How Progressive are Multidimensional Accounts of Autonomy? Transnational Feminist-Friendly Amendments and a Critical Intersectional Relational Autonomy

Abstract

“Multidimensional” accounts of autonomy offer multiple, rather than unitary, dimensions along which to measure autonomy. Such accounts are significant developments in the literature and help generate nuanced, degree-based frameworks. However, transnational feminists—those theorizing feminism in light of (post)colonialism and global neoliberalism—may raise concerns about multidimensional accounts for women in the Global South. For instance, there may be worries about the generalized focus and implicit individualism that still lurks. Sympathetic to both non-unitary autonomy and transnational feminist projects, I argue that multidimensional accounts can be salvaged from such critiques with two amendments. First, they can adopt what I label a “critical” relational framing, and second, they can include intersectional identities. Using commercial surrogates in India as an example, I show how these amendments to multidimensional theories might better serve these women of color, and indeed all persons.

Autonomy is a widely invoked conceptual and practical tool in feminism. Since women have long been denied autonomy, one feminist aim is to show that women are autonomous and also—importantly—to recognize oppressive forces that hinder their autonomy (Mackenzie and Stoljar 2000). Some recent accounts capture these competing considerations by being “multidimensional,” offering three (Mackenzie 2014) or four (Killmister 2018) measures by which to assess autonomy. From a feminist perspective, there is much to support in these groundbreaking theories with their various axes that plot different aspects of autonomy in degrees.

However, while welcome, these theories do not go as far as they might when we consider transnational feminist insights. Just as transnational feminists have voiced concern relating to women of color in general feminist theory (Narayan 1998; Mohanty 2003; Jaggar 2005; Weir 2017; Khader 2019), I argue in this paper that similar worries extend to multidimensional autonomy theories. These include (a) theorizing in a generalized and a primarily Western way, and (b) a lingering abstraction and individualism. As a result, I argue that multidimensional autonomy is inadequate from a transnational feminist perspective. Such theories can, however, be redeemed.

To do this, I offer a “transnational feminist-friendly” reframing of multidimensional accounts of autonomy. I argue that the problems for the accounts are reduced by amending them so that they (a) have a “critical” relational component and (b) include intersectional identity in their theories. While not all transnational feminists will unanimously endorse them, these proposals do uphold key considerations of transnational feminisms.

First, I show that a critical relational approach can integrate structural relations into autonomy in a way that responds to criticisms raised by some transnational feminists. This is by taking seriously the relational underpinnings of transnational feminisms and by targeting our theoretical gaze on how to improve relations. The critical relational proposal advocates more fully a relational ethos, thereby addressing the problem of individualism.

Second, I show how intersectionality emphasizes particularities between and within different groups, including in “doing” autonomy. Those with intersectional identities occupy spaces between oppression and power, have unique knowledge about bolstering their conditions, and manifest autonomy in a multitude of ways not confined to pre-defined Western modes. The intersectional identities proposal adds specificity, thereby addressing the issue of the generalized Western woman slipping into theory.

If my analysis is correct, reframing multidimensional accounts in these two ways would make them more accurate and useful (Narayan 1998) to women of color, and to all persons. Indeed, the same is true of any account of feminist autonomy. Further, my argument allows a new direction to some transnational feminist positions. It expands current theories that are more freedom, rather than autonomy, focused (Weir 2008, 2013, 2017), and it responds to those that reject these kinds of multidimensional approaches altogether (Khader 2020).

I proceed in four sections. First, I outline multidimensional autonomy and its strengths. Second, I discuss criticisms of the approach. Third, I offer guidelines for “transnational feminist-friendly” amendments. Finally, I present the positive proposals and apply them to one practice discussed by transnational feminists: commercial surrogacy. This is to show how critical intersectional relational autonomy is not a mere abstraction but has concrete benefits for those it hopes to serve.

Multidimensional autonomy and its benefits

Multidimensional accounts, with their various axes or dimensions, offer a novel way in which to conceptualize autonomy.¹ Take theories by Catriona Mackenzie (2014) and Suzy Killmister (2018). Though by no means the only such accounts, both are explicitly multidimensional, relatively recent in the literature, compellingly argued, and speak to a range of feminist concerns. As such, they serve as representative examples.

Mackenzie, for instance, argues that autonomy has three axes. The first axis, self-determination, captures external and structural conditions for autonomy. This includes freedoms, such as political or personal liberties, and freedom from harms. However, Mackenzie—following capability theory—is especially interested in opportunities to be, do, and value. The second axis, self-governance, measures internal skills and capacities to make choices in line with one's practical identity. In particular, self-governance refers to common notions of competence and authenticity in autonomy. The third axis, self-authorization, identifies normative conditions about agents. Such conditions are threefold: (i) answerability or responsibility for oneself (accountability); (ii) favorable attitudes, such as trust, respect, and esteem, towards oneself (self-evaluation); and (iii) being regarded by others as having social standing as an autonomous agent (social recognition).

Killmister, on the other hand, proposes four dimensions of autonomy. The first dimension, self-definition, captures “commitments” (to certain goals, values, beliefs) that are part of one's personal identity. When an agent forms a commitment to being or doing x, then x has normative force for that agent. The second axis, self-realization, is a type of practical agency. This represents how far one deliberates, forms intentions, and then intends to perform actions to discharge one's commitments. The third axis, self-unification, assesses whether personal identity and practical agency coalesce and do not contradict in obvious ways. This axis goes beyond mere deliberation, intention, or action (the focus of the first two axes) as it records whether our commitments cohere with our embodied or phenomenal experiences. Last, self-constitution is about the person as a whole. It identifies whether individuals are beings that take on their own commitments at all. It also ascertains whether dimensions 1–3 are performed to a basic extent, such that there is minimal autonomy.

I take multidimensional accounts to be helpful in at least three ways from transnational feminist perspectives.² First, multidimensional accounts are not excessively demanding in who counts as minimally autonomous, which avoids ruling many people out as non-autonomous. For transnational feminists, this is positive for supporting the intuition that those living under constrained conditions can still have autonomy (Narayan 1993; Mohanty 2003; Weir 2013; Khader 2019; Dokumacı 2020).³ Recognizing that women in the Global South are autonomous—which is still often denied—has long been the calling card of transnational feminists.

Under Killmister's self-constitution (axis iv), for instance, a woman who takes up undesirable work could be someone who meets minimum autonomy requirements. Such a woman could form commitments, deliberate, intend, act, and be unified to some extent (2018, 69). Critically, we could even suppose her to be deferential to her husband and take on his commitments about going to work rather than forming her own. For Killmister, being deferential may itself be a value she commits to, and deliberation does not need to be very deep (72). If the woman defers in this way, she still passes the self-constitution threshold. Similarly, under self-realization (axis ii), some autonomy in oppression is possible. The woman may be “significantly unfree” (51) because of limited options, poverty, and sparse employment prospects without failing to put her intentions into action. For Killmister, she succeeds at their enactment so as long as she does not intend something that the constraints of her situation make impossible.

Mackenzie's model, likewise, recognizes that autonomy can be present under oppression just as it can be absent in relative freedom. For example, it allows that at least some individuals can be self-governing (axis ii) and self-authorizing (axis iii) in oppressive contexts (axis i), while others can be free and have many opportunities (axis i) without meeting self-governing (axis ii) or self-authorizing (axis iii) conditions (2014, 40). Mackenzie acknowledges that some minimal level of achievement is required under each dimension, if autonomy is to retain its theoretical value, but determining what this threshold is differs by domain (2015, 55). In the democratic citizenship domain, for instance, high self-determination, with plentiful opportunities to be and do are required; while, in the medical domain, it is high self-governance, especially competence, that is key (2014, 40). For legal and personal domains, the threshold is different still.

Second, multidimensional accounts attempt to show that autonomy is complex. For transnational feminists, this takes seriously the idea that a plethora of factors, not a single explanation—often reduced to local culture for women in the Global South (Narayan 1998; Mohanty 2003; Jaggar 2005)—can affect one's life and options.

Mackenzie, for instance, argues that low opportunities, such as insufficient capabilities for health, nutrition, or employment (axis i) can increase exploitation, coercion, and vulnerability in multiple domains, including medical ones (2014, 26, 29). In turn, these factors can impinge upon normative attitudes towards the self (axis iii) when making medical decisions. A person's sense of responsibility for their decisions (36) may begin to wane, or they may come to mistrust and lose basic confidence in their judgments, in part because of their limited external conditions. Finally, one's authenticity (axis ii) over decisions can be affected (40–41). A desire may not be authentically a person's “own” if they are alienated from it, if reflection is inadequate and insensitive to how it was formed, and if it does not fit her practical identity (32) because of these internal effects.

Killmister's account similarly helps interpret the ways in which contextual phenomena—like material deprivation, double binds, and adaptive preferences—can undermine autonomy. For example, a woman could value motherhood as critical for having a good life, form a commitment to that value (2018, 24) and set goals (27) around it. One aim may be that she wants to care for her children, full-time, at home, and failing to do so would be detrimental to her (axis i). Yet poverty may mean she has to compromise the time she spends with her children by finding employment, working two badly paid jobs, or carrying out work that she finds degrading (146–47). If oppressive conditions compromise commitments in this way, this woman's self-unification (axis iii) is reduced (148). In other words, contextual phenomena of these types impair attainment on one of the dimensions of autonomy (144).

Third, whatever determination is made about the autonomy of persons, both accounts seek to engage those persons in ethical ways. This matters for transnational feminists, who emphasize that women of color must be treated in respectful ways (Narayan 1993; Mohanty 2003; Khader 2019). Paternalism and disrespect towards women in the Global South has been rife; avoiding these outcomes is paramount.

Killmister's account, for instance, allows interventions that target failures of some aspects of autonomy while being respectful to those we acknowledge as agents. For those who hold false beliefs but meet conditions of self-constitution (axis iv), for example, she proposes greater information provision to correct such beliefs. Likewise, she only advocates “nudging”—governmental action likely to improve well-being but which involves an element of deception—if agents already hold a commitment to not-being-against-being-deceived.

Mackenzie, meanwhile, argues that scaffolding be put in place to enable the autonomy of those who do not meet the threshold of self-governance or self-authorization (2014, 40). This suggests a non-paternalistic approach insofar as not demanding a specific choice (2014), as well as an account where having autonomy is not determinative of being treated with respect (2015). Rather, Mackenzie proposes supporting or enabling self-governance or self-authorization in non-disrespectful ways, whatever persons choose.

Evidently, there is much to support in multidimensional approaches, especially in their nuance about autonomy, the existence of autonomy in constrained conditions, and in treating others respectfully.

Transnational feminist critiques

Despite these welcome benefits, multidimensional accounts are also subject to some of the same criticisms that transnational feminists have levied against feminist theory in general. In this section, I draw out two such critiques and highlight how multidimensional autonomy falls short. This is not to dismiss the approach but to identify its unrealized potential to go even further in accounting for autonomy in oppressive circumstances given transnational feminist insights. The criticisms are that there is (a) greater abstraction from context and more individualism than necessary and (b) less focus on complex and specified identities than would be conducive to a fuller understanding of decisions in constraint.

Abstraction and individualism

Transnational feminists have sought to reframe theory in ways that have radically destabilized the notion of atomistic individualism; that individuals are fundamentally isolated beings (Mohanty 2003; McLaren 2017). They take selves to be deeply embedded in relations not just with other humans and institutions—which many western feminist theories acknowledge—but also with nature, land, spirituality, nonhuman others, etc. (McPherson and Rabb 2011; Weir 2017; TallBear 2019).⁴ For some, they are “co-constituted” in these relations (TallBear 2019, 36); “intrinsically connected” (Weir 2017, 277) to them. This type of rejection of atomistic individualism yields revolutionary political implications. For instance, non-individualism of this kind requires genuine, rather than token, solidarity and collectivity in feminism (Mohanty 2003; Fultner 2017), and it lends itself to structural solutions that can often be overlooked in more individualist theory.

While deciphering what these differences mean for autonomy is less obvious, I attempt to bring this out—first for relational freedom, where there has been relatively more elaboration in the literature, and then for relational autonomy, where there has been comparatively less. When discussing freedom, Alison Weir, for instance, describes Indigenous feminisms as adopting a position of “radical relationality” (2017, 276). This is a freedom “situated in relation to land (in the broadest sense), and is deeply connected to struggles for sovereignty [… and is] a quality of ethical relationships” (277, original emphasis). Indigenous women's individual freedom is “intrinsically connected” (277) to communal sovereignty. This is a deeper rejection of atomistic individualism than the Western relational self without this shared goal, casting the net of relationality more extensively. It has implications for freedom itself: “the possibility of a different conception of individual freedom in relationship” (278), and for what freedom entails, requiring connection with land and others.⁵

Radical relationality has insights for autonomy too. It resists the individuality or instrumentality that can still be associated with relational autonomy, argues Weir.⁶ For instance, radical relationality does not seek to restructure relations to enable the “maximal pursuit of individual ends” or to facilitate “the capacity to discover and to follow one's own will or law” (277), which Weir takes to be the respective views of Iris Marion Young and Jennifer Nedelsky. These are more individual or instrumental purposes and notions of self. Rather, radical relational autonomy implies restructuring relations for—what I will argue are more revisionist—non-individual purposes (communal goals of freedom of land, etc.) and as ends in themselves (for better relations because these are already part of self).

Elsewhere, Weir (2008) offers a helpful expansion of this idea through Cynthia Willett's work. She argues that freedom (which, here, she takes to be relational autonomy) is not to be understood simply “in the context of relationships one has (the project of most theorists of relational autonomy).” Rather, it is a much more positive conception—it “is precisely the capacity to be in relationships that one desires: to love whom and what you choose to love” (16, original emphasis). This is not the capacity to pick and choose relations in a vacuum or, more charitably to relational autonomy theories, one's ends despite being embedded in relations. Instead, it is the capacity to live fully, in ways people find mattering to them, in deep connection with others. It is to fully be ourselves as human beings (Horkheimer 1972); an ultimately radically relational state of affairs.⁷

I call this position, roughly, “being-with.” Being-with, I propose, is entwined in the notion of autonomy; it is a more radically non-individual concept than standard relational accounts. This, in turn, I offer, is likely to require more demanding political interventions. Being-with suggests more revolutionary measures, such as about land or other groups, than any isolated, individual ends permit.

This discussion on different types of relationality also applies to multidimensional accounts. Both Killmister and Mackenzie take selves to be relational in the non-transnational sense, but not necessarily in the transnational one. This reflects the Indigenous women described less well than a radical relational self does, and I argue that it fosters a less innovative approach for enabling being-with. To see this more clearly, consider the extent to which autonomy itself is relational on multidimensional views.

The accounts are reluctant to take autonomy to be inherently relational; the disinclination being either total or partial. Killmister's view is instrumentally relational in that other people are needed to learn about autonomy and to exercise the skills of autonomy. However, being in a particular kind of relation with others is not definitive of what autonomy is; rather autonomy is always about the capacities of the individual (2018, 144). This differs from the inherent relational conception of a radical approach to autonomy that I have outlined.

Killmister's position differs too from the (Western) relational idea that autonomy is realized with others through a dialogical disposition to answer for oneself (Westlund 2009). Mackenzie, on the other hand, does mention Andrea Westlund's dialogical approach under the self-authorization axis (2014, 36), so there is some scope for a more non-instrumentally relational autonomy. Indeed, elsewhere, Mackenzie intimates that relational autonomy—which she takes her view to be—simply is socially constitutive (2015, 48), which implies it is more similar to a radical view.

However, Mackenzie—as well as Killmister (2018, 144)—ultimately reject a stronger connection between external conditions and internal states. External and internal are merely causally interdependent not intrinsically connected (Mackenzie 2014, 24).⁸ In short, both accounts—understandably given current frameworks of many relational models—are at pains to avoid an excessively strong connection between the external and internal. Yet, however the constitutive versus instrumental debate is settled, the focus on it misses an important opportunity: to recognize radical relationality and to reimagine the connection between the external and internal. I argue that such an approach does not primarily ask us to consider challenging the external or to be as progressive as we might. I contend that we need to adopt radical relationality and reinterpret the relationship between external and internal factors for autonomy to do this.

Killmister's interventions, for instance, are based on individual-targeted solutions. Policy-based interventions, like nudging, damage autonomy if—as is almost inevitably the case (Thaler and Sunstein 2008)—they are implemented more broadly than at the level of the individual (Killmister 2018, 180).⁹ Interventions concerning peoples or land, taken to be fundamental to selves on a radical relational view (Weir 2017) and for being-with, may be less pronounced when interventions focus on individuals.

Mackenzie, meanwhile, does conclude that “structural social and political change” may be required to tackle restrictions in self-determination (2015, 63). This might include expanding opportunities to be and do. As this centers on the capabilities people can actually function in, rather than a negative idea of freedom from harms, this endorses a more positive conception for autonomy, just as radical relationality does. However, this stops short of the deeply relational approach that radical views imply. It enables people to be and do within their current embeddedness rather than a richer notion of thoroughly being-with others as they find meaningful.

In another example, as already argued, both accounts suggest that autonomy is impaired, even if not negated or determined, by negative externalities. However, the focus on how these diminish an individual's autonomy re-emphasizes and re-centers the individual; the framework that a radical relational view seeks to resist. For instance, contrary to there not being any damage to an individual's autonomy in multidimensional accounts—because such accounts separate internal and external axes (Johnston 2017)—there is a failure that is ultimately associated with the person.¹⁰ This is because their overall autonomy is reduced compared with someone without restricted self-determination. It does not suggest the structural, social, or political parts themselves are at fault simpliciter. This effectively gives prominence to, and concentrates solutions on, the individual. In so doing, it curtails interventions that could tackle structures in deeper ways for being-with.

Transnational feminists certainly want to protect and promote women's individual sense of agency (Mohanty 2003; Welch 2013; Khader 2019). However, the focus on agency, rather than autonomy, by various transnational feminists is a deliberate strategy. It is to broaden the discussion from mere individual indicators and solutions regarding autonomy to contextual and social ones (Madhok 2007). While relational and multidimensional accounts do better than traditional autonomy theories on this front—by including external factors or capabilities—an individualized focus still remains in two ways. First, multidimensional theories may lead to modest types of interventions relative to even less individualized transnational feminist accounts. Starting with a radically relational view is likely to allow more extensive interventions that enable being-with. Second, at the same time, multidimensional accounts still end up with a view of reduced autonomy of the person overall. The focus of the discussion remains the individual, despite admirable efforts to avoid over-individualizing. Not all transnational feminists will want to discuss autonomy beyond individual capacities, as I will go on to discuss. For now, though, I suggest, there are stronger, non-instrumental, ways to conceptualize autonomy that I claim better fit the transnational critique of challenging individualism.

Identities in oppression

Western feminisms, in general, have been charged with essentializing the group “women”, whereby all women are assumed to share the same oppressions. Chandra Talpade Mohanty (2003, 33–37) argues that part of the reason for this is that Western feminisms are inattentive to their power and privilege in theorizing. They employ methodologies that uncritically justify concepts, experiences, values, etc. as being universal when they are not. The methodologies and their resulting theories develop in the West and then are, wrongly, regarded to extend to all people in all contexts. For instance, gendered division of labor anywhere is thought to be oppressive, because it is so in the West (35). As transnational feminists argue, however, women are socially located in complex spaces, with reference to race, sexuality, nationality, disability, caste, class, religion, coloniality, and more. As such, there are both different kinds of oppressions faced by members of the group “women” and varied experiences of gendered practices that may not be oppressive.

Relatedly, transnational feminists challenge the attitude of superiority in the Western assumption that “third world” women are not subjects of power, whereas “first world” women are (Narayan 1997; Mohanty 2003; Khader 2019). The former are regarded as dupes of culture or patriarchy because they behave or choose differently to the latter, who supposedly evade such forces. Yet identities in any location are multiple and also likely to bring both privileges and disadvantages given their complexity. It is not the case that women are entirely either victims (as “other” women are often regarded) or autonomous agents (as Western women are comparatively presented). Rather, they will each occupy various non-binary positions.

The multidimensional accounts I have been considering do not squarely deal with differing identities in these ways, but they do appeal to the broader notions of practical identities and changing selves. That is, they require a notion of identity: having a sense of “who one is” allows us to claim desires, commitments, etc. are meaningfully “ours”, even when our identities evolve over time and are relational. Further, as already discussed, multidimensional theorists are able to identify autonomy in oppression, which recognizes the possible agency of women in the Global South. Likewise, as noted above, in offering individual-specific remedies to improve autonomy, they show there is sensitivity to “woman” being a heterogeneous group. However, differing identities and their connections to autonomy have been less well covered. In particular, the accounts presuppose that the factors that indicate autonomy for some women in one context must be the same for different women in any other context. Yet such a general account is prone to smuggle in Western biases, undermining its usefulness.

One could read Mackenzie picking out the actual opportunities to be and do that are available to individuals (axis i) as her indirectly connecting how identity-specific oppressions can affect self-governance. However, as her philosophical interest is domain-centric (law, political participation, bioethics) rather than woman-centric (woman located in particular spaces), there is no direct exploration of such identity-based specifics for different kinds of women. This, no doubt unintentionally, leaves us with a generalized notion of “woman,” which potentially misses a way of gauging autonomy that is specified to women's identities, histories, etc. As Pinar Dokumaci (2020) explains about “pious feminist” women in Turkey, such women neither fully adopt the language, knowledge, value sets, etc. of “compliant” (religious) or “resistant” (secular) feminists. Rather, they carve their own approach to being feminist women that spans across and beyond multiple feminisms. Given this, it is plausible that they hold their own conceptions of autonomy that are similarly untethered to a single existing model too. As transnational feminists have argued (Mohanty 2003, 40–41), women may navigate their identities, lived experience, narratives, theories, in unique ways that signify autonomy. Yet it remains invisible to us if we take up a generalized view.¹¹

In a different example, Killmister stresses the importance of intention formation and acting on intentions for autonomy (axis ii). This differs from Saba Mahmood (2005) who argues that pious women in Cairo, Egypt, perform acts of piety out of habit long before anything like intentionally embracing piety. One's socially located identity matters to how one engages with decision-making in the first place, and habit, according to Mahmood, may well be a form agency takes.¹² Assuming a predefined manifestation for aspects of autonomy without specificity to the local, then, can obscure indications of agency. Killmister categorically and exclusively focuses on Western liberal contexts in her theory (2018, 86), so there is no fault of hers in cross-applying her account. The worry is that the theory—with its liberal Western leanings—may inadvertently be thought of as falsely universal and get co-opted, as it is, to southern contexts. For transnational feminists, avoiding the generalized Western woman in theory from the outset can mitigate this.

In sum, the claim has not been that multidimensional accounts necessarily exclude women in the Global South from being autonomous. Rather, it is that these accounts are both more individualized and more generalized than many transnational feminists may take to be correct or useful.

Developing multidimensional accounts in transnational feminist-friendly ways

The critiques considered thus far could lead us to reject multidimensional accounts altogether.¹³ However, as the theories retain much merit, I believe this is too hasty. A different strategy—the one I adopt here—is to develop positive amendments that overcome these problems. In this section, I set out some guidelines and a purpose for theorizing about autonomy to help identify what I suggest are “transnational feminist-friendly” amendments. In the final part of the paper, I offer two proposals that satisfy these conditions and also address the concerns raised.

Guidelines

In order to limit neo-colonialism in theory, a fundamental aim for transnational feminists, I suggest that any amendments should adhere to three guidelines.

First, there needs to be a shift away from ethnocentrism—the idea that the standards of one's culture are best—and eurocentrism—the assumption that Western European culture is best. With regards to autonomy, we might, for instance, reject views that reflection by oneself is better than discussing with others, that rationality is superior to affective responses, and that the mind is more attuned to what we want than the body. The former of each formulation is associated with Western norms, and so regarded better by default (ethnocentrism).¹⁴ Likewise, we should be dubious if it turns out that, on a framing of autonomy that is partial to and values certain traits, the white, male, secular, New York, banker is always more autonomous than the brown, female, religious, Jullundur-based, surrogate. There is something askew if Westerners are ceteris paribus more autonomous by default due to a specific configuration of autonomy (eurocentrism). Mindful of these biases, a less-Western-centric conception that is transnational feminist-friendly would be unlikely to automatically take Western framings or agents as the optimal sort.

Second, given that the West has determined the orthodoxy for autonomy, we should be willing to be more open or imaginative about autonomy where helpful. One need not keep the “referent point” for a concept fixed just because it has always been that way, for instance; rather, one must be willing to “unfix” it, if doing so permits a transformative and non-dominating feminist theory (Wynter 1982). In particular, if the concept of autonomy is to be genuinely useful to women of color—such as in recognizing complex instances of agency—and if it is important to highlight histories—such as the hierarchies synonymous with colonialism—then structures cannot be ignored. That is, we might resist the idea that autonomy is merely a capacity of individuals simply because that is the way it has usually been framed. Focusing purely on the individual and whether they are autonomous as the referent point detracts from the feminist and postcolonial structural lens that underpins many transnational positions.

Third, and finally, the concept of autonomy should not be too abstracted from the real world. A lack of concreteness threatens to make autonomy over-demanding and/or over-valued in actual state of affairs. On over-demandingness, having too stringent an account would mean few people can attain it, rendering it unhelpful and insensitive to current conditions. Expecting secondary or tertiary levels of education, for instance, more likely excludes those in the Global South, who may not meet these conditions but be highly in tune with what they desire nonetheless. On over-valuing, only valorizing autonomy falsely aggrandizes it and misses negative consequences in reality (Khader 2019). There can be losses of social connection and support if autonomy is conceptualized and realized as greater independence, for instance. Alternatively, harms may accrue from being too subjected to social whims and not pursuing goals different to one's community if autonomy is overly relational (Khader 2019). A non-abstracted version, such as one that resists high thresholds and reductive framings, is more achievable and attuned to the world.

Purpose

Philosophers theorizing about autonomy usually have a purpose for their accounts in mind, even if only implicitly. For instance, some advocate autonomy theories for identifying persons who are subject to the moral law (Kant 2012), others for showing which citizens can participate in the political realm (Christman 2009), and others still for distinguishing which agents are sufficiently self-controlled to count as authors of their personal decisions (Frankfurt 1971; Watson 1987; Bratman 2007).¹⁵ Drawing on transnational (Narayan 1998) and radical Black activist (e.g., Olufemi, 2020) feminist insights about bettering the lives of women of color, I propose an ameliorative purpose for autonomy.

An ameliorative account of autonomy (AA), following Sally Haslanger's enquiry into gender and race (Haslanger 2000), would seek to improve, rather than merely describe or conceptually clarify.¹⁶ Amendments under AA are for enhancing the situations of women in the Global South in ways meaningful to them. For instance, where there are gaps between what a person takes to be meaningful and what she does, and the deficit is explainable by historic or ongoing identity-based injustices, like racism-sexism in (neo)colonialism, attending to those inhibiting forces in ways that matter to her is prompted. There are also notable constraints on amelioration for transnational feminists. Improving harmful structures, rather than determining or attaining optimal structures (Khader 2019), is important.¹⁷ Likewise, increasing opportunities, rather than banning options (Narayan 1993), is to be preferred.

Versions of AA, of course, are neither new nor immune to criticism.¹⁸ But if the improvements suggested do not pay attention to the guidelines of AA, they would not be transnational feminist-friendly, according to my account. For instance, if the proposals decree women of color to always be less autonomous than white men because typically masculine traits are valued, they would not meet this condition. Similarly, if they fail to foreground structural changes for justice improvement alongside discussion of individual capacities, they would not be acceptable. Likewise, if they exclude some women from being autonomous because unjust structures exist, they would not be valuable for transnational feminists.

Critical and intersectional relational autonomy

I now turn to the amendments themselves and offer two proposals: “critical” and “intersectional” relational autonomy. In broad terms, my proposals aim to recognize individual women's abilities, while also commenting on deeper structural improvements, without the outcome that women are deemed non- or less- autonomous by default because they live under oppressive conditions.

In addition to outlining the conceptual features, I also demonstrate how the proposals seek to improve the real lives of women in the Global South in ways meaningful to them; the declared purpose of AA. To do this, I draw on one practice that transnational feminists have examined in some detail: commercial surrogacy in India.¹⁹ This is to show that AA is not simply an abstraction but is concrete. It centers enquiries into autonomy on what matters to and serves those in constrained situations. The example of surrogacy, then, is in defense of AA, rather than to advocate a particular model of surrogacy. Further, the theory is not limited to commercial surrogacy, nor to this specific context; rather, AA applies more broadly beyond this example and location.

Critical relational autonomy

I begin with the critical relational proposal. This adopts, but also builds on, radical relationality in two ways: it affirms the deeply relational interpretation of autonomy and it adds an impetus for others to combat relations that undermine this. Through this, I show that any threats to positively being-with is harm to relational persons, not a reduction in autonomy.

First, a critical view adopts a radical position that humans are fundamentally relational. Any conception of an autonomous life is categorically not one that is removed from or despite others, but that is always in good connection with them. This requires that the inherently relational nature of autonomy be made visible and explicit for all. There are deep interdependencies that connect all agents including—through interpersonal, institutional, natural world, nonhuman other, and structural ties—you and I, whether agents know of the significance of these relations or not. That these relations enable people to be-with is what is key under critical relational autonomy.

Second, identifying these relations helps diagnose harmful relations to autonomy, but, more notably in a critical approach, it also presses to improve those relations. The critical focus makes autonomy not simply about a capacity of individuals to live as most meaningful to them, as in a radical position. Instead, it is also an active and practical aim for those to whom they relate. Those who stand in relation to others have reason to advocate for change in ways meaningful to those others to facilitate being-with. This is distinct under a critical position.

Holding these differences in mind, a critical position allows that oppression harms a person being-with, but that this does not mean that they are non-autonomous. When oppression exists, it impacts what people can be or do. This, however, triggers a practical demand to tackle it, not a referendum on the degree of the person's autonomy that exists.²⁰ In a reimagined picture, the relationship between autonomy and oppression is inverted. Oppression does not soak up autonomy like the proverbial sponge thereby reducing it, as under multidimensional accounts; rather autonomy—a notion of being-with—acts as a mirror to demand change to the oppression which threatens it.

It is more apt on my view, then, to discuss this in terms of oppression harming the person as a fundamentally relational being. It is well understood as a harm to the person, since oppression—at heart—targets the self (Lugones 2003, 58). It reduces, immobilizes, presses down on persons as members of groups (Frye 1983). This relationship does not change one's status as a person qua person on my account.²¹ Yet it does threaten it—other people and systems can certainly make one's life go better or worse—and it is disrespectful to a person's status. Using this language enables the claim that the harm is to persons rather than to their autonomy.

To be clear, on my view, poor relations do not equate to (i.e., lead to a negation of) autonomy or, differently to multidimensional accounts, impair (i.e., lead to a reduction of) autonomy. Instead, they are to be identified, and improved, as they encroach upon thoroughly relational persons. Relations that harm agents’ goals and projects are of ethical import because they harm not just fuller expression of oneself but the (intrinsically relational) self itself. This is recognizing oppressive relations not as deficits of an individual's autonomy but as harms to the person, and then working to reduce those harms. More pithily: the problem is the world, not the person. My approach decouples harmful relations and autonomy reduction, differently to multidimensional accounts, but does not wholly detach those relations. On the contrary, like the multidimensional view, it sees the significance of them. But it is in a more profound way, as it stresses the gravity of the harm of those relations to us as human beings.

A critical relational approach is also more extensive than multidimensional accounts. It does not just seek to tweak relations but transform them, including by overhauling dominating structures and systems, to live fully as deeply relational beings.²² While this has parallels to a radical relational view, a critical relational approach goes further. As human beings are inherently relational, with all of us connected to each other, the critical approach is the demand for us to contribute to challenging those relations to support other people's being-with. It recognizes our roles in this endeavor too. How might we demand more of ourselves, others, institutions, structures, to enable relational goals and life plans of persons to whom we stand in relation? Importantly, any such demands are initiated from those whose lives they are about and action is in conjunction with them. Together, this reflects an ameliorative, not merely a descriptive or conceptual, purpose.

There are two further reasons for why critical relational autonomy is a useful approach for transnational feminist concerns.

First, critical relational autonomy is not too externalist and does not revert to an internalist-only view. That is, some transnational feminists resist “socially constitutive” (Khader 2020) accounts of autonomy, or favor “extremely ‘thin’” accounts of autonomy (Narayan 1993, 431).²³ This is not because they reject relationality. Rather, it is because “thick” accounts connect autonomy with non-oppressive external conditions. When applied to women in the Global South who may lack these conditions, this risks denying them the very autonomy that may protect them against paternalism, thereby increasing their susceptibility to greater harm. Externalism encourages paternalism.

As intimated earlier, multidimensional accounts, on some views, can evade this concern by distinguishing the internal and external more sharply (Johnston 2017). Indeed, there has been concession that theories like Mackenzie's (when taken to be socially constitutive) could successfully do this (Khader 2020). This allows outsiders to offer opportunities to people under oppressed conditions but not override their wills. Yet, Serene J. Khader (2020) argues that this tactic raises other worries for the accounts. One needs to show why self-governance trumps other axes, like self-determination (where oppression is captured), such that it matters most in judgments of autonomy. Doing this, however, ultimately means that self-governance equates to autonomy; an internalist-only position.²⁴

The account I am proposing recognizes, like multidimensional accounts, that it is important to capture external features in a conception of autonomy. It also endorses that those under oppression or in harmful relationships can be autonomous nonetheless. Their agency is neither obliterated nor determined by less than perfect conditions. However, as discussed, my account goes further. It dissociates any negative relations as reducing autonomy itself; rather, the issue is that the relations themselves are problematic in some way. This avoids concluding that such women have reduced autonomy at all, and so further limits the risk of paternalism.

As requirements under a transnational-feminist friendly account for self-governance are minimal, many—including oppressed people—will achieve it and so evade having their wills overridden. If internalists take this to be the most important aspect about autonomy, say in preventing state coercion, they can seek out merely self-governing type conditions. Doing this would be enough to not be paternalistic for them. However, under AA this is not the only element or case that matters; rather identifying and improving relations for being-with do too. Indeed, in the state coercion example, the state itself stands in relation to many oppressed persons and so is a candidate for improvement.

If the goals of AA are accepted, it is more conducive to consider structures as part of the framework of autonomy because it also instigates discussion about what to do. The worry with an internalist-only account is that relations often get sidelined; it is more likely, I suggest, that we pick out instances of autonomy without identifying or seeking to address injustice, if we do not overtly consider them together in the same account.²⁵ Of course, this is not necessarily the outcome, but it is more probable when done separately than together. Since the aims of AA require something broader than merely discerning individual self-governance, a critical relational autonomy is apt. Paternalism is resisted but other aspects—like harmful relations—are challenged too. The external matters.

Moreover, while internalist-only accounts can attain these non-deterministic outcomes too, they do so at a cost of other advantages. Even though they do not seek to reject relationality, for instance, they effectively regress to a more individualistic autonomy. Khader (2020, 8–9) notes that some internalist-only (i.e., reflection alone) conceptions could acknowledge relationality by “allow[ing] that some forms of socialization are especially compatible with reflection and hold[ing] that caring deeply about others can be reflectively endorsed.” While this is true, it is a more impoverished notion of autonomy than how (for example) Dokumaci understands the pious feminist women in Turkey. The women of her study are “relationally constituted pious sel[ves],” not “socially embedded, individual, pious sel[ves]” (2020, 252).²⁶ Such women see their autonomy as perpetually interdependent with their relations, such as religious and feminist values and with the views of others. Meral, one of the interviewees, explicitly connects her autonomy with the restructuring of relations with laws and policies prohibiting headscarves in public places, for instance. These are women, I contend, who see their very autonomy as deeply entwined with their relations. Internalist accounts with a relational component that merely values certain socialization for reflection or allows for reflective endorsement of caring about others less fully embody this type of understanding of relationality than a (radical or) critical relational autonomy.²⁷

The second reason a critical relational autonomy is useful is that it applies broadly, not merely to those who endorse it. This is not to falsely universalize interpersonal and structural relations as important, which many transnational theorists would balk at. Rather it is to recognize them as truly foundational; it is an ontological claim about what people are like. While it would be wrong to suggest that everyone has radical relationality involving “connection to land” or “connection to the living enspirited world, in all of its diversity and processes of change,” which Weir identifies for some Indigenous peoples (2017, 276), it is right that everyone is radically relationally connected. An inherently relational autonomy, thus, applies to all because such connection is fundamental for all.

For instance, the critical relational account does not apply only to women, who are stereotyped as more relational and so more likely to value a relational approach than men. Nor does it apply only to women of color, who are stereotyped as such and, in addition, as more culturally rooted than white women. Rather, it also applies to those often considered most autonomous, typically by their omission, when discussing autonomy in relation: white men. This group, who are stereotyped as disvaluing interpersonal connection, is harmed by its absence too. White men in the West notoriously suffer deficits in emotional care and connection because there is gendered value placed on stoicism and not expressing one's vulnerabilities. Their relations are not conducive to being-with, so they are harmed and something needs to be done.²⁸ Relations are fundamentally important to all, and, on a critical view, poor relations harm many types of people being-with.²⁹ Ultimately, the world, as it is, with its multitude of anti-being-with relations, requires significant change for us all.

Critical relational autonomy and surrogacy

One aspect of the dominant discourse in surrogacy that a critical relational autonomy helps challenge is that Indian surrogates are to be understood in an “either/or” way. That is, either (through an individualized, ethnographic approach) autonomous or (through a universal, moral framework) exploited and non-autonomous (Banerjee 2010; Bailey 2011). Critical relational autonomy allows comment on both.

For instance, surrogates in India may be highly reflective, but this is not (contra internalist-only models) the end of the matter. If better relations are important within the account of autonomy, as per the critical approach, there are improvements to identify and make in the situation of the surrogates. Likewise, surrogates may face oppression, but this is not (contra multidimensional views) a reduction of their overall autonomy under critical relational autonomy. There are changes to propose to relations, but this does not reflect damage to or failings of the surrogates. Finally, the women may want certain relational ends but this is not (contra radical relationality) the only focus. We are inherently connected to those women in varied, complex relationships. This invites deeper interpersonal and systemic improvements under critical relationality. Let us consider these in turn.

The evidence in the literature is that surrogates reflect in ways that show they have minimum levels of competence to satisfy internalist theories. However, many surrogates also express regret and sorrow at severing ties with the child (Sama 2012; Saravanan 2013). Withholding this relationship harms these women as persons; their relations, which are bound up with their very autonomy in a critical approach, suffer. This instigates seeking change—such as to the compulsory end to the surrogate-child relationship imposed in such arrangements, for instance. People as people are harmed by not being able to live fully (in some form) within fundamental relations that just matter to them. This is so much so that relations, and how to improve them, becomes part of every discussion, including that of a theory of autonomy.³⁰ To not do this, as internalist accounts suggest, curtails an opportunity to improve the lives of women in the Global South in ways that matter to them and that better fit how they understand themselves.

That surrogates make assessments and form judgements about their situations, likewise, meets many internal requirements of multidimensional accounts. The surrogates weigh up their options and take surrogacy to be best for them in very difficult circumstances, for example (Pande 2010a). Yet suggesting there is less autonomy overall for the women due to these contexts—since they have fewer opportunities or changed goals, for instance—as under multidimensional accounts, does a disservice to the women. The women have views on what needs to change, including having greater employment options or welfare provision. Recognizing their autonomy as intact, including relative to non-oppressed persons, is politically important in making systemic change happen. Critical relational autonomy, which both seeks to practically improve relations and sees oppression as a harm to the person rather than a diminished autonomy, does this.

Finally, the surrogates have goals and plans that are highly relational: they focus on protecting their families, and decisions are localized around their immediate lives. But it would be a mistake to suggest that they and we are disconnected in our day-to-day lives. Some surrogates in the Global South, for instance, note disparities between them and intended parents in the Global North (Pande, 2010b).³¹ These inequalities emerge from particular colonial histories, from which those in the Global North benefit. There are also ongoing colonial attitudes toward the Global South as a site of cheap labor, including for surrogacy. Pressing for change in these relationships—such as reparations for resetting unjust colonial relations that could fund better terms and conditions, jobs, or safety nets that the women want—is action that those in the Global North can take. This would support the autonomy of the women to whom they stand in relation as colonial benefitters and “service users.” This differs to a radical relationality that does not connect others, narratives, frameworks, between all and beyond the immediate in the same way, and does not insist on others’ involvement in such change given these ties.

Critical relational autonomy helps to avoid the pitfalls of the siloed (individual ethnographic or universal moralized) approach that some transnational feminists have identified as problematic in current evaluations of surrogacy. It is, thus, well-placed to navigate the autonomy of the surrogates and serve the surrogates in an ameliorative way.

Intersectional relational autonomy

I argued earlier that social identity is often overlooked when theorizing autonomy. I now argue that multidimensional accounts would benefit from recognizing intersectional identity more overtly. This would enable a richer understanding of decisions within constraint and a search for varied signs of agency.

Intersectionality describes the unique oppressions that result from membership of various subordinated social classes (Crenshaw 1989; Wynter 1990). The oppression is unique because it is neither reducible to oppression experienced as a member of any single social class (Crenshaw 1989) nor merely additive or intensified versions of any single oppression (Khader 2013). Through the case of intimate partner violence (IPV), for example, Kimberlé Crenshaw (1991) describes how Black women in the US hold intersectional identities. They may want to report IPV to authorities (important for their membership of the group “women”). Simultaneously, they may not want to fuel stereotypes of Black men as violent (important for membership of the group “Black”). Thus, as Black women, they face distinct limitations regarding IPV compared with white women or Black men.

To meaningfully include intersectionality in the discussion of autonomy, I propose to: (1) recognize the varied quality of options at the intersection, which foregrounds their complexity; and (2) highlight intersectional agency, which identifies possibly obscured agency.

First, on quality, Crenshaw's example has already shown that fewer good options limit what Black women can do in IPV. However, there may be positive effects of intersectionality for autonomy too. Diana Tietjens Meyers (2000), for instance, argues that understanding one's complex identities is important for greater self-knowledge. “Knowing oneself”, including knowing what one might do when intersectional identities conflict, provides greater visibility of what is meaningful to persons. Uma Narayan (1993) suggests that those who are cognizant of their oppression hold epistemic privilege. Such a position enables a unique and realistic perspective of one's options. These views indicate that, while intersectionality often highlights harms to persons, it can also be helpful for autonomy. Agents knowing their complex selves and situations better fosters autonomy insofar as it increases the chances of “being true to oneself.”

Second, on agency, the women navigate their worlds in ways specified to their multiple identities. Khader draws on Naila Kabeer's work to show that women in Bangladesh, who do not have legal divorce, may appeal to “verbal divorce” instead. This form of divorce publically functions to admonish men to uphold their duties in marriage, thereby achieving what they want while remaining in their communities (Khader 2019).³² The women in Crenshaw's example too might want interventions that stop their partners’ harmful treatment of them in ways that do not compromise their identities. They may wish to “report” them to others that are properly sensitive to their racist-sexist contexts, and want official authorities to not be excessively interfering and violent in their interactions (1991). These women demonstrate high levels of agency—whether they (e.g.) resist, modify, accept, reform their situations—despite less good options and with particularity to their identities.

There are both conceptual and political advantages to including intersectionality in AA. One conceptual benefit is that there is greater visibility of the complex kinds of opportunities afforded to, and desires of, particular agents.³³ Another—the one I will focus on here—is that it permits conceptual space for different manifestations of autonomy, not merely those commonly noted, such as reflection (Mackenzie 2014) or deliberation (Killmister 2018). Intersectionality, I argue, primes theorists to be open to the ways those with multiple identities could navigate their options, express their desires, or understand autonomy within their contexts. For example, some may live in non-individualist societies without a norm to reflect in isolation at all, so a test for reflection per se (or for reflection to be adequate and sensitive to desire formation) is not as meaningful (Mackenzie 2014). Others may be in contexts where it is the norm to work out desires only in collaboration with others, not via pre-known intentions and attempting to enact them (Killmister 2018). Indeed, these non-reflective and collaborative approaches may not be so distinctively “other-context” but apply to us all (Doris 2018). These examples suggest that it is better to adopt a stance of epistemic humility; one should be open to varied indicators of autonomy more generally in any context. Ultimately, intersectionality allows pluralism in recognizing ways to demonstrate autonomy, and avoids prematurely foreclosing signs of agency.

There are also political benefits. One example is that the concept reveals that those at the intersection may have a better sense of how to improve their autonomy than we, as outsiders, do (Lugones and Spelman 1983). While Killmister argues that we must provide limited autonomy-enhancing supports to others, a worry is that it is still more top-down than bottom-up; “us” giving them predefined support rather than “them” telling us what they want. For instance, the women in the Crenshaw and Khader examples hold preferred ways of dealing with their predicaments that would meet their needs more than, or only by, a generally prescribed and formal intervention, such as providing more information, calling the authorities, or legalizing divorce. Intersectionality within AA opens up space for the unique perspectives and needs of the women, making their agency and desired support more discernible.

Intersectional relational autonomy and surrogacy

How does each aspect of the intersectionality proposal help clarify autonomy in surrogacy? On the first part—the quality of options—surrogates in the Global North and South, and indeed within those locations, hold differing identities. Given this, contrary to common perceptions and as others have argued, the practice of commercial surrogacy may affect them and their options differently (e.g. Khader, 2013).³⁴ However, it is the second aspect of intersectionality—intersectional agency—that is most significant for AA. Details about interlocking identities, oppressions, and opportunities offer a richer way to discern autonomy. For instance, surrogates in several studies are relatively poorer, from lower caste groups, and are less educated than many around them—yet their agency is apparent in varied ways. Intersectional account of autonomy enables us to see particular women as not just passively having constraints wash over them, but as operating in sophisticated ways within them, even when we might not interpret it as such given our preconceptions.

For example, surrogates may decide what to do collectively with others, develop strategies of engagement, and see possibilities for themselves that belie narratives of them as non-agents. Amrita Pande (2010a) describes how surrogates at Hope Maternity Clinic in Anand, Gujarat, make demands for better terms based on their shared experiences. They negotiate that contracts should stipulate that intended parents, rather than the surrogate, pay the broker's fee. This is because it is a substantial cost for surrogates to bear while comparatively little for intended parents (990). While ambivalently accepting surrogacy, within their situations, these women restructure their relationship to the clinic, intended parents, each other, and themselves. They work out what they desire collectively, moving as a group, without reverting to any specific individual, and with focus on changing the structures and frameworks around them. This is a collective autonomy not reducible to individual reflection (Mackenzie 2014) or commitments (Killmister 2018), and it is not deferential (Killmister 2018), but it is a form of autonomy characteristic of this surrogate group nonetheless.

In a different example, Sayani Mitra and Silke Schicktanz (2016) show how surrogates in Delhi and Kolkata experience loss and grief when the conception phase of IVF fails. They are not expected to feel anything about this “liminal stage” of the process and their crying is disregarded as naive and as misunderstanding the technology (6), such that they are silenced (7) by the clinics. Yet they weep and informally attend to the loss with each other and their husbands. Despite their lacking precise vocabulary and full awareness of the technicalities, and their feelings not being recognized, validated, or even permitted by the clinic, they process an embodied experience others deny them in their own way. The surrogates are not assumed to be subjects entitled to feel this sorrow; in this regard, they lack social recognition as agents by the authorities (Mackenzie 2014). They cannot express or seek clarity from the clinic or intended parents about their plight as they do not allow space for it; so, they attempt something that their situation makes impossible (Killmister 2018). Yet, they are certainly agents who find a narrative to make sense of their very real anguish amongst each other.

In highlighting intersectional autonomy, AA encourages a closer look at perhaps unseen ways in which these women engage with their environments and compels theorists to identify instances of agency they may not immediately recognize as such. It does not presuppose one approach to discern autonomy pre-theoretically or based on an idealized form. It does not shoehorn these experiences as, in the end, individual reflection or deliberation, but accepts them as different but nonetheless valid. Once seen, this agency demands that we pay attention to what the women want and how they wish to engage in those spaces, whether that involves resistance, acceptance, or anything else.

Moreover, intersectionality requires paying attention to the least well-off in societies. It presses theorists and practitioners to consider the breadth of how the marginalized understand and exert themselves. This necessitates framing autonomy not from a privileged perspective, but as it is in reality for the majority of world's people. It also facilitates asking the right question: what is it that the least well-off seek for their lives to go better and to live more fully? If we focus on this group and see them as agents, it is more likely that we do the right things in the right ways when engaging with those already oppressed.

The progressive potential of AA for multidimensionality

Multidimensional accounts of autonomy offer nuanced ways in which to plot autonomy. Nonetheless, using a range of transnational feminist concerns, I argued that these could be critiqued, amongst other things, for their generalized focus and their less relational approach. These end up (inadvertently, I believe) being less useful to women in the Global South. I proposed two solutions that would help take account of transnational feminist concerns. First, by utilizing intersectional identities, multidimensional accounts can avert a generalized use of the Western woman, further nuance our assessments of autonomy by spotting hitherto unrecognized agency, and pay more attention to the solutions women want themselves. Second, in being critically relational, multidimensional autonomy can be less individualistic still and press us to improve structures that inhibit autonomy on the women's own terms within our discussions about autonomy.

AA has benefits that are likely to appeal to transnational feminists, should they wish to retain a concept of autonomy. AA highlights discussion of how to change structures rather than focusing solely on individual capacities (it is non-orthodox). It avoids the idea that the existence of background injustices equates to someone being non- or less autonomous, which can be the case with multidimensional and other ameliorative accounts. Likewise, it does not expect high levels of education or assume autonomy is without costs (it aims to limit abstraction). Finally, it resists the notion that Western men and women are automatically more autonomous than Southern women, which is a biased view (it is less Western-centric).

I have not argued that multidimensional accounts fail to identify surrogates as autonomous. Rather, my claim has been that multidimensional accounts can be further bolstered by my two positive proposals. To answer the question at the outset of this paper—how progressive are multidimensional accounts of autonomy?—my answer is: they are progressive, and more so than what has come before, but they could be more progressive yet.