Castoriadis versus Michels

Biographical comparison

‘Why Michels?’ one might ask. It is significant to note at the outset that Castoriadis does not engage directly with Michels’s thesis on oligarchy, although one may trace an indirect engagement with it all throughout his writings. Thus, it is not my aim here to discuss an existing set of statements that Castoriadis makes about Michels, but to reconstruct a selection of arguments from Castoriadis’s work with the aim of imagining what his response to Michels would have looked like. It is equally significant to note the general absence of an engagement with Michels in Castoriadis scholarship. With one exception, ² the secondary literature on Castoriadis has ignored Michels’s thesis on oligarchy (and, indeed, the arguments of elite theory more generally). This is a real problem when it comes to understanding and evaluating Castoriadis’s theory of democracy. For when students of Castoriadis’s thought ignore Michels’s thesis, they turn a blind eye to the elephant in the room; they overlook or consciously evade what is arguably the most powerful challenge to radical democratic theory in general and Castoriadis’s theory of democracy more specifically. Such oversight or evasion can result in an uncritical celebration of Castoriadis’s political thought.

While there are numerous thinkers within the tradition of elite theory whose ideas could have served as the focal point of the present inquiry, I elect to focus on Michels for three reasons. For one, Michels observes the oligarchic tendency across an array of institutions. Despite focusing on political parties, he applies his thesis to unions, firms, and national governments as well. In this manner, Michels’s approach resonates with Castoriadis, who also thought about the contest between the few and the many across multiple institutional registers. Secondly, Michels develops his thesis on oligarchy in direct criticism of radical democrats, or thinkers comprising an ideological category under which one might justifiably place Castoriadis’s political thought and action (Scaff, 1981: 1280; Nye, 1977: 6). However, the third and most important reason for bringing Michels and Castoriadis into dialogue along the lines I have in mind is biographical in nature. Although Michels and Castoriadis are not exactly contemporaries, the striking convergences and divergences in their lives justify the effort to see them as interlocutors in ‘the transhistorical agora of reflection.’ ³

In a first phase of their lives, both Michels and Castoriadis advocate passionately for socialism. In 1903, Michels joins the German Social Democratic Party, but his strong support for syndicalism and his emphasis on the goal of a classless society does not sit well with the party’s established leaders, Karl Legien and August Bebel (Scaff, 1981: 1271). ‘In particular,’ Juan Linz states, ‘[Michels] deplored the way calculations of parliamentary advantage dominated party life and led to the abandonment of every vigorous idea and every energetic course of action’ (Linz, 1968: 265–6). Similarly, in the winter of 1942–3, Castoriadis joins the Trotskyist Parti communiste internationalist (PCI). However, as early as 1946, he begins criticizing the USSR for having created a new exploitative class, and in 1949, along with Claude Lefort, he forms the journal Socialisme ou Barbarie, which breaks away from the PCI due largely to the PCI’s continued defense of the USSR (Dosse, 2014: 50–8). Thus, much like Michels, Castoriadis joins an ostensibly revolutionary party, becomes critical of it, and then breaks from it on the grounds that the party has been disingenuous in its commitment to the revolution project.

In what may be considered a second phase of their lives, both Michels and Castoriadis turn their focus to theoretical labors, even if they continue to intervene and influence through their academic writings. By 1907, exasperated it seems from the lack of receptivity to the syndicalist idea, Michels leaves both the SPD as well as the Italian Socialist Party, of which he had also been a member (Scaff, 1981: 1272). His break with syndicalism appears to have occurred during this period, which was one marked by ‘critical self-appraisal, withdrawal from practical politics, and a shift in theoretical perspective’ (Scaff, 1981: 1272). The theses he develops in his magnus opus register the change in his perspective. ‘Political Parties,’ Philip Cook writes, ‘is a product of the Turin period. No part of it seems to have been written before his resignation from the socialist party. The rejection in Political Parties of the possibilities for democracy, syndicalism, and socialism documents a critical turning point in his life’ (Cook, 1971: 778). It is important to emphasize that Michels’s disillusionment during this time coincides with a turn to social scientific inquiry. ⁴

Castoriadis’s more politically engaged period comes to a close in the late 1960s and early 1970s. Socialisme ou Barbarie publishes its last issue in 1965 and dissolves entirely in May 1967, around the same time that Castoriadis immerses himself in theoretical labor and in ‘learning a new discipline, psychoanalysis’ (Dosse, 2014: 149). The journal discontinues for two reasons. For one, Castoriadis radicalizes his critique of Marxism to the point of alienating Marxists within the journal. A large part of this critique involves the rejection of class analysis and an affirmation of the significance of the symbolic realm in political life; here, Castoriadis crosses paths with Michels, who also emphasized the power of the symbolic over class interest. ⁵ Secondly, the journal discontinues due to a disagreement between Castoriadis and Lefort on the internal organization of the journal and the role it ought to play in broader events. Castoriadis argues for a concept of leadership within the journal; Lefort argues for a leaderless form of association. Castoriadis argues for a more direct involvement of the journal’s members in political organizing; Lefort argues that the journal should limit its activity to the production of ideas that the workers must judge without any further mediation by intellectuals. Amidst acrimonious debates, as well as Castoriadis’s belief that a burgeoning culture of ‘privatization’ had made it too difficult for a journal of the sort to successfully incite revolution, the journal dissolves (Dosse, 2014: 149).

Now, in what may be understood as a third phase of their lives, the courses of the two thinkers diverge radically. Michels turns to supporting fascism. Castoriadis remains committed to democracy and revolution. Michels officially joins Mussolini’s Partito Nazionale Fascista (PNF) in 1928 (Di Nucci, 1992: 91). Citing Carl Schmitt’s study of dictatorship, Michels explains that the ‘duce’ (leader) acts with decisiveness and clarity, unfettered by the ‘intrigues and of “connections” dear to majoritarian and democratic regimes’ (Michels, 1949: 121). He highlights Mussolini’s work ethic and compares him to historic figures like Louis XIV, Frederick the Great, and Napoleon (p. 128). He sees in Mussolini ‘not only the leader of a great party’ but also ‘the leader of a great state,’ and one who has brought the identification of the political party and the political leader to ‘its maximum development’ (pp. 136–7). Beyond the leader’s organic connection with the party, Michels sees in the figure of Mussolini the embodiment of Italian popular will:

It may be said that tempestuous times always find the men, their exponents, who faithfully express them, and that Italy found in its new ‘Duce’ or leader, at least in many ways ‘its living and active incarnation.’ And it appears to be true that when Mussolini speaks, he translates in a naked and brilliant form the aims of the multitude. The multitude itself frantically acclaims, answering from the profundity of its own moral beliefs, or, even more profound, of its own subconscious. (1949: 126)

Corrado Malandrino has tried to explain how Michels passes from a disinterested commitment to social science to an endorsement of fascism. He argues that the period 1913–23 could ‘be seen as the gestation period in Michels’s thought of a nationalist outlook’ (Malandrino, 2013). More precisely, the pivotal moment during these years occurs when Michels develops the ‘law of transgression’. Stemming in the ideas of William Wundt and Ludwig Gumplowitz, the ‘law of transgression’ posits that nations have an intrinsic tendency to grow beyond their hitherto existing national boundaries and that they do so at the cost of denying other nations the right to national self-determination. The ‘law of transgression’ thus transfers the assumption of historical determinism evident in Michels’s oligarchy thesis onto the realm of international relations and steers Michels’s thought to a justification of imperialism. ‘There is then good reason to assume,’ Malandrino writes, ‘that in the first years of the postwar period, gradually, and in correlation with the development of concrete political events that led to the rise of fascism, the axiom of national patriotism shuttled Michels first into the hive of Mussolinism, and then on to fascism’ (2013: 196–7). ⁷ In Mussolini and fascism, Michels appears to have seen a confirmation and a realization of the social scientific truths he had discovered earlier.

Unlike Michels, Castoriadis remains committed to the revolutionary project throughout his life. Whereas the social scientific apparatus Michels develops precedes his turn to fascism and apparently leads him to an endorsement of it, Castoriadis’s psychoanalytical research in the 1970s and his studies of ancient Athens during the 1980s did not attenuate his commitment to socialism and democracy. And while it might have been somewhat harder for former socialists to embrace fascism in Castoriadis’s time than it was during Michels’s time – given that fascism had not yet shown the full extent of its destructiveness – it is noteworthy that Castoriadis makes absolutely no turn to the right in his lifetime. He does not heed the call of liberal conservativism, as many former revolutionaries in France did (Negt, 1983). The task of this essay is to explore whether the philosophical concepts Castoriadis develops can help explain this steadfast commitment.

Michels’s thesis

What exactly does the ‘iron law of oligarchy’ mean? Michels holds that all forms of organization in modern times entail an oligarchic structure of authority. ‘In every organization,’ he writes, ‘whether it be a political party, a professional union, or any other association of the kind, the aristocratic tendency manifests itself very clearly’ (1912: 70), and, in another passage, he pithily states: ‘…who says organization, says oligarchy’ (1912: 364). Moreover, he rejects the possibility that things could be otherwise. Oligarchy is not choice but ‘a condition induced by the “tactical” and “technical” necessities which result from the consolidation of every disciplined political aggregate’ (1912: 365). These necessities pertain, of course, to the decisiveness, that is, the discipline and speed, with which organizations must act in complex competitive environments, and their need to either grow strategically, incorporating more and more human energy, or lose out in the struggle for power and influence.

The main theoretical implication of Michels’s thesis pertains to democracy. For if oligarchy constitutes the invariant principle of social organization, then democracy does not. Democracy cannot name the principle of an actually existing structure of authority. ⁸ If oligarchy is an invariant, then democracy names: either (1) an ideological construct that functions to ‘veil from superficial observers’ ⁹ the reality of oligarchy; or (2) a cultural tendency or institutional qualification that increases in the scope of those who could become part of the elite few. But, either way, democracy does not name a principle of organization. ¹⁰

Michels’s thesis rests on two premises. For one, it rests on the idea that ‘the mass’ (or, what one might refer to as, ‘the many’ or ‘the majority’ of citizens) is politically incompetent. It cannot act as a governing power, as it is ‘altogether incapable of undertaking the most necessary resolutions’ (1912: 65). Echoing Benjamin Constant (1988: 312), he states that the mass cannot deliberate due to its numerical size and geographical scope (Michels, 1912: 65). Echoing Gustave Le Bon, Michels argues the mass as a collectivity lacks sound judgment, and that it readily succumbs to passions, and it is thus easily controlled by ‘all kinds of men who are exceptionally bold, energetic, and adroit’ (Michels, 2012: 64). ¹¹ Moreover, anticipating Ortega y Gasset’s point about the ‘hermetism’ of mass man’s soul, Michels argues the mass is comprised mostly of individuals who are ignorant and indolent when it comes to thinking about politics and who are thus incapable of acting responsibly in moments of political decision (Michels, 1912: 65). ¹² Finally, Michels holds, the mass does not really want power. ‘Though it grumbles occasionally,’ he writes, ‘the majority is really delighted to find persons who will take the trouble to look after its affairs’ (1912: 88).

However, beyond just arguing that the mass are politically incompetent, Michels expresses a faith in the competence of the few who rule. It is at this point that his observation about the inevitability of oligarchy turns into a ‘moral justification’ of it. He writes:

The incompetence [of the masses] furnishes the leaders with a practical and to some extent with a moral justification. Since the rank and file are incapable of looking after their own interests, it is necessary that they should have experts attend to their affairs. (1912: 111).

Castoriadis and the competence of the many

In contrast to Michels – but, as we shall see, not without an interesting point of convergence in their thinking – Castoriadis treats democracy as if it refers to a principle of organization both on the level of political regimes and on the level of organizations in civil society. He does so without pause from the very first writings he pens to the very last. ¹³ A few statements from his writings should suffice to demonstrate the point. In 1956, Castoriadis affirms that the central task of social theory is to ‘aid in the growth of democratic-collective forms of organization run by the mass of workers themselves,’ and he asserts that the masses are not simply the ‘infantry’ of class action but a subject that can ‘develop astonishing capacities for self-organization, self-direction, and self-leadership’ (1988c: 24). In 1957, he holds that ‘the political structure of socialism will be largely based on collectivities of common work’ which will in turn be ‘the fertile soil on which direct democracy can flourish, as the ancient city or the democratic communities of free farmers in the United States of the nineteenth centuries were in their times, and for similar reasons’ (1988b: 99). In May 1968, in a text written during the uprising, he advocates for the ‘widest possible direct democracy,’ which he continues to understand through the ideas of council democracy and workers’ self-management (1993a: 135). In the 1980s, he undertakes an extensive study of ancient Athenian political institutions in search of the roots of the democratic imaginary and the cultural catalysts of its appearance in human history (2004, 2008, 2011). And in 1997, during the last year of his life, at the age of 75, he argues: ‘The only possible form of democracy is direct democracy, namely, a democracy where people decide for themselves and not by means of irrevocable representatives’ (1997: 9).

What kind of conceptual resources does Castoriadis’s work contain for challenging Michels’s first premise, or the idea that the masses are politically incompetent? Castoriadis’s distinction between ‘the political’ (le politique) and ‘politics’ (la politique) presents a first, crucial point of engagement with Michels. ‘The political,’ in Castoriadis’s lexicon, refers to the instance of ‘explicit power’ that exists in every society (1991a: 156). In turn, ‘explicit power’ denotes a coercive apparatus without which, Castoriadis holds, no society could endure, since it would not be able to protect itself against external and internal threats, or the effects of natural disasters. ¹⁴ ‘Politics,’ on the other hand, denotes an expression of collective activity which aims at the creation or perpetuation of autonomy. It names a form of ‘lucid,’ ‘reflective’ and ‘deliberative’ activity, a historically rare type of activity, and a type of collective activity almost always intertwined with the radical transformation of society. Politics, thus defined, exceeds the scope of leadership. It addresses the full set of society’s institutions, in which government and all instances of leadership are but subsets. And the type of creation that politics implies takes the form of a massive social historical process. It is massive in the sense that it can only be performed by the mass, or what Castoriadis calls the ‘anonymous collective,’ which acts despite and generally against the directions of established leadership. ‘Politics’ and ‘the political’ thus refer to different forces in the configuration of the polity, and politics denotes the type of radical, transformative process that can only be driven by the many. ¹⁵

Now, Castoriadis’s conception of ‘politics’ affirms the creative force of the many against the instituted authority of the few, and in this sense it constitutes a profoundly democratic concept. His concept of politics implies that it is not the actions of elites which shape a political order, but rather the multiplicity of energies discharged by the politically anonymous mass that does so. It therefore also implies not only that ordinary people matter in the shaping of a political order but that they matter most, that the shaping of the political order is ultimately their doing, whether they know it or not. Thus, in counting the anonymous collective as the real subject of political creation, Castoriadis’s concept of ‘politics’ imparts political competence to the masses; it imparts to them the competence to create the institutions of society, a set which of course includes political institutions as a subset.

However, there are at least two objections that may be raised to challenge, if not the validity of Castoriadis’s distinction per say, then at least the use of his distinction between ‘politics’ and ‘the political’ for the purposes of critiquing Michels. On the one hand, Michels’s statement about the inevitability of oligarchy relates specifically to the logic of authority structures within organizations. It is a statement focused on organizational technique, not on the creation of social and political institutions. Thus, while it might be right to think about political forces as forces that exceed the scope of leadership as Castoriadis does, this does not suffice to falsify the observation about how organizations are actually structured because it does not address the question of internal structure directly. Castoriadis gave a lot of thought to the issue of organization in his writing on worker self-management – and I shall return to this latter. But the concept of politics, which is so central to his democratic theory, operates on an entirely different level of analysis from the one needed to address Michels’s iron law of oligarchy thesis.

On the other hand, to claim that the many shape the polity does not amount to a demonstration of their capacity for self-rule. Even if one were to accept the claim, it would not follow that the many are capable of creating and sustaining an ‘explicit power’ which operates democratically. In order to argue that the many have such a capacity for self-rule, Castoriadis (or anyone thinking with Castoriadis) must be able to imagine a transformation of the many from a multitude, understood as an instituting power with disruptive force, into a demos, understood as an instituted power that embodies, secures, and retains the instituting power of the multitude it once was. Otherwise, it cannot be said that the many display a capacity for self-rule but only a capacity for resistance, which is different from the claim one would have to derive from Castoriadis’s concept of ‘politics’ if he or she were to use it in disputing Michels’s first premise. Among other things, an argument which purports to show that the many have the competence for self-government would require a detailed discussion of historical cases in which they have actually displayed such competence.

Now, Castoriadis does make reference to some historical examples in order to support his argument empirically. He holds that, in modern times, the many demonstrate extraordinary capacities for self-rule in the following instances: the US town halls during the 19th century, the Paris Commune of 1871, the Soviets in 1917 before they came under Bolshevik control, the German Revolution of 1919, the Spanish Civil War of 1936, the Hungarian Revolution of 1956. Yet it is not clear that the examples Castoriadis mentions demonstrate the self-organizing capacities of the many, at least not to the extent that they would have to do so if they were to explain his commitment to direct democracy. He does not discuss these historical examples in detail, and he thus invites his readers to simply assume that the examples support his argument about the historical feasibility of democracy. If one looks outside of Castoriadis’s work for the historical details he himself does not provide, one can easily cast doubt on whether the examples he mentions demonstrate the existence or even the possibility of direct democracy. The US townships were numerous and governed by the many, ¹⁶ but they were small in size ¹⁷ and, more importantly, they operated within a national framework that would have to be defined as ‘liberal oligarchic’ from Castoriadis’s standpoint. The other examples were too short lived. It is true that their trajectories were cut short by violence, a fact which must not be overlooked and which raises questions about whether the tendency toward oligarchy results from natural processes, as Michels suggests, or whether it is imposed and sustained through violence by already established oligarchies. But the short duration of these examples also makes it hard to explain how anyone could have such a great confidence in the capacity of the many for self-rule.

Castoriadis’s critique of oligarchy

The second relevant line of argumentation that one can draw out of Castoriadis’s work – what in the introduction I called Step 4 of my argument – focuses not on affirming the organizational competence of the many, but on displaying the irrationality of bureaucratic organization, or the irrationality of exactly those oligarchic structures of authority to which Michels’s thesis refers. Unlike Weber who sees in bureaucracy the rule of no one, the rule of an impersonal power, Castoriadis defines bureaucracy as a permanent division of labor between the few who act as ‘directors’ in an organization and the many who act as ‘executants’ of the former’s will. Thus, from his viewpoint, who says bureaucracy, says oligarchy. He critiques the irrationality of bureaucracy across its multiple manifestations in the government, in the operation of unions, in the workplace, and in political parties, though due to spatial limitations, I will only discuss the latter two.

In the workplace, bureaucratic oligarchy is irrational because it excludes ‘executants’ from decision-making and it thus fails to incorporate the ensemble of local knowledges that the many in each case possess. ¹⁸ The practical consequence associated with this exclusion is that it creates an ‘implicit struggle,’ a continuous tension, within the place of employment or the organization: workers find it harder to derive a sense of meaning from what they do; they do not feel motivated to do anything more than they must to keep their jobs, and this in turn increases the need for coercion on the part of management. In the workplace, the division between directors and executants is thus irrational because it wastes a large amount of creative energy, which could be channeled into a more productive organization of work. Castoriadis argues that the more rational form of organization would be based on workers self-management, whereby all members of an organization participate directly in deciding on the conditions under which they work.

The aim of Castoriadis’s critique of bureaucracy in the workplace centers on the problem of expert authority. Castoriadis denies that experts in and of themselves have the epistemic authority to make decisions on behalf of the whole. That is to say, he rejects the principle of technocracy. Against the ascription of political authority to experts, he draws attention to the fact that experts often have profound disagreements among themselves. Indeed, as the level of expertise among them increases, the possibilities of deeper disagreements might also increase. And, one might add, experts do not always act with the long-term interests of the organization in mind. Their instrumental rationality might serve designs that are rational only for those at the very top, not for the company as a whole. For Castoriadis, the logic of democratic organization retains a place for experts, as he does not deny disparities in knowledge, but the knowledge of experts never acquires absolute status and it does not afford them a right to decide on behalf of the whole. It seems then that his radical discounting of the epistemic authority of the ‘directors’ would explain why he has such a strong commitment to direct democracy.

Yet I do not believe it does. Castoriadis’s critique of bureaucracy cannot explain his unwavering support for direct democracy, since it cannot even explain an unwavering support for democracy in the workplace. Castoriadis offers a powerful critique of bureaucracy, but evidence exists to show that workers do not want direct democracy in the workplace, even if they might want to be consulted by management on certain issues. Carole Pateman, for instance, observes that workers generally do not want to participate in higher level decisions – i.e. decisions such as those pertaining marketing and investment – within their companies (1970: 82). ¹⁹ The unwillingness of workers to create institutions of self-management in their places of employment, a fact which Castoriadis is entirely aware of as early as 1961, seems like it would suffice to smother his conviction in the possibility of direct democracy in the workplace (1988a: 227).

Beyond the organization of the workplace, Castoriadis’s critique of bureaucracy extends to a critique of political parties, the focal point of Michels’s study, although this time Castoriadis does not only target the organizational irrationality of parties but a wider set of flaws in them. It is important to note that Castoriadis fundamentally agrees with Michels on the oligarchic nature of parties. Yet, whereas Michels ties leadership to concrete disparities of skill among their members and even passes to a moral justification of leadership on the basis of these disparities, Castoriadis denounces both the party structure and the ideas that authorize its leaders. For Castoriadis, political parties, like all bureaucracies, are not meritocratic: it cannot be taken for granted that politically competent individuals can rise to positions of leadership within parties. ‘The main point about contemporary parties,’ he writes, ‘is that they themselves are bureaucratic apparatuses dominated by self-co-opting clans’ (Castoriadis, 2007: 132). Against the idea that the orientation and problem-solving capabilities of parties can change as new faces come to occupy leadership positions within them, Castoriadis’s statement suggests that the upper echelons of party leadership tend to retain the same people or to recruit individuals who have the same hue as the individuals who occupied them previously. ²⁰ He also holds that political parties are indisposed toward solving problems. Rather, he argues, they ‘transfer all problems onto the false terrain upon which these problems can neither be solved nor even posed: the parliamentary arena’ (Castoriadis, 1993a: 145). ²¹ The main interest of elected officials is to win reelection, and they are likely to cater to the needs of donors who can help reelect them rather than to work toward the implementation of policies that address their society’s deep-seated problems. Finally, for Castoriadis political parties are not powerful. He argues that ‘…real decisions are made behind closed doors, backstage, or in places where those who govern meet informally’ (Castoriadis, 2007: 125). Castoriadis goes on to say:

It is well known that [political decisions] are not really made in those official places where they are supposed to be made. By the time they are presented to the Council of Ministers or the House of Representatives, the game is over…the motives behind them (the real ones, in any case) are secret, and in most instances legally beyond reach. (Castoriadis, 2007: 125)

Since Castoriadis does not develop his critique of political parties in an extensive manner, it is important to ask whether support for these critiques can be found in the empirical studies of other scholars. The claim that parties are ‘bureaucratic apparatuses dominated by self-coopting clans’ finds support, for example, in studies which show the prevalence of patronage structures in political parties (Kopecký et al., 2012), as well as in redistricting practices which aim at securing the electoral success of incumbents (Bullock, 2010). Other studies may be cited to substantiate his claim about the powerlessness of parties. Morris Fiorina, for example, observes the growing organizational and ideological powerlessness of political parties, and he ties it to a decline in political responsibility. ‘In contemporary America,’ he writes, ‘officials do not govern, they merely posture’ (Fiorina, 1980: 44). And the critique that political leaders make decisions ‘behind closed doors,’ or away from the kind of public deliberative settings valorized by classical liberalism, has a long pedigree. Carl Schmitt (1988) argued that the more consequential political decisions are not made by representative chambers but by powerful interests operating outside of such chambers, and, more recently, scholarship which examines the power of lobbies has worked toward confirming this view (Herrnson et al., 1998). Thus, one can find support for Castoriadis’s critical positions in empirical studies. Ultimately, however, Castoriadis’s critique of the bureaucracy and political parties does not explain why he was militant of direct democracy for the simple reason that thinkers of other political commitments, such as liberals or Thoreauian anarchists, can share the same critical perspectives about political parties without feeling logically, morally, or politically compelled to commit themselves to a direct democracy.

Conclusion

It has not been my task to judge which of the two thinkers, Michels or Castoriadis, is right, even though this question hovers over the foregoing inquiry and sneaks into it at every point. The question I have sought to address is whether Castoriadis develops ideas that can explain his unwavering intellectual and political commitment to direct democracy, and, more specifically, to reflect on whether these ideas can do so in light of Michels’s ‘iron law of oligarchy’ thesis. In my judgement, while the ideas Castoriadis develops work toward nourishing such a commitment, they do not explain it in the strict sense of the word ‘explanation.’ That is to say, his commitment to direct democracy does not follow from these ideas as a true conclusion follows from true premises within an inferential structure. The direct engagement with Michels’s thesis enables us to see that Castoriadis does not succeed in demonstrating the political competence of the modern masses, which a strong belief in the feasibility of direct democracy in modern times seems to require. Nor does he really demonstrate that oligarchy constitutes a contingency rather than a necessity of modern organizational life, at least not in the passages I have examined here. Rather than deriving it as one derives a true conclusion from true premises, Castoriadis upholds his commitment to direct democracy in spite of the strong, in spite of the seemingly unassailable, position that Michels defends.

The study of Castoriadis’s perspectives on democracy in juxtaposition to Michels’s thesis on oligarchy raises the question of whether political convictions ever have the structure of valid inferences. Indeed, Castoriadis’s case suggests that they do not. It suggests that people who have strong political commitments must play the game of sustaining them despite their ostensible falsifiability. Conversely, it suggests that when thought attempts to situate politics on seemingly unassailable conceptual foundations, on ‘iron laws,’ it can very quickly pass to a normative justification of authoritarianism. The foregoing study of Castoriadis and Michels suggests that, even among great thinkers like Castoriadis, political commitments are a matter of historically constructed subjectivity and personal, existential decisions about what ‘ought to be,’ rather than a matter of inferences drawn out of what supposedly ‘is.’ Castoriadis in fact says as much. He explains that if he has committed himself to democracy and the project of autonomy it is only ‘because [he] will[s] it and know[s] that others will it as well’ (1987: 373).