Equal Societies, Autonomous Lives: Reconciling social equality and relational autonomy

Relational egalitarians flesh out the idea of what it means to treat persons as equals in society. On many influential accounts, relating as equals requires, among other things, to respect others as equal, autonomous agents (Hojlund, 2021; Scheffler, 2015; Schemmel 2021a; Schuppert, 2015). However, the attempt to subsume respect for personal autonomy under social equality is prima facie suspect: people can presumably endorse social inequalities. We can ask two sets of questions when confronted with situations such as these: first, are these truly autonomous choices? For instance, following substantivist approaches to autonomy, some may question whether choosing subservience or a subordinate status can be an autonomous choice because one would thereby fail to be self-respecting, to see what is effectively in their own interest, or because this choice is often made against a social background that diminishes their available options or controls over their life (Babbitt 1993; Hill, 1991; Oshana, 2006). And, second, these types of cases raise a challenge for relational egalitarians in that it is unclear what it means to treat presumably autonomous agents as equals here. On one hand, respecting their autonomy requires that one should respect their decisions to endorse social inequalities. On the other hand, relational egalitarians should be capable of criticizing social hierarchies and should aim to equalize them. How, then, can we resolve this apparent tension? In this article, I show how debates between relational autonomy theorists hold important lessons for relational egalitarians. The connections between the two theoretical families have not been extensively studied (for recent exceptions, see Stoljar Mackenzie Meyers, 2002, 2005; Westlund, 2009). Similarly, relational egalitarians argue that egalitarians should be concerned first and foremost with how people are treated and regarded in society (Anderson, 1999; Fourie et al., 2015; Lippert-Rasmussen, 2018; Schemmel, 2021a). I argue that by connecting social equality and relational autonomy, it is possible to resolve this apparent tension between respect for autonomy and the protection of social equality. More precisely, I argue that relational egalitarians should adopt a constitutivist, externalist understanding of autonomy. I point out that a constitutivist, externalist understanding of autonomy is not designed to evaluated particular individual decisions, but rather to identify the required external conditions to guarantee a substantive level of personal autonomy. The externalist approach consequently allows to say both that some non-egalitarian choices can be deserving of respect and that egalitarian should adopt a structural perspective to promote personal autonomy globally. Below, I first briefly introduce relational egalitarianism and relational autonomy. Second, I distinguish between three ways of conceptualizing the connection between social equality and autonomy: the instrumental approach, the deontic approach, and the constitutivist position. I argue that both the instrumental and deontic approaches fail to provide convincing answers to the tension between autonomy and (social) equality. Consequently, I show that if relational egalitarians want to argue that social equality requires, among other things, to treat autonomous agents as equals qua autonomous agents, they should adopt a constitutive, externalist conception of autonomy. Relational egalitarianism and relational autonomy share many aspects. As Stoljar and Voigt (2021b) points out, first, they both consider that theories of justice "should take the fact of people's social embeddedness as their starting point" (2021b, 1). Second, for both relational egalitarianism and relational autonomy, they do not refer to single theories but are "umbrella terms that encompass different conceptions of the way in which social conditions should be included in an analysis of autonomy and equality" (2021b, 1). However, both theoretical families have fundamentally different aims. Relational autonomy theorists identify the conditions under which an agent is autonomous. In contrast, relational egalitarians identify the conditions ensuring that people are treated as equals in society. Nonetheless, given their family resemblance, their connection is fruitful especially since many relational egalitarians presume that social equality requires, among other things, that autonomous agents should be treated as equals in society. Moreover, beyond this family resemblance, the move toward relational autonomy provides relational egalitarians with precious resources to think through what it means to respect and foster personal autonomy in society. This, in turn, enables us to clarify what it means to treat persons as equals. There are two things to note here, however. First, this is not to say that autonomy is the only reason explaining why social equality is required or valuable. The claim here is simply that since many relational egalitarians aim to respect autonomous agents equally, it is relevant to explore what adopting a relational conception of autonomy would mean for our conception of social equality. Second, it does not follow that we should treat nonautonomous or nonagential persons as unequals in society (see Bengtson Lippert-Rasmussen, 2022).1 Relational egalitarianism emerged in reaction to the perceived dominance of distributive egalitarianism in political philosophy (for general overviews, see Fourie et al., 2015; Nath, 2020; Voigt, 2020). Broadly, relational egalitarians argue that equality is not reducible to distributive concerns. For them, equality is about how social, political, and interpersonal relationships are structured. Relational egalitarianism comes under negative and positive forms. Negatively, it states that certain types of relations are incompatible with the equal social status of all. More positively, it highlights that certain background conditions must be met to ensure that persons are treated as equals and relate to one another as such. There are debates concerning which inegalitarian relationships ought to be criticized and, conversely, which relationships should be equalized. However, relational egalitarians are typically concerned with self-regarding attitudes, interpersonal relations, and institutional structures. Self-respect and the ability to regard oneself as an equal is often presented as deeply connected to relational egalitarianism (Schemmel, 2011, 2019; Stoljar Wolff, 2015). For instance, Stoljar and Voigt argue that to be able to stand as an equal, it is necessary to see oneself as an equal: a requirement which has both institutional and interpersonal, affective dimensions (Stoljar Schemmel, 2011; Viehoff, 2014). Second, political institutions are themselves required to follow some relational requirements both because the actions of a democratic state reflect how citizens consider one another (Anderson, 1999, p. 314), and because the state is an agent to whom relational requirements apply directly (Schemmel, 2012). This may require, for instance, that states be equally responsive to the interests of all their citizens, that they be accountable to them, and that they be in a position to justify their decisions adequately. Third, and finally, political institutions may also be evaluated based on the attitudes they express vis-à-vis citizens (Anderson Dworkin, 1977; Schemmel, 2012; Schemmel, 2021a; Voigt, 2018). Therefore, relational egalitarianism points toward a very broad understanding of equality that includes three different levels: how persons relate to themselves, how they relate to each other, and how they relate to socio-political institutions. However, as will be discussed throughout, it is then unclear how relational egalitarians can effectively reconcile these different levels. Ensuring that states treat all as equals seems to require respecting that individuals may endorse interpersonal social hierarchies and consider themselves as unequals. In a manner akin to the debate between relational and distributive egalitarians, relational autonomy was developed in response to the perceived dominance of an overly rationalistic and atomist models of autonomy such as the ones associated with Kant (1948 1785); Rawls (1999 1971); or Hobbes (1966 1642) (for overviews, see Lee, 2023; Stoljar, 2018b). A recurring argument shared by relational autonomy theorists is that this move is inherently problematic: we cannot abstract away from the social embeddedness of the self and presuppose that agents are isolated, self-sufficient beings. Rationalist and atomist approaches fail to capture how our projects and ambitions are informed by our social context; how social norms can impact our capacity to be autonomous and our available choices; and, more deeply, how some relationships may be constitutive of personal autonomy (Stoljar, 2018b).2 Broadly speaking, theories of relational autonomy are relational due to three shared assumptions. As Mackenzie (2014, pp. 21–22) points out, relational autonomy theorists reject the view that persons are purely rational contractors, self-sufficient, or independent, but rather accept as a given that individuals are socially embedded, dependent on one another, and thus vulnerable. Second, this social, cultural, and historical embeddedness means that a person's identity does not develop in the abstract, but is affected and even perhaps constituted by the relations they engage in. Third, these considerations support the general conclusion that "social conditions restricting the exercise of self-determination are unjust" (Mackenzie, 2014, p. 22).3 In other words, by giving a greater weight to human vulnerability and our embedded social identity, Mackenzie underlines that relational autonomy theorists develop analyzes that are responsive to the negative effects of unjust social relations such as unjust social hierarchies of standing, authority, and esteem that track people in different spheres of life (within the family, in the workplace, in civil society, etc.). In that sense, for Mackenzie, relational autonomy theories play the role of diagnostic tools that highlight how "social domination, oppression, stigmatization, and injustice can thwart individual autonomy" as well as to propose solutions on how social relations could be reformed to protect and foster personal autonomy (Mackenzie, 2014, p. 23).4 Therefore, at first sight, relational egalitarianism and relational autonomy should be mutually supportive. Both theoretical families tend to be deeply concerned with intrapersonal attitudes, interpersonal relationships, and socio-political relations. There is, of course, a debate between internalist and externalist theories of autonomy—that is, on the question of whether autonomy can be captured by considering the internal structure of a person's will or psychology only, or if we should also consider external social and political structures (for a recent overview, see Mackenzie, 2021, pp. 33–39). However, some elements of self-authorization, self-respect, and the presence of interpersonal relationships that foster our individual capacities for agency are typically central to relational autonomy. However, as discussed in Section 2 below, despite this initial resonance between the two theoretical families, some tensions remain between equality and autonomy on many accounts of relational autonomy. is the bridge between liberal insistence on respect for personal autonomy and egalitarian conceptions of social justice in much of contemporary liberal egalitarianism: it is the former that ultimately underpins the latter, (…) by continuously ensuring (roughly) equal substantive liberty to live one's life as one sees fit. (Schemmel 2021a, p. 9) However, it is somewhat unclear how we should understand this appeal to agency rather than personal autonomy in the work of Anderson and Schuppert.5 Agency is typically understood as the capacities to plan, weigh competing options, be rational and be capable of self-control (Stoljar, 2014, pp. 246–249; Stoljar 2018a; Mackenzie, 2015). However, autonomy requires more. What autonomy requires depends on the theory of (relational) autonomy one prefers. Roughly, proceduralists add authenticity conditions to agential capacities—such as reflexive endorsement (Friedman, 2003). Weakly substantive approaches maintain that autonomy also includes some weak normative content, like self-regarding attitudes such as self-respect (Benson, 2005a; Benson, 2011; Meyers, 1989). And externalist conceptions maintain that autonomy should moreover include an adequate range of external options and conditions (Mackenzie, 2021; Oshana, 2006; Stoljar, 2018a). Stoljar takes the example of Jehovah's Witnesses imprisoned in concentration camp in Nazi Germany to illustrate the importance of external options. Many imprisoned Jehovah's Witnesses continued to live according to their religious convictions despite imprisonment (Stoljar, 2018a p. 231; see also Bettelheim 1960, pp. 122–123). In this case, the Jehovah's Witnesses remained agents, and were even autonomous in some sense; they continued to live an authentic moral life. Nonetheless, to say that they were fully autonomous would miss how one essential aspect of their autonomy was infringed on: as political prisoners, they lacked access to a set of adequate external options (Stoljar 2018a, p. 244ff; see also Raz 1986, pp. 373–374). Accordingly, for externalists, authenticity and weak normative content is insufficient to consider that an agent is substantively autonomous. to be capable of functioning as an equal citizen involves not just the ability to effectively exercise specifically political rights, but also to participate in the various activities of civil society more broadly, including participation in the economy. And functioning in these ways presupposes functioning as a human being. (…) To be capable of functioning as a human being requires effective access to the means of sustaining one biological existence (…) and access to the basic conditions of human agency—knowledge of one's circumstances and options, the ability to deliberate about means and ends, the psychological conditions of autonomy, including the self-confidence to think and judge for oneself, freedom of though and movement. (Anderson, 1999, pp. 317–318) Similarly, Schuppert discusses gender hierarchies to illustrate what it means to stand as an equal (Schuppert, 2015, p. 123). He writes that social equality is violated by gendered hierarchies because: "repeated instances of gendered misrecognition can both undermine a person's self-respect and lead her to adopting 'conforming' behavior in line with existing biases" (Schuppert, 2015, p. 123). He goes on to argue that subscribing to a gendered hierarchy can lead to "self-respect undermining adaptive behaviour." Of course, I cannot here delve into the debate on adaptive preferences, since this would take me too far from the main question at hand. However, notice that standing as an equal seems to come with substantive constraints on what can be considered to be an autonomous, authentic behavior for Schuppert. However, this attempt to subsume (2) under (1) can lead to puzzlement. People can presumably choose to live non-egalitarian and hierarchical lives autonomously. Respecting personal autonomy would then require accepting that certain social hierarchies may be legitimate if they are autonomously chosen, which sits uneasily with the idea that equality requires that all be treated as equals. Indeed, as shown above, the position of Schuppert supports the conclusion that some social hierarchies are incompatible with personal autonomy and social equality. However, this may be too hasty: after all many have shown that social relations are deeply connected to personal autonomy. Accordingly, it is fruitful to consider the different ways in which it is possible to connect social equality and relational autonomy to see whether it is possible to resolve the apparent tension between SEC and AC. There are at least three ways to connect social equality and relational autonomy. In this section, I consider each in turn. I start with the instrumental approach. Second, I consider procedural and weakly substantive approaches. Third, I evaluate constitutivist positions. I argue that only the latter are promising to resolve the tension between SEC and AC. A first way to connect both social equality and relational autonomy would be to say that social equality is instrumentally valuable at least in part because it promotes autonomy. Stated negatively, this claim holds that social inequalities are wrong at least in part because they negatively affect personal autonomy by, for instance, damaging self-regarding attitudes or one's ability to reflect critically. Some claims by relational autonomy theorists support this point (Johnston, 2021, p. 139; Mackenzie, 2014, p. 18). Consider, for example, the work of Govier (1993) and Benson (1994). They highlight that for a decision to truly count as one's own, it has to be grounded in certain positive attitudes toward one's capacity to reflect critically (Govier, 1993, p. 101; see also Cudd, 2006, pp. 176–178). This supports the idea that social relationships that damage some self-regarding attitudes could be objectionable given that they undermine or prevent autonomy. Although he focuses on agency rather than personal autonomy, the position of Litalien (2021) is helpful to consider how this argument could play out when it is extended to personal autonomy. He argues that social inequalities are wrong at least in part because they affect individual agency negatively. He defines agency as a two-pronged capacity for intentional action: "(1) a capacity for forming intentions (and other intentional states) and (2) a capacity to carry out those intentions, to have an effect in the world" (Litalien 2021, p. 89). Although he is open to the possibility that the wrongness of relational inequalities may be overdetermined (Litalien 2021, p. 87), he maintains that the instrumental account is the best account to explain why social inequalities are wrongful. we can hypothesize that, globally, relational equality promotes agency, even if it might be true that, locally, certain inegalitarian relations might promote agency better than relational equality would in the same context. This implies that it would be a mistake, as a matter of political philosophy, to assess each relation individually. (…) We should (…) favor the policy (or policy package) that promotes agency globally. What I suggest here is that it is a policy (or policy package) that aims at relational equality that will best further that goal. (2021, pp. 92–93). To illustrate (3), one could think of a society of benevolent masters where society is divided into two groups such that one group, the "carers," are entrusted with the care of the members of the second groups, the "carees," unbeknownst to the carees. The carers are respectful, they consider the strong interests of the carees, and the carers act to promote the well-being of the carees when they can. To do so, the carers can paternalistically intervene in their carees choice without their carees' knowledge when they judge that it would be better for the carees to purse some options compared to others to ensure that they will be better able to carry out their intentions in the future. If it turns out that by carefully separating the carers and the carees such that the position of the carees is, in fact, improved in the sense that they are better able to have an effect in the world than they would without the carers, the situation arguably remains objectionable from a relational egalitarian perspective (and, in fact, from the perspective of many autonomy theorists such as externalists). It remains that the carers that are imposed on the carees can dominate them. However, from an instrumental perspective, we should prefer policy package (3). If we accept the claims that social inequalities are wrong to the extent that they negatively affect individual agency and that we only have an indirect reason to aim for social equality because egalitarian relations generally promote individual agency, then we have no principled reason to prefer (2) over (3). However, as mentioned, Litalien is more interested in agency rather than autonomy. Could someone object that once we move from agency to autonomy, then we would have to prefer (2) over (3)? This of course would depend on the particular conception of personal autonomy that one adopts. However, from this detour through an agency, the instrumental approach seems like an unpromising avenue to try and reconcile social equality and personal autonomy. From a purely instrumental perspective, as argued, nothing prevents us from preferring a policy package that is partially egalitarian, but which also bends the rules when we have good reasons to believe that putting the demands of social equality aside will foster personal autonomy. At best, it only provides us with a contingent defense of relational equality; from an instrumental perspective, aiming to protect and foster agency and aiming for social equality are simply two distinct goals, which can come apart and conflict with one another, although they may happen to coincide in practice. Accordingly, we have good reasons to look beyond instrumentalism to consider how social equality and personal autonomy can work in tandem. A good way to go about is to consider the different ways in which autonomy theorists have conceived of the importance of social relations for personal autonomy. A topic to which I now turn. Consequently, it is warranted to consider other ways of connecting social equality (and social relationships generally) and personal autonomy. Two options are accessible: one may contend that the connection is noninstrumental in nature, or that some socio-political relations are constitutive of it. Starting with the first option, procedural and weakly substantive accounts (Friedman, 2003; Meyers 1989; Meyers, 2002, 2005; see also Christman, 2009) point toward the idea that the relation between equality and autonomy is not only instrumental. Briefly, procedural approaches are content-neutral—they do not put "direct normative constraints" (Benson, 2005b, p. 133) on the content of an agent's preferences—and they insist on how preferences are formed, adopted or reflexively endorsed. Autonomous agents are agents who reflect critically about their preferences, motives, and desires in such a way that they can consider them to be their own. Weakly substantivist approaches, on their end, also do not put "direct normative constraints" on the content of an agent's preferences, but add some normative content to their conception of autonomy, such as requiring that autonomous agents have some moral attitudes like self-respect (Govier, 1993; McLeod 2002; Benson 2005b; Richardson 2001; for an overview, see Stoljar, 2018). Following these procedural and weakly substantive approaches to autonomy, one could move beyond instrumentalism to highlight that the connection between equality and autonomy is a deontic constraint where some norms simply delineate how we ought to treat autonomous agents in society. As Miklosi puts it, from a deontic perspective, the significance of egalitarian relations is not situated in "their contribution to personal or impersonal value, but in their being the fitting response to the equal moral status of each person as a being with a certain kind of moral authority. On this view, equality in relationships is significant not (or not only) as something to be promoted, but as a constraint, something that is to be respected" (Miklosi, 2018, p. 131).6 Of course, here Miklosi is interested in the equal moral status of all rather than autonomy. Nonetheless, a deontic approach could be used to underline that respect is owed to autonomous agents and to their autonomous choices (pro tanto). Following this deontic approach, social equality would not (only) be instrumentally useful to promote social equality, but it dictates how we should treat autonomous agents and autonomous choices. In a word, it requires respecting the choices of autonomous individuals, even if they are substantively non-egalitarian. Below, I consider procedural and weakly substantive views in turn to show how they resonate with this deontic requirement. To respect autonomous agents and their autonomous choices would simply be the fitting way to treat others as equals, even if their choices are not substantively egalitarian. Autonomy has something to do with the respect owed to persons as persons. Many philosophers agree that respect is owed to persons simply by virtue of their potential for being autonomous, whether or not this potential is ever actualized. Nevertheless, someone's actual manifestation of autonomy may warrant yet another form of respect, also connected to sheer personhood. (…) On a content-neutral view, showing respect to people's actual choices is (…) owed generally to the choices made by anyone with the competency to choose and act self-reflectively; what she chose would not matter. (Friedman, 2003, p. 22) the notion of self-determination has been elaborated in terms of a certain sort of self-reflection that involves, one might say, self-monitoring and self-regulation. According to a generic version of this view, to realize autonomy a person must first somehow reflect on her wants, desires, and so on and take up an evaluative stance with respect to them. She can endorse or identify with them in some way or be wholeheartedly committed to them, or she can reject or repudiate them or be only halfheartedly committed to them. (…) When she chooses or acts in accord with wants or desires that she has self-reflectively endorsed, and her endorsement is somehow a part cause of her behavior, then, (…) she is behaving autonomously. (Friedman, 2003, pp. 4–5) To clarify, although the agent choosing to be subservient is, in a sense, less autonomous than others, we can still say that they are autonomous if they have the sufficient autonomy competencies, a concept introduced by Meyers (1987, p. 619; 1989, p. 208; Meyers, 2005, p. 49).8 Even if some can be "more" autonomous, by having more available options, for instance, individuals are still autonomous agents when they meet the specified threshold. What they choose to be or do does not change the fact that they could be autonomous agents (depending, of course, on the details of the case). She buys the clothes he prefers, invites the guests he wants to entertain, and makes love whenever he is in the mood. She willingly moves to a new city in order for him to have a more attractive job, counting her own friendships and geographical preferences insignificant by comparison… She does not simply defer to her husband in certain spheres as a trade-off for his deference in other spheres. On the contrary, she tends not to form her own interests, values, and ideals, and when she does, she counts them as less important than her husband's. (Hill, 1991, p. 5; see also Westlund, 2003, pp. 485–486; Friedman 1985) This procedural account resonates strongly with the liberal approach to social equality. This approach is most clearly developed by Schemmel (2021a). He subscribes to a liberal view of justice explaining why certain socio-political relations are unjust but abstains from telling individuals how they should live their lives (Schemmel, 2021a, p. 5). Hence, Schemmel subscribes to the idea that relational egalitarianism, when it is understood as a theory of justice, does not require to equalize any and all relationships: some interpersonal relations that depart from the (social) egalitarian ideal are not unjust on this view (Schemmel, 2015, p. 153). From a liberal standpoint, since justice is concerned with how to organize the "institutional set-up of society," then it is not unjust for people to enter unequal relationships, which lie outside the basic structure. Following this deontic understanding of social equality, it means that relational egalitarians should promote the relations necessary to ensure that all have the opportunity to develop the relevant autonomy competencies—by, for instance, providing free, public education—and should otherwise protect the rights and freedoms of all equally without promoting certain ways of life. Here to treat others as social equals is reduced to equally respecting the autonomous choices of all (pro tanto): we should respect the individual choice to endorse and act on a plan of life, but should otherwise remain largely neutral concerning the content of these plans of life. However, this position would render relational egalitarians vulnerable to a charge often raised against proceduralist accounts of relational autonomy: this content-neutral approach is too limited to fully capture the ways in which different oppressive social norms can be problematic (see notably Benson, 1991; Mackenzie, 2021; Stoljar, 2018b). This deontic position creates an i