the laws are his and he is bound to obey them by virtue of his fidelity to his city. This fidelity has a number of roots, including his gratitude for the city’s role in his education, his material and cultural sustenance, and the freedom it gave him to live a full and truly human life. But above all, he is faithful because he voluntarily joined himself to the collective enterprise that is the polis and, as such, recognizes his own freedom in the works of that collective enterprise, whether he happens to disagree with some of those works or not. By this example, we can see that it is possible for the sting of obedience to decisions that are not of one’s own choosing to be ameliorated by identification with the collectivity that is their author, and to identify oneself with that collectivity even when its will and judgements do not accord with one’s own.

Voluntary submission to collectively determined laws cannot be guaranteed in all cases, however, and individuals will sometimes be compelled to obey. This means the use of coercive power. Autonomous society cannot be entirely free of such power. It can only aim to be wise and lenient in its use, seeking to persuade and entreat rather than coerce, compelling only when it seems unavoidable and clearly in the collective interest as determined by the collectivity itself. We have already discussed the motives for such liberalism. Ultimately, they arise only from a commitment to autonomy, a commitment that cannot be guaranteed but must be fought for, fostered and preserved.