One of the great achievements of Jürgen Habermas’ discourse theory is his dialogical recasting of the Kantian idea of moral autonomy.

His dialogical recasting of the Kantian conception enables a way of thinking about autonomy that avoids some of the most serious objections that have been directed against the normative idea of autonomous agency. For the most part, these objections, which were picked up and developed by communitarian, feminist and poststructuralist critics in the 1980s and 1990s, have to do with connotations of self-ownership and self-sufficiency that the concept of autonomy acquired in the course of Western modernity. Habermas’ conception of moral autonomy is able to avoid these unwelcome connotations, while nonetheless accommodating the self-determining component of the ideal of autonomous agency (Cooke 1998).

In a legal-political context, his dialogical recasting of autonomy has the potential to redress certain features of liberal and communitarian approaches to democratic politics that Habermas finds troubling. Liberal approaches prioritize individual rights to freedom of choice over the communal freedom of democratic participation. This leads to the permanent threat of social disintegration since legal subjects who exclusively pursue their private interests erode the basis for social solidarity. Taking his lead from Kant (and Rousseau), Habermas argues that coercive law can preserve its socially integrating force only if the addressees of the law understand themselves as its authors (Habermas 1996: 33). Self-authorship calls for collective self-legislation: ‘It is only participation in the practice of politically autonomous law-making that makes it possible for the addressees of the law to have a correct understanding of the legal order as created by themselves’ (Habermas 1996: 121, emphasis in original).

Communitarian approaches, and the civic-republican models of politics Habermas connects with them, suffer from the opposite weakness. They prioritize realization of a shared conception of the common good over the freedom of individuals to determine for themselves their aims in life and the paths that will enable them to pursue them. He holds that this does not ‘sit well with the conditions of cultural and societal pluralism that distinguish modern societies’ (Habermas 1996: 279).

In terms of freedom, liberal approaches view legal-political institutions as placing restrictions on individual liberty; in consequence, they seek to protect negatively construed individual freedoms against the interventions of political authority by way of a system of subjective rights. Civic-republican approaches view legal-political institutions as a means for collective ethical self-realization and prioritize the positively construed liberties of communal political participation, downplaying or disregarding the importance of subjective rights to life, liberty and property (Habermas 1996: 268). Habermas’ discourse theory of law and democracy seeks to overcome the competition between the negative and positive liberties that he discerns in liberal and civic-republican models of democratic politics. Doing so entails reconciling private and public autonomy at a fundamental conceptual level. This is his co-originality thesis. According to this thesis, private and public autonomy are internally connected and evenly balanced (Habermas 1996: 84, 100, 104, 107, 118, 127, 129). I support his concern to take equal account of (what I prefer to call) the individual and communal moments of freedom and regard his dialogical recasting of Kantian autonomy as a fruitful starting point. I argue, however, that he fails to realize the promise contained in this recasting and, as a result, fails to establish his co-originality thesis. The main problem, I contend, is an unsatisfactory account of private autonomy.

1

Habermas’ rearticulation of Kantian moral autonomy is part of what has come to be known as discourse ethics (Habermas Reference Habermas1990). From the outset, Habermas has described himself as following Kant in the attempt to answer the question of what it means to act rightly in a moral sense. Like Kant, he limits morality to the class of universally justifiable normative judgments, leaving aside ethical questions of ‘the good life’. For Habermas as for Kant, particular needs, desires and values are ethical matters; the answers to ethical questions are inescapably context-bound and thus not susceptible to justification on grounds of their universality.

Like Kant’s moral philosophy, too, Habermas’ discourse ethics is deontological, cognitivist, universalist and formalist. It is deontological in the sense that it attributes an imperative, binding force to moral norms that is analogous to the unconditional character of truth claims. It is cognitivist in the sense that it answers affirmatively the question of whether we can rationally justify normative statements; indeed, it is not just cognitivist, connecting morality with the acquisition of knowledge and understanding, but has an in-built epistemic dimension, since moral knowledge and understanding is understood in a subject-and context-transcending, truth analogous, sense. It is universalist in the sense that it is construed in terms of universalizable interests. Finally, it is formalist in the sense that the distinction between valid and invalid norms is not made on the basis of their particular content but is rather decided in a process governed by a formal principle of universalization.Footnote ¹ This principle, ‘U’, requires valid moral norms to be universalizable, in the sense of acceptable to everyone as equally good for everyone. There is an evident analogy between ‘U’ and the Kantian categorical imperative, especially in its first formulation: ‘Act only in accordance with that maxim through which you can at the same time will that it become a universal law’ (G, 4: 421).Footnote ²

Habermas conceives of moral autonomy as a form of freedom that involves rational acceptance of the validity of particular moral norms: it is self-determination on the basis of moral insight. Again, there is an evident analogy with the Kantian view – autonomy for Kant, too, requires the self (‘autos’) to be governed by a law (‘nomos’) that is moral in an objective sense.

However, there are also some important differences between Kant’s moral philosophy and Habermas’ discourse ethics, and by extension, between their respective conceptions of moral autonomy (Habermas Reference Habermas1990: 203–4). Most importantly for present purposes, discourse ethics is dialogical in a double sense. First, norms are valid only if they could be vindicated by an agreement reached among participants in real argumentations (guided by idealizing suppositions). By contrast Kant assumes that individuals can test the validity of their maxims of action ‘monologically’, in isolation from others. Furthermore, the validity of moral norms is dialogical in the sense that it is dialogically produced: the validity of moral norms is not just tested in an (idealized) procedure of argumentation, it is generated within an (idealized) procedure of argumentation. In Kant’s case, by contrast, the validity of moral norms is not produced by way of argumentation, but is genetically prior to and conceptually independent of it.Footnote ³

In line with his dialogical rearticulation of Kant’s theory of moral validity, Habermas also offers a dialogical reinterpretation of moral autonomy. In his account, moral autonomy, like moral validity, is produced dialogically. More precisely, it is generated within intersubjective deliberations about the validity of moral norms. This means that moral autonomy, for Habermas, is genetically and normatively a social construct. In contrast to Kant, for whom moral autonomy is primordial and belongs inalienably to every human being, moral autonomy for Habermas has no pre-social justification. Thus, it cannot be used to ground pre-political natural rights that place restrictions on the exercise of popular sovereignty (as has often been the case in the liberal tradition of political thinking). This point is crucial for Habermas’ account of law and politics. It is crucial because it allows him to avoid the problematic normative prioritization of private autonomy, and corresponding downgrading of public autonomy, that he sees as characteristic of liberalism. Against this, Habermas asserts the co-originality of private and public autonomy (Habermas 1996: 84, 104, 107): both are forms of freedom the justification of which is purely social (indeed, he construes both as social in origin: as generated by actions within human society). In addition, as mentioned, they should be construed as internally connected and as related to each other in an evenly balanced way.

Habermas criticizes Kant for conceiving of legality as a limitation of morality (Habermas 1996: 113, cf. 105). He connects this with Kant’s genetic and normative prioritization of morality over law (Habermas 1996: 106). Kant takes the view that there is a primordial ‘natural’ right owed to each human being by virtue of their humanity: the right to equal individual liberties backed by authorized coercion (Habermas 1996: 100). This primordial human right is grounded in the autonomous will of individuals ‘who as moral persons, have at their prior disposal the social perspective of a practical reason that tests laws’ (Habermas 1996: 93). Habermas acknowledges that Kant connects private autonomy (here understood as moral autonomy) with popular sovereignty by way of a system of public laws that secures the freedom of each member of society as a human being, together with the equality of each member with every other. Public laws acquire legitimacy only as acts of the public will of autonomous and united citizens – in Habermas’ terms, through the exercise of public autonomy (Habermas 1996: 93–4). Habermas commends Kant for seeking to provide an account of the system of law in which private and public autonomy are connected. He argues, however, that Kant fails to make it an internal, conceptual connection. Kant assumes that no one exercising their autonomy as citizens could agree to laws infringing on their moral autonomy as warranted by natural law (Habermas 1996: 101). But, despite Kant’s efforts to do so by way of the construct of the social contract, he is unable to show an internal connection between moral (‘private’) autonomy and public autonomy. Habermas attributes this to his genetic prioritization of moral autonomy, which goes hand in hand with a normative prioritization: Kant follows ‘a path of justification that progresses from morality to law’ (Habermas 1996: 101). In consequence, Habermas not only sees an unacknowledged competition in Kant’s legal theory between morally grounded human rights, which guarantee private autonomy, and the principle of popular sovereignty, which engenders public autonomy (Habermas 1996: 94); he also observes that the competition is weighted in favour of private autonomy, which is legitimated independently of public autonomy (Habermas 1996: 100–1); by contrast, public autonomy has no legitimation independently of private autonomy. In order to overcome this competition between private and public autonomy, Habermas proposes a revised Kantian account of their relationship. In his alternative account, the scope of citizens’ public autonomy is not restricted by primordial natural or moral rights, nor is the individual’s private autonomy merely instrumentalized for the purposes of popular sovereignty.

I will come back to the co-originality thesis in section 3. For the moment, it is sufficient to note three interconnected features of Habermas’ dialogical recasting of Kantian moral autonomy: First, it denies to autonomy any pre-political justification in natural law; in this sense it is a form of social freedom; second, it construes autonomy as intersubjectively constituted, in the sense that it comes into existence by way of interactions between human subjects in processes of deliberation; third, it construes autonomy processually as a form of freedom that is generated within processes of deliberation.

In a larger project, I propose a reconceptualization of individual autonomy as ethical self-determining agency that shares these three features. However, I do not accept Habermas’ sharp distinction between moral questions, which he sees as subject to the principle of universalizability, and ethical questions, which he considers inescapably context-bound. My alternative dialogical conception calls for reflection on the validity of particular beliefs, interests and values from the point of view of subject- and context-transcending ideas of the good. In Habermas’ terminology, therefore, it is a version of ethical autonomy (I have more to say about this in section 3). Nonetheless, it shares the three salient features of Habermasian moral autonomy: it is social, constitutively intersubjective and processual.

The conception of ethical autonomy I propose plays a central role in a model of pluralist politics I consider suitable for contemporary democracies, in which citizens (broadly understood) hold diverging, possibly clashing ideas of what it means to lead an ethically good life (Cooke Reference Cooke, Kaul and Salvatore2020a). In the history of democratic modernity, encounters between citizens with conflicting ethical perspectives have often been seen as a source of social divisiveness, both by political theorists and those actively engaged in political affairs. They have also been seen as a threat to individual freedom, necessitating its protection, typically through a system of rights. It is indisputable that divergences of ethical perspective have historically given rise to terrible social conflicts and continue to be socially divisive. Nonetheless, the political pluralism I envisage views even discordant encounters between diverging ideas of the good life as potentially strengthening social bonds rather than weakening them and conducive to individual freedom rather than threatening it.

My proposed conception of ethical autonomy is based on an account of humans as agents who engage reflectively with the questions of what it means to live an ethically good life and how best to do so. Their reflections are guided by ideas of the good, which shape their identities and self-understandings and provide them with ethical orientation in their day-to-day lives. These ideas of the good may be multiple and even conflicting; moreover, they are normally tacit as opposed to explicitly articulated. The key point is that their evaluation and, as necessary, modification, change or abandonment, requires reflective engagement within intersubjective encounters. Ethical autonomy develops in and through these encounters. Ethical autonomy is self-determining agency, in the sense that ethical agents are responsible and rationally accountable. By this I mean that individual human agents take responsibility for the validity of the ideas of the good orienting their particular beliefs, interests and values, and for the ethical thinking and behaviour that follow from them. In addition, they must see themselves as rationally accountable for these life-orienting ethical evaluations, in the sense of being able and willing in principle to explain to others why they regard them as valid.Footnote ⁴ The same holds for their ethical judgements, decisions, actions and everyday conduct of life.

Validity is at stake in such intersubjective ethical encounters. In order to retain the epistemically construed, cognitive element which, like Habermas, I see as indispensable for normative ideas of self-determination, I build a reference to ethical validity in a subject- and context-transcending sense into the concept of ethical autonomy. Ethical validity is defined as at once inherently subject- and context-transcending and unavoidably subject- and context-dependent. This double characteristic helps to explain why ethical autonomy requires interaction with others.

The thesis of the inherently subject- and context-transcending nature of ethical validity is based on a presumption that validity in this sense is ultimately inaccessible to humans. Whereas as participants in intersubjective ethical deliberation we must assume that the exchange of reasons could lead to responses to questions about the good life that are better justified in a subject- and context-transcending sense, we cannot assume that they will ever produce definitive answers (Cooke 2006: 148–52).

The thesis of the unavoidable subject- and context-dependence of ethical validity stems from its inescapable mediation by language, culture, individual psychology and a variety of socio-political factors. This means that even epistemically and existentially significant experiences of validity (‘truth’) are mediated experiences and can be evaluated only as particular articulations of it. When combined with an argument in terms of the evaluative horizon of democratic modernity, the subject- and context-dependence of ethical validity entails its dependence on justification (Cooke 2006: 132–3). Justification is conceived as the intersubjective exchange of reasons in public processes, in which all participants are concerned to find the right responses to the concrete judgements, decisions and actions that are being discussed and evaluated in a given instance. In this respect, my thinking converges with Habermas’ view of public justification.

Reflective evaluation of the questions of what it means to lead a good life, and how best to do so, requires interaction with others who are similarly ethically engaged – who share a concern with ethical validity in a subject- and context-transcending sense.Footnote ⁵ This concern is shared generally, by all who see themselves as ethically self-determining agents. It relates to the validity of the ideas of the good that orient a particular human subject’s judgements, decisions and actions. By contrast, the concern of specific subjects with the validity of their particular judgements, decisions and actions is not generally shared, for their validity depends on a host of factors relating to the particular situation in which these individuals find themselves. Nonetheless, at any point in the critical exchange the question of the universal validity of the underlying ideas of the good life may rise to the surface and become a focus of discussion and contestation. In this account, as we can see, evaluation of ethical validity claims depends on agonistic processes of contestation and response, in which individuals challenge each other about what it means to lead an ethically good life. This is always potentially a learning process. However, ethical learning is not exclusively intersubjective but may also take place through self-reflection: the Socratic maxim ‘know thyself’ is an important component of it. In short, I conceive of ethically self-determining agency in terms of ethical learning through self-reflection and agonistic processes of contestation and response.

An evident difficulty here is that particular ethical reasons may not be readily accessible or, indeed, even intelligible, to others. These difficulties are exacerbated under cultural conditions of value-pluralism. Since mutual intelligibility and accessibility is a precondition for intersubjective engagement with the validity of ethical judgements, decisions and actions, how then is ethically self-determining agency possible? In other work, I respond more fully to this objection (Cooke 2013: 256–65, 270–4; 2016; 2020a). An important element in my response is to distance myself from the conceptual connection between validity and idealized rational agreement that Habermas asserts in the case of moral validity (Habermas 2003). I argue that meeting conditions of general intelligibility and accessibility may depend on epistemically significant shifts in affect and cognition. These shifts in affect and cognition are not restricted to deliberative contexts; indeed, they are typically set in motion not by arguments, but by experiences independent of deliberation. In order to take account of this, ethical validity, though construed epistemically as subject- and context-transcending, is not defined in terms of an idealized universal rational agreement but is, rather, transcendent even of such agreement (Cooke Reference Cooke2020b).

Furthermore, I contend that the challenges of general intelligibility and accessibility, though they must be taken seriously, do not render the process of intersubjective deliberation fruitless; rather they call on those engaged in it to be receptive to the epistemic significance of ethical experiences that are formative for other identities, to be hermeneutically sensitive in discussions of such experiences and to acknowledge the relative fragility, in the sense of epistemic contestability, of their own particular ethical convictions and commitments.

I have sketched an account of ethically self-determining agency in terms of reflective engagement with questions of the ethically good life by way of self-reflection and processes of agonistic contestation and response in which ethical validity is at stake. In this account, each individual shares a concern to lead an ethically good life with all other individuals who are able and willing to see themselves as ethically self-determining agents, although each of them may have very different ideas about what an ethically good life is. This presupposes that each individual acknowledges the equal status of every other individual as an ethically self-determining agent. Such acknowledgement of equality is fragile. It is vulnerable to implicit social biases (relating to gender, ethnicity, race, social class and so on); to the competitiveness structurally built into the capitalist economic system which, in modernity, has provided the material basis for democratic association; and to human failings such as envy and malice. For this reason, reciprocal acknowledgement of each human subject’s status as an ethically self-determining agent must be formally supported by law: by institutionalized legal recognition of the equal social standing of each individual as an ethically self-determining agent.

In addition to legal-political recognition of each subject’s status, ethically self-determining agency depends on other socio-cultural conditions, enabled by a corresponding system of social institutions. Basic needs for food, accommodation and clothing must be met, there must be an economic redistribution system that guards against significant discrepancies in wealth and social status and an educational system that grants all children access to schooling from childhood to maturity. In addition, a variety of contextually specific socio-cultural conditions may also contribute to the further development of ethically self-determining agency.

Social institutions serve not only to stabilize and foster socio-cultural conditions that enable ethically determining agency and contribute to its further development. They also serve functions of ethical orientation and guidance. In the next section, I explain what I mean.

2

What are social institutions? Classical sociological accounts of social institutions define them as socially constructed, supra-individual entities (Berger and Luckman Reference Berger and Thomas1967). Examples include families, parliaments, religious congregations, trade unions, sports clubs, courts of justice, the internet, non-government organizations, schools, the World Bank, the printed media, the United Nations and cultural agencies.

I follow Luc Boltanski in using the term ‘institution’ to refer to entities that primarily serve the semantic function of shaping and stabilizing social meanings (Boltanski Reference Boltanski2011). They have other functions, in particular, policing and administration, but their primary role is to form and stabilize meanings. Boltanski is correct to highlight the semantic shaping and stabilizing functions of social institutions. However, he pays little attention to the specifically ethical character of the meanings they shape and stabilize. By contrast, I want to emphasize the role of social institutions, individually and in configuration, in constituting webs of ethical meaning. In my account, social institutions are incorporations of – often diverse and sometimes conflicting – ethical values. As such, they have (more or less stable) ethical identities. The incorporated ethical values shaping institutional identities form a multi-layered and multi-dimensional ethical sedimentation. This ethical sedimentation is the complex historical product of human interactions within the institution, as well as interactions between that institution and other institutions and with its non-institutional environment; however, it may pass unnoticed by the institution’s members (broadly understood). Nonetheless, by way of their webs of ethical values, social institutions, more or less implicitly, provide general ethical orientation and specific guidance: they point their members in certain ethical directions, thereby impacting on their particular ethical values and inflecting their particular identities as ethical beings; at the same time, they provide specific direction in the conduct of everyday life by way of concrete manifestations such as laws, ordinances, rules, policies, doctrines, codes of behaviour, recommendations and other prescriptions.

The authority of social institutions resides in this ethically orienting and directing power. Its concrete manifestations are authoritative for particular human agents in particular life-situations, whenever they perceive them as important aids in their endeavours to live an ethically good life and as powerful motivations to live such a life. Authority becomes authoritarian when the institution legitimating the exercise of authority, or those in leadership roles within it, take control of the truth of these ethical values in a way that renders them incontestable. At issue is less the substance of the ethical orientation or guidance that is offered than its contestability by the human agents concerned – the main issue is whether it is offered or imposed. Authoritarian authority always imposes ethical direction. It is hostile to ethically self-determining agency because it compels its subjects to accept unquestioningly certain ways of thinking and acting, disallowing the kind of contestation that is a condition of ethical learning.

If social institutions are to exercise power that is authoritative, but non-authoritarian, they must be open to transformation in response to the ethical challenges they encounter. These challenges may be addressed to various aspects of the institution’s ethically inflected identity: at its operation, its organization or its incorporated ideas of the good. This means, in turn, that social institutions must see themselves, and be seen by their members, as in a permanent process of constitution: they must recognize the inherent instability of their institutional identities. Social institutions must acknowledge, furthermore, that the process of constitution is ethically motivated: driven by a concern to inflect the institution in question with particular ethical values. Since in the societies of democratic modernity, the particular ethical values orienting and guiding the members of social institutions are often plural and sometimes conflicting, the process of constitution will be agonistic rather than harmonious. Nonetheless, the institution’s members, insofar as they are able and willing to see themselves as ethically self-determining agents, will consider themselves part of a common project of constitution — as co-authors both of a common good that defines the (unstable) identity of the social institution in question and of their own ethically self-determining agency. In short, for institutions to be non-authoritarian, yet authoritative, they and their members must engage in a perpetual process of mutual ethical identity-constitution.

From this we can see that my sketch of autonomy as an intersubjective social process goes hand in hand with a sketch of social institutions as the site of the non-authoritarian authority that contributes substantively to this process.

In political contexts, I use the term ‘common good’ to refer to the normative purpose of an ensemble of ethically inflected, legal-political institutional arrangements designed to contribute equally to the development of each member of the political association as an ethically self-determining agent. These legal-political institutional arrangements establish a system of law that guarantees the equal status of citizens as ethically self-determining agents, both in their memberships of particular social institutions and as participants in the collective process of constructing political common goods. They also contribute to the constitution of the identities of citizens as ethically self-determining agents. They do so by incorporating ideas of the good that offer citizens substantive ethical orientation and guidance, which they can accept or challenge in processes of critical engagement with their own ethical ideas and practices. From this we can see that legal-political institutions potentially enhance individual autonomy in two interconnected ways.

First, by formally enabling ethical autonomy, in the sense of securing the conditions necessary for its constitution, by granting each citizen a legally guaranteed equal status and ensuring the socio-cultural conditions necessary for its development.Footnote ⁶

Second, by improving the substantive quality of ethically self-determining agency. By way of concrete laws, ordinances, policies and other kinds of prescriptions and recommendations, they offer citizens general orientation and specific guidance in relation to what it means to live an ethically good life, which citizens may accept, reject, contest or ignore.

From this we can see that authority, too, is an epistemic concept; furthermore, that it, too, is permanently in process of constitution: it is constituted continually, in a dynamic movement, in which human agents, individually and collectively, engage with specific ethical prescriptions and recommendations, affirming or challenging them, modifying them or rejecting them, but in every case contributing to critical evaluation of their claims to authority.

3

In his writings on law and democracy from the mid-1990s onwards, Habermas offers an account of legal-political institutions that in important respects is similar to the one I have sketched.Footnote ⁷ In his account, legal-political institutions serve purposes of citizen engagement in collective processes of law- and decision-making. He calls this collective self-legislation or public autonomy. As mentioned earlier, public autonomy is held to be co-original with private autonomy, by which he means internally connected and evenly balanced. For reasons I now explain, I understand this as the claim that the relation between private and public autonomy is one of dialectical interplay. By ‘dialectical interplay’ I mean a movement between two distinct elements in which each depends on the other for its full development and thus changes substantively through interaction with it. In other words, each element has a distinct transformative, generative power, contributing significantly to the constitution of the other. Put differently again, private and public autonomy are reciprocally constitutive.

At first glance, Habermas’ remarks on the co-originality thesis allow for a different, weaker reading. On this reading, he asserts the reciprocal dependency of private and public autonomy as opposed to their reciprocal constitution. Closer consideration shows, however, that only the stronger reading will meet the requirement that both forms of autonomy are evenly balanced – that both have equal normative weight. Habermas makes clear that this requirement must be met: ‘the private autonomy of citizens must neither be set above, nor made subordinate to, their political autonomy (Habermas 1996: 104). Kant’s legal-political theory fails to meet it, since it puts morally grounded freedom ahead of political will-formation (Habermas 1996: 101).

The weaker reading is suggested when he writes that private and public autonomy ‘mutually complement’ each other (Habermas 1996: 99, 106) or ‘mutually presuppose’ each other (Habermas 1996: 84, 128). On this reading, private autonomy is a necessary condition of public autonomy in the sense that only agents whose equal status is guaranteed by a system of subjective rights could engage in collective self-legislation. Conversely, public autonomy is a necessary condition of private autonomy in the sense that recognition of each citizen as a bearer of rights must be secured through the exercise of public autonomy, typically generating a system of subjective rights.

The stronger reading is suggested when he speaks of reconciling private and public autonomy at ‘a fundamental conceptual level’ (Habermas 1996: 84), of showing their ‘internal’ relationship (Habermas 1996: 104), and of the ‘dialectical relationship’ between private and public autonomy (Habermas 1995: 130). Not only is there textual evidence for the stronger reading; only the stronger reading would suffice for the purpose of a balanced relationship, since if private and public autonomy were not dialectically connected, private autonomy could have normative priority over public autonomy, even when constrained by a system of rights. This would weaken the normative status of public autonomy, making it dependent on an idea of private autonomy that has independent normative weight – which is Habermas’ objection to Kant, and more generally to the liberal tradition of political thinking. My contention, accordingly, is that in order to achieve a balanced relationship between private and public autonomy, it is insufficient to establish a relationship of mutual dependency. The success of the thesis requires Habermas to establish their reciprocal constitution.

The account of ethically self-determining agency in a political context that I sketched in section 2 illustrates the required kind of dialectical interplay. In my account, the authoritative outcomes of democratic decision-making processes (‘public autonomy’) can contribute in a constitutive way to the further development of ethically self-determining agency (‘private autonomy’). By this I mean that authoritative outcomes (laws, ordinances, policies and other prescriptions and recommendation) can enable ethically self-determining agency not only externally, by establishing social institutions such as a system of rights that provide a secure social basis for its development; they can also enhance it internally, contributing to the constitution of its ethical substance by providing concrete ethical guidance. Conversely, ethically self-determining agency can contribute to the constitution of the ethical substance of public autonomy by way of agonistic citizen interventions that assert the importance not only of particular egocentric or strategic interests but also of particular ethical beliefs and values.

My account proposes an ethical version of the dialectical interplay between ‘private’ and ‘public’ autonomy (in my terminology, between the individual and communal moments of self-determining agency); however, as we shall see, a non-ethical, moral version is also available. The key point is that for the co-originality thesis to succeed as a thesis about the evenly balanced, mutually dependent relationship between private and public autonomy, each of the two moments – what Habermas calls private and public autonomy – must be made dependent on one another for their full development. Do Habermas’ conceptions of private autonomy and public autonomy meet this condition?

The answer to this question is not straightforward. One difficulty is Habermas’ lack of clarity as to what he means by private autonomy. Sometimes he adopts a Kantian legal-theoretical perspective, in which it means subjective freedom of choice (Willkür); sometimes he adopts a Kantian moral-theoretical perspective, in which it means moral autonomy. In the initial part of his discussion, when addressing the strengths and limitations of Kant’s approach, private autonomy clearly means moral autonomy (Habermas 1996: 93–106). In his subsequent discourse-theoretical justification of basic rights, it clearly means freedom of choice (Habermas 1996: 118–31).

Another difficulty is that he switches between a functional, objectivating, perspective on law and the subjective perspective of agents who are subject to the law, without making clear that he is doing so (Peters Reference Peters1994; Cooke Reference Cooke1994). From the beginning, he states emphatically that he is not attempting a normative justification of law, but rather offering a functional explanation that follows Kant’s account of the legal form (Habermas 1996: 111–12).

According to Habermas, Kant characterized the legal form through three abstractions referring to the subjects addressed by the law. Law abstracts from moral motivation, making free choice a sufficient source of law-abiding behaviour. Law abstracts from motivation in general, simply demanding conformity to rules. Law abstracts from concrete moral and ethical identities, concerned only to regulate behaviour externally. Habermas does not attempt to justify Kant’s view of the legal form, which he takes to define the modern form of law, claiming that it is not a principle one could justify, either epistemically or normatively (Habermas 1996: 112). Rather, he holds that it has emerged through a process of social evolution, giving us to understand that this is irreversible and should be regarded as a learning process (Habermas 1996: 111). Functional explanation looks at the law as it serves purposes of social coordination, regulation and organization. In modern societies, it functions to offset deficits arising from the collapse of traditional ethical life. With this collapse, morality is detached from tradition and custom and becomes postconventional, supported by reason alone. This leads to an overburdening of individual and collective moral judgement and decision-making, making law necessary as a relief mechanism. In addition, law must assume a form that meets the organizational needs of increasingly complex, advanced capitalist societies (Habermas 1996: 113–18). Whereas functional explanation looks at law as a medium and institution from the outside, a normative justification looks at law from the inside, taking the subjective perspective of those who are subject to it and considering why they have good reasons to accept and obey the law.

Confusingly, however, given his insistence that he is concerned with functional explanation, he elaborates his co-originality thesis under the heading ‘A Discourse-Theoretical Justification of the System of Basic Rights’, implying that here he is adopting a subjective rather than functional perspective (Habermas 1996: 118–31, my emphasis). The language he adopts in his normative justification of the system of basic rights is further cause for confusion. He appeals to an idea of private autonomy as subjective freedom of choice (Willkür) in language evoking ideas of self-ownership and self-sufficiency closely associated with liberal conceptions of freedom as absence of interference. Thus, he describes private autonomy as freedom from communicative freedom, by which he means discursive freedom, the forms of freedom acquired by agents in intersubjective processes of deliberation. It is freedom from the obligations of communicative freedom: ‘the negative freedom to withdraw from the public space of illocutionary obligations to a position of mutual observance and influence’ (Habermas 1996: 120). At first glance, formulations such as these could be read as expressing the functional, objectivating standpoint of modern law, in line with his Kantian abstraction from the capacity for moral autonomy and corresponding view that free choice (Willkür) is a sufficient source of law-abiding behaviour. However, this reading is at odds with the intention suggested by the section heading, which is to offer a justification of the system of basic rights. Moreover, he is clearly adopting a subjective rather than functional perspective when he writes that the agent ‘who simply decides as she wishes is not concerned whether the reasons that are decisive for her could also be accepted by others’ (Habermas 1996: 119–20). Or again, when he writes that ‘[l]egally granted liberties entitle one to drop out of communicative action, to refuse illocutionary obligations; they ground a privacy freed from the burden of reciprocally acknowledged and mutually expected communicative freedoms’ (Habermas 1996: 120). Remarks such as these, which express a subjective perspective (lack of concern by the subject as to whether her reasons are acceptable to others, freedom from the burden of having to give reasons) point in the direction of an idea of private autonomy that seems indistinguishable from the liberal conception of individual freedom targeted by his co-originality thesis. To all appearances, therefore, the idea of private autonomy we find in his justification of the system of basic rights is not dependent on public autonomy for its further development. It is not open to substantive transformation in dialectical interplay with the outcomes or exercise of public autonomy; rather it is a form of freedom that is guaranteed by rights and protected by them against external interference. It is noteworthy that he approvingly cites Hannah Arendt’s characterization of the law as a shield or ‘protective mask’ that is held up over the physiognomy of persons to protect their freedom of choice from interference by others (Habermas 1996: 531–2, n. 42).Footnote ⁸

Things look different in the case of public autonomy. For Habermas, again following Kant (and Rousseau), public autonomy is the freedom acquired when citizens understand themselves as both the authors and the addressees of the law (Habermas 1996: 104, 120, 454). He claims, however, that discourse theory enables us to make better sense of this intuition than either Kant or Rousseau. For this purpose, he introduces, in a first step, a general discourse principle, which states: ‘Just those action norms are valid to which all possibly affected persons could agree as participants in rational discourses’ (Habermas 1996: 106). In contexts of democratic deliberation, where rational discourses are concerned with action norms in legal form, the discourse principle takes on the specific character of a principle of democracy. Democratic legal norms are justified by balancing moral, ethical-political and pragmatic reasons; in this justificatory context, the three moments of practical reason (Habermas 1993) are systematically interconnected. In consequence, their validity (legitimacy) is determined not by the moral principle of universalizability, ‘U’, but by the democratic principle of general rational acceptability (Habermas 1996: 108). According to the principle of democracy, ‘only those statutes may claim legitimacy that can meet with the assent (Zustimmung) of all citizens in a discursive process of legislation that in turn has been legally constituted’ (Habermas 1996: 110). What Habermas calls public autonomy (sometimes referred to as ‘civic’ autonomy, ‘legal’ autonomy or ‘political’ autonomy) is the freedom of citizens who engage in collective processes of public deliberation concerning the rationality of general or particular legal norms and public policies. It refers both to the exercise of such citizen engagement, and the valid laws and policies that result from it and is effectively synonymous with popular sovereignty.

There is an evident analogy between Habermas’ conception of public autonomy and his conception of moral autonomy. Both share the three dialogical features I identified earlier. Like moral autonomy, public autonomy is a social form of freedom, with no pre-political origins; it is constitutively intersubjective, coming into existence by way of interactions between human subjects in processes of deliberation; it is processual, a form of freedom that is generated within processes of intersubjective deliberation. In addition, like moral autonomy, public autonomy is construed cognitively and, apparently, epistemically (Cooke 2013: 265–74). Just as individual human subjects acquire moral autonomy by way of a discursive process that enables them to bind themselves freely to insights of practical reason, they acquire public autonomy by way of a discursive process that enables them to bind themselves freely to formal principles concerning the exercise of power, which are likewise understood as insights of practical reason (Habermas 1996: 121; 2008: 77–9).

However, the evident analogies between public autonomy and moral autonomy do not amount to an argument for their co-originality: demonstration of an analogy does not establish a reciprocally constitutive relationship or even a relationship of mutual dependency. Indeed, it is far from clear that Habermas seeks to show the co-originality of public autonomy and moral autonomy. Although, as mentioned, in his initial discussion of Kant he uses the term private autonomy to refer to moral autonomy, for the most part when he speaks of private autonomy he means subjective freedom of choice (Willkür). Moreover, even if he were concerned to establish the co-originality of public and moral autonomy, his conceptualizations of each would prevent him from doing so. This is due to the fundamental difference in the principles of justification operative in political-legal discourses and in moral discourses. As we have seen, in political-legal discourses, which are the site for the constitution of political autonomy, a context-bound democratic principle that balances moral, political-ethical and pragmatic reasons determines the validity of laws: here, rational acceptability refers to an agreement reached by all citizens within a particular jurisdiction. In moral discourses, which are the site for the constitution of moral autonomy, a context-transcending moral universalization principle (‘U’) determines the validity of norms and principles; here, rational acceptability refers to an agreement reached by all concerned that the interests in question are universalizable, whereby no human subject may be excluded in principle from participation in the deliberative process. In consequence of these two fundamentally different principles of justification, there can be no reciprocally constitutive relationship between political and moral autonomy. Democratically valid reasons cannot ground moral validity, for in moral discourses ethical and pragmatic reasons have no bearing. On the other side, moral reasons are insufficient for the purposes of democratic validity, for here ethical and pragmatic reasons also come into play. For example, a democratically valid legal ordinance temporarily restricting the movement of citizens outside their places of residence during a pandemic, which is based on a complex of moral, ethical and pragmatic reasons such as equal respect for persons, an implicit or explicit view of what a good life for humans consists in, health and safety considerations, an assessment of the kinds of restrictions all citizens are likely to accept temporarily, etc., can make no claim to moral validity. Conversely, the moral principle of equal respect for persons is just one – albeit crucially important – factor in determining legal-political validity.Footnote ⁹ Habermas repeatedly emphasizes this difference between law and morality and, by extension, political and moral autonomy. Thus, he criticizes theorists such as Robert Alexy, who construe the claim to legal validity as a claim to moral validity (Habermas 1996: 230–3; cf. Cooke 2007). For the same reason, he would have to reject Rainer Forst’s version of the co-originality thesis.

Forst offers a variation on the co-originality thesis that provides a unified normative ground for private and public autonomy (Forst Reference Forst2012: 113–16), aiming thereby to remedy what he sees as the insufficiently robust normative underpinning of Habermas’ version (Forst Reference Forst2016: 20).Footnote ¹⁰ The unified ground is a moral right to justification. This is based on the concrete respect individuals owe to themselves and others as equal normative authorities, and institutes citizens collectively as the normative authority for decisions and actions that are valid in political-legal contexts.Footnote ¹¹ It calls for procedures of democratic law-making in which the claims and arguments of all those addressed by laws can adequately be raised and considered according to norms of generality and reciprocity. While Forst, as I argue in my larger project, succeeds in establishing the mutually constitutive relationship between public and private autonomy, the cost of his ultimately moral approach is a depoliticization of ethical concerns, giving rise to the likelihood of citizen disaffection; a related cost is that it makes politics inhospitable towards ethical learning, thereby diminishing the vitality of the political public sphere. From Habermas’ perspective, the cost is a problematic subordination of law to morality.

Although Habermas does not establish a reciprocally constitutive relationship between public and moral autonomy – and to all appearances, does not seek to do so – he does establish a reciprocally constitutive relationship between public autonomy and human rights (Habermas 1996: 101, 103). In this case, by contrast with public and private autonomy, their mutual dependency entails a relationship of mutual constitution. However, while he treats the internal connection between public autonomy and rights as synonymous with the sought-for internal connection between public and private autonomy (Habermas 1996: 104), it is not. Let us look at this part of his argument more closely.

The argument runs as follows: ‘there can be no law at all without actionable subjective liberties that guarantee the private autonomy of individual legal subjects; and no legitimate law without democratic law making by citizens in common who, as free and equal, are entitled to participate in this process’ (Habermas 1995: 130). From this we can see (i) that democratic self-legislation (public autonomy) is dependent conceptually on citizens who are free and equal citizens (the bearers of subjective rights) and (ii) that participation by free and equal citizens (the bearers of subjective rights) in the process of democratic law-making contributes to the constitution of valid law (public autonomy). The specific contribution with which Habermas is concerned is to the constitution of a system of subjective rights. The reciprocally constitutive relationship between public autonomy and a system of rights is elaborated in his justification of rights. The basic premise of this part of his discourse theory is that the institutionalization of the legal medium merely establishes the legal code: it does not yet determine the familiar liberal basic rights (such as rights to personal dignity, to life, liberty and bodily integrity, and to freedom of movement). As he puts it: ‘the basic rights inscribed in the legal code itself remain unsaturated … They must be interpreted and given concrete shape by a political legislature in response to changing circumstances’ (Habermas 1996: 125).

However, the reciprocally constitutive relationship between public autonomy and subjective rights does not establish an evenly balanced relationship of mutual dependency between public autonomy and private autonomy. His argument shows how public autonomy contributes to the constitution of particular subjective rights such as rights to dignity, freedom of movement, etc., but it also shows that it may leave the substance of private autonomy untouched (Willkür). Indeed, this appears to be his position: public autonomy, by constituting certain subjective rights, formally guarantees private autonomy and provides it with a protective cover, but it does not impact on the substance of citizens’ choices and decisions.

I have argued that the ‘even weight’ or ‘balance’ component of Habermas’ co-originality thesis requires him to establish the reciprocal constitution of private and public autonomy. We can now see that, in order to do so, Habermas would have to provide an account of ‘private autonomy’, the further development of which depends on collective citizen participation in processes of democratic legislation and decision-making. I suggested that my idea of ethically self-determining agency, together with my account of non-authoritarian institutions, would enable him to do so. This is because like Forst and Alexy, I posit a common normative ground between the individual and communal moments of freedom (between what Habermas calls private and public autonomy). Unlike Alexy and Forst, however, the normative ground I posit is ethical rather than moral. Habermas is suspicious of what he calls the ‘ethical constriction’ of political discourse (Habermas 1996: 279, 283). Thus, he is likely to be wary of attempts to give an ethical reading of the reciprocally constitutive relationship between public and private autonomy and to find my proposal uncongenial. But we have also seen that he rejects attempts to give a moral reading of the relationship. He seems to be caught between a rock and a hard place. Appeal to a traditional liberal conception of private autonomy, enabled and protected by a system of rights, is not a viable solution to this difficulty, if he wishes to uphold his thesis of an evenly balanced, internal relationship between private and public autonomy. What he requires is a conception of private autonomy, the further development of which depends on its interplay with public autonomy (and vice versa). His dialogical recasting of the Kantian idea of moral autonomy harbours the promise of such a conception. It is in this spirit that I propose a dialogical recasting of ethical autonomy. However, Habermas does not take this step. Were he to do so along the lines I suggest, an ethical reading of the co-originality thesis might look more attractive. His worry about the ‘ethical constriction of politics’, which he shares with Forst (Forst Reference Forst2012), is that an ethical reading of freedom in a legal-political context permits the imposition, by socially dominant groups, of particular visions of the good life on all citizens, violating their moral autonomy (Forst) or their self-determining agency/free choice (Habermas). In the case of the ethical model I propose, this worry is unfounded. It is addressed by my processual interpretation of the political common good as perpetually under construction by the ethically self-determining agents for whom it is good, and by the corresponding conception of political authority as in significant measure constituted on an ongoing basis by all citizens, whose ethically self-determining agency, in turn, is significantly constituted by political authority. The advantage of an ethical reading, acknowledged implicitly by Habermas when he includes ethical and pragmatic as well as moral reasons in his account of the normative grounds of democratic legitimacy, is that it conceives of valid laws and political policies as matters of importance for citizens as particular individuals in particular situations, who in addition to being motivated by particular egocentric or strategic interests, are deeply attached to particular beliefs and values. It is precisely because laws and policies on the one side, and the beliefs and values of citizens on the other, are matters of subjectively felt deep concern, that each side must relate dialectically to the other, open to further development and transformation as a result of the encounters. As things stand, this deep concern, which is ethically grounded, is missing in Habermas’ account of the relationship between private and public autonomy.Footnote ¹²