The Convention for the Rights of Persons with Disabilities and Article 12: Prospective Feminist Lessons against the “Will and Preferences” Paradigm

Human rights have recently impacted on current conceptualisations of the rights and obligations owed to individuals with impairments, culminating in the UN Convention for the Rights of Persons with Disabilities. Particularly significant is Article 12, where interpretations have heralded a “will and preferences” paradigm which rejects substituted decision-making mechanisms, even in situations where an individual should make personally harmful or unwise decisions about their treatment, care, or relationships. This paper explores problems with “strict” and “flexible” interpretations of Article 12, focusing specifically on safeguarding issues in cases of relational abuse, exploitation, and coercion. Drawing analogies with feminist arguments opposing violence against women in the domestic sphere, I challenge the private/public and individualistic account of autonomy which is implicit in interpretations of the “will and preferences” paradigm, and suggest that proponents of Article 12 should consider the possible justifiability of expanded protectionist measures in cases of abuse involving individuals with impairments.


Introduction
The impact of human rights in the area of mental health law, particularly mental capacity law, is evident in the increasing prominence of the United Nations Convention for the Rights of Persons with Disabilities (CRPD). The CRPD's concept of legal capacity is thought to fundamentally challenge OPEN ACCESS current medico-juridical conceptualisations of the rights and duties owed to individuals with impairments. Traditionally, the concept of mental capacity has guided mental capacity law to clarify when the right of autonomy warrants respect, whereby a certain level of decision-making competence determines when deference to an individual's subjective choices is required. Others can and are obliged to make decisions on their behalf should this threshold not be met using the principle of best interests. In short, the concept of mental capacity has the dual function of protecting an individual's right of autonomy and safeguarding their welfare should they be found to lack capacity.
By contrast, legal capacity is an intrinsic characteristic that applies universally to individuals, regardless of the nature and level of impairment. Substituted decision-making regimes have questionable validity through this crucial shift towards universal legal capacity, and the human rights of persons with disabilities will demand supportive decision-making mechanisms to secure their equal rights and respect under the law. Much scholarly attention has scrutinised the legal and moral permissibility of substituted, best interests decision-making mechanisms in current legislation, such as the Mental Capacity Act 2005 in England and Wales, Ontario's Substitute Decisions Act 1992, and Americans with Disabilities Act 1990 [1][2][3] . Different analyses have also tried to articulate the specific institutional and legal requirements of supportive decision-making instruments [4,5]. In practice, Ireland has formally committed to introduce legal capacity legislation wholly compliant with the CRPD [6]. Most agree that a human rights perspective on mental health law demands greater deference to the decisions by individuals with impairments, as discussed in Article 12 (4) of the CRPD, where legal capacity requires that the "rights, will and preferences of the person" are respected-or what has been called, the "will and preferences" paradigm ( [7], p. 84). Article 12 symbolises a move away from a protectionist perspective towards disabled persons, towards a rights framework which demands policies that respect the autonomous choices of individuals. It is to recognise the "dignity of risk" so that individuals with impairments are allowed to make unwise choices without paternalistic incursions into their lives.
The normative and practical import of the CRPD is undeniable. However, the purpose of this paper is to highlight certain contradictory assumptions lurking behind interpretations of Article 12, of which are most apparent in their analysis of permissible safeguarding mechanisms in circumstances of relational abuse, manipulation, and undue influence. Exploring these contradictions in more depth raises questions as to whether current interpretations of Article 12 provide sufficiently robust safeguarding mechanisms, particularly in situations where the concerned individual "consents" to or is "happy" with the abuse by their carer, family members, or immediate support network. The paper argues that feminist challenges to the law's complicity in violence against women are instructive and may justify an expanded scope for protectionist measures. Feminists, unlike interpreters of Article 12, are willing to accept that this conclusion may follow from their critique of liberal assumptions surrounding the private/public distinction and individualist accounts of selfhood and autonomy.
The structure of the argument is as follows: Section 2 outlines the principle interpretive claims of the CRPD. Accounts of Article 12 emphasise that the "will and preferences" of the individual is binding on others which indicates the impermissibility of substituted decision-making regimes. Section 3 probes and utilises feminist legal sources to challenge the implicit commitment to the public/private dichotomy within these Article 12 interpretations. Section 4 exposes further the oddly individualistic picture of the rights-bearer and autonomy at the heart of Article 12. I suggest that a number of the assumptions grounding the CRPD's human rights approach to mental health law are incoherent, and require closer critical examination if the normative claims about legal capacity are to be found convincing, particularly in contexts where individuals with impairments are abused, manipulated, coerced, or exploited. My analysis will conclude that the foreclosure of any substituted, best interests' decision-making is not entirely justifiable if we are to draw the full implications of a thoroughgoing relational analysis of the law and concept of autonomy.

CRPD and Article 12
The concept of legal capacity within the CRPD is grounded in two aspects of equality: (i) the protection of formal equality; (ii) provisions to secure substantive equality. Formal equality refers to the fair treatment of individuals irrespective of contingent features of their personhood, such as ethnicity, gender, physical or personal attributes. Equal treatment in this respect requires abstracting from an individual's defining characteristics, focusing instead on universal, value-neutral features we all share, such as our common humanity 1 . Formal equality in this sense forms the backbone to the concept of legal capacity. Unlike the notion of mental capacity, legal capacity is inherent to all individuals, regardless of mental and cognitive abilities, and is a "universal attribute inherent in all persons by virtue of their humanity" ( [13], para. 8). Article 12 further states that "persons with disabilities enjoy legal capacity on an equal basis with others in all aspects of life" [14]. Importantly, such formal equality entails that individuals with disabilities are rights-bearers with legal agency which deserves state recognition ( [13], para. 5), thereby signalling "the change from welfare to rights" where "the equality idiom [applies to] both same and different persons with disabilities" ( [15], p. 45). This marks an important shift from mental capacity regimes where, although individuals may be entitled to certain basic rights, their legal agency-such as the right to make legally binding decisions in contracts, about marriage, treatment, and care arrangements-tends to be contingent on the individual meeting a certain threshold of decision-making competence. By contrast, emphasis on the disabled person as rights-bearer helps guarantee their legal status and agency on par with able-bodied individuals ( [2], p. 31).
But the formal equality embedded within the concept of legal capacity can only be secured through provisions of substantive equality. Whereas formal equality requires a focus on commonality and the universal, substantive equality requires a more particularistic, contextualised focus so that existing institutional supports and resources are tailored specifically to the differential needs of individuals and groups. The same treatment of certain individuals by societal and institutional structures can signal a failure to treat them equally in a substantive sense. For example, consider how public spaces have the potential to include and exclude. The use of steps rather than slopes, or the absence of elevators in public transport, illustrate how the same treatment of individuals has differential, unequal consequences: the assumption that we are all the same (i.e., able-bodied) effectively functions as barriers to the equal treatment, access, and inclusion of those with physical impairments. Thus, a crucial part of the CRPD is the protection of substantive equality so that institutions must provide "reasonable accommodations" to individuals with impairments ( [13], para. 33). 1 I set aside the issue of the "status" of individuals with severe cognitive impairments. Suffice to say that proponents of the CRPD (and the normative intent of the CRPD) seek to challenge claims that those with significant impairments lack the status of personhood. But for more of this debate, see [8][9][10][11][12].
Legal capacity challenges the contingent link between decision-making capacity and legal agency under mental capacity regimes. Supportive mechanisms must be available to individuals with impairments to enable them to exercise their agency and secure both formal and substantive equality. Decisional support implies a range of macro-and micro-duties, extending from state policies and legislative changes, to interpersonal advocacy and networks of support [16]. Most importantly, Article 12 contains the crucial clause that the "exercise of legal capacity" must be ensured to "respect the rights, will and preferences of the person" [14]. This signals the rejection of substituted decision-making mechanisms where other individuals can make best interests decisions on behalf of another with a finding of mental incapacity. According to the General Comment issued by the Committee on the Rights of Persons with Disabilities, the principle of "best interests" is non-compliant with the "will and preferences paradigm" and violates the right of individuals with impairments to enjoy legal capacity "on an equal basis with others" ( [13], para. 21). There are flexible and strict interpretations of the "will and preferences paradigm". The General Comment signals a strict interpretation that rejects any justification for best interests' decision-making on behalf of individuals with impairments. Respect for the legal capacity of disabled individuals is coextensive with deference to their rights, will and preferences with regards to their choices about health, treatment and care. In short, the subjective preferences of the individual are prior to any other welfarist considerations or third-party obligations to intervene. Michael Bach and Lana Kerzner in their report for the Ontario Law Commission offer a more flexible interpretation of Article 12, proposing three decision-making statuses. Legally independent status refers to those individuals who possess "requisite decision-making abilities" that are tracked by "understand and appreciate" tests traditionally associated with tests of mental capacity (i.e., the ability to understand, retain, use and weigh information) ( [2], pp. 83-84). An individual's need for support and assistance to exercise these abilities introduces two other decision-making statuses. Supported decision making status "distributes decisions-making abilities required for competent decision-making processes across an individual and his/her supporters" but crucially, this is "as directed by the individual's will and/or intention" ( [2], p. 24, emphasis added) . By contrast, facilitated decision making status applies in temporary circumstances where an individual's impairments prevent them from acting independently and they also lack supportive others who have personal knowledge to determine their will and preferences. This status permits some substituted decision-making based on the principle of best interests. Ultimately, however, the "best interpretation of will and preferences" remains standard for these instances where substituted decision-making is required ( [17], p. 6).
The CRPD has many welcome features, such as the normative focus on the social inclusion of and equal respect for individuals with impairments, the critical challenge to and scrutiny of substituted decision-making mechanisms in existing mental health legislation in national jurisdictions, as well as the outright resistance to outdated, harmful presumptions about the lesser status of disabled persons. Nonetheless, contradictory assumptions at the heart of the "will and preferences" paradigm are exposed when probing questions of justifiable safeguarding interventions in cases where individuals with impairments are abused by their immediate support/care relationships. Both strict and flexible interpretations acknowledge the reality of unacceptable levels of abuse, coercion, and exploitation of disabled persons, but each struggle to provide a plausible solution mainly due to problematic inconsistencies in their assumptions. On one hand, the private sphere of individuals with impairments must be protected, including their right to make choices about their relationships, even if harmful to themselves. Yet, on the other hand, public institutions, society, the state, etc. have positive obligations to ensure the promotion of individuals' legal capacity. In other words, the "will and preferences paradigm" asserts the traditional liberal separation between private and public, yet recognition of the positive, supportive interventions that are needed to exercise these legal rights and capacity demands undercutting the very same private/public dichotomy.
The next sections argue that certain lessons can be applied from interpretations of the European Convention of Human Rights as well as feminist examinations of violence against women to mitigate this contradiction. There are important parallels between the women's and disability rights movements: both feminists and the social model of disability maintain that oppression, inequality, and discrimination occur through social, structural failings, where society and its various institutions entrench patriarchy and disablism. Both focus on the social construction of identity in terms of gender and able-bodiness respectively. However, feminist discussions of violence against women avoid the contradiction that besets interpretations of Article 12, mainly because they accept the consequences of their premises. As we see below, feminists argue that the social construction of identity (and the inequality that results): (i) demands the breakdown of the public-private distinction which may sanction intervention into one's family life and private choices; (ii) can (and does) compromise the internal decision-making processes of an individual to the point that it impedes one's autonomy. Though understandably resistant to these implications, I suggest that interpretations of Article 12 ought to consider feminist responses to violence against women in order to provide convincing safeguards in cases of relational abuse involving individuals with impairments.

The Private/Public Distinction
Within the strict reading of the "will and preferences" paradigm, individuals with impairments are entitled to certain resources, goods, and supportive mechanisms necessary to exercise their capacity. This entitlement nonetheless does not translate into a positive duty to intervene in private, potentially harmful choices of the individual. To use a case study, imagine P has a combination of learning and cognitive impairments; she lives with her father who struggles to provide the necessary care and support P requires. He has a fractious, deteriorating relationship with social care workers, claiming their attempts to provide support are both unnecessary and invasive. This ongoing conflict confuses P herself-on one hand, she expresses her interest and preference for engaging in the activities offered to her by social care workers (such as attending college, going on outings), yet she is loyal to her father who claims these individuals are interfering with their lives. He forbids P from seeing these social care workers, increasingly isolates her from contact (i.e., prohibits her from answering the door or telephone by punishing her if she disobeys, repeatedly tells her that he is the only one who can take care of her, that she cannot survive without him, that she must rely on him solely). There is a worry amongst social care workers that, should this enmeshment and isolation continue, the various skills she has acquired through outside contact will be suppressed, meaning her future autonomy might be at risk. Attempts to have outside mediation have been unsuccessful through the obstruction of P's father. The very few times social care workers have managed to speak to P, she tells social workers that she strongly desires to attend the various activities she had access to previously, as she enjoyed experiencing and learning new things. But when asked about whether she would like to live elsewhere apart from her father she says she "wishes to stay with her father since she is completely helpless without him".
These types of situations frequently occur where individuals with impairments choose to remain with carers who neglect, coerce, or abuse them, even against the supportive advice by others. Indeed, carers of disabled persons can misuse power to systematically isolate individuals away from others [18][19][20]. Yet, interpretations of Article 12 assert that the will and preference of P is legally binding should all else fail (see [2], pp. 144-45), given that any substituted decision-making-and by default, third-party interventions-would violate P's legal capacity. Lurking behind this conclusion is an adherence to the liberal distinction between public and private spheres-partly due to how rights language typically asserts the protection of the individual and certain areas, such as family life, from outside intrusion. Or if such interventions are permitted, they must be strictly limited in accordance with the principle of proportionality ( [21]; [22], p. 30).
Given how pervasively individuals with impairments have been imposed upon with some undesirable paternalistic treatment, most advocates of Article 12 welcome the move to defer to a person's expressed will and preferences. The CRPD's change "from welfare to rights" is argued to protect the right of autonomy of disabled persons ( [15], p. 45). Yet, some circumspection about the liberal concept of individual rights is necessary, especially in light of how it constructs and shapes our notion of what is private and what is public. In the first instance, feminists contest the value-neutrality of individual rights discourse, noting that rights can embed ideological power which in turn systemically excludes or oppresses certain groups. Private property rights, for example, have historically protected men of certain classes to the disenfranchisement of women, indigenous groups, and the disabled [22]. Liberal rights that rest implicitly on a public/private distinction further determine contestable spheres of regulation and non-regulation. The boundary between the regulated and non-regulated initially appears as a value-neutral description of some sort to characterise areas where the law cannot intervene. The private sphere is often described as descriptive fact or a feature of biology-the individual is literally a bounded self, the family is a bounded unit, the house is a bounded structure-yet history indicates that non-regulated areas of society are political, legal constructions to legitimate contestable norms and inequitable practices. Feminists argue that these descriptions reflect complicity towards certain existing social arrangements and power relations which run through the private sphere ( [22], p. 29). As Nicola Lacey puts it, "non-regulated areas may be seen as areas in which the legal system implicitly legitimises sexism and racism, given the social facts of their existence" ( [22], p. 30).
Moreover, the private/public distinction, though pervasive in the liberal imagination, remains mythical in legal practice. Regulation of the family is a case in point. A mundane example would be truancy laws in the UK, where parents can be fined for decisions to excuse their child from school during term-time for a family holiday [23]. If the private/public, non-regulated/regulated distinctions hold, parents would presumably have the right to make decisions which affect their children, including the right to decide when their family goes on holiday. But truancy laws and the authority of schools regulate if and when children can be excused from school. For examples more pertinent to my argument, consider legal interpretation of Articles 3 and 8 in the European Convention of Human Rights (ECHR). Comparisons between the ECHR and CRPD are speculative by nature, due to the fact that the latter lacks the legal status of the former 2 . Nonetheless, closer examination of ways that relational abuse has been addressed under the ECHR is instructive for my analytical purposes. Discussing Article 3, Jonathan Herring points out that the prohibition of torture and inhumane or degrading treatment or punishment is an absolute right, making it impermissible for this right to be infringed in any manner. The conventional public/private dichotomy is subsequently challenged, given that the state's duty to intervene overturns a family's right to privacy in cases of intimate abuse where one person's Article 3 rights are infringed ( [24], pp. 282-83). Herring puts this point even more strongly, "in an intimate abuse case the state cannot justify its failure to protect a victim's Article 3 rights by referring to that person's right to respect private life. But what about cases where the victim "consents" to or is "happy" with their mistreatment? Even as these borderline cases are unlikely to qualify as instances of inhuman or degrading treatment under The family life for which Article 8 requires respect is not a proprietary right vested in either parent or child: it is as much an interest of society as of individual family members, and its principal purpose, at least where there are children, must be the safety and welfare of the child. It needs to be remembered that the tabulated right is not to family life as such but to respect for it. The purpose, in my view, is to assure within proper limits the entitlement of individuals to the benefit of what is benign and positive in family law. It is not to allow other individuals, however closely related and well-intentioned, to create or perpetuate situations which jeopardise their welfare ([25], para. 43).
Two points are notable in Munby J's interpretation of Article 8: firstly, the right to private and family life is a qualified right which must be balanced against another individual's Article 8 right. As such, it is a balance between two competing negative rights. But the second aspect goes further: the positive obligation of the state to intervene in situations where the family fails to protect the bodily and psychological integrity of the individual (in other words, when Article 8 paragraph 2 is satisfied) qualifies this right to private and family life 3 . Where competing negative rights of private individuals and positive obligation of the state require balancing, Herring argues that the law must consider the values that underline these rights.
In the case of Article 8 the underlying value is that of autonomy: the right to pursue your vision of the "good life". A judge could then consider the extent to which the proposed order would constitute a blight on each of the party's opportunities to live the good life and make the order which causes the least blight. Applying that in this context I would argue that although removing the victim from intimate abuse from an abusive carer will infringe the carer's autonomy, it will do so to a much lesser extent than leaving the victim to suffer abuse would do. But what if the victim does not want the assistance? Here there is a balance between protecting the current autonomous wish of the victim, with the increase in autonomy they may experience if they were removed from the abuse ( [24], p. 283).
According to this analysis, the application of Article 8 has to move beyond the public/private dichotomy in order to fulfill requisite obligations in certain situations. This may very well mean intervening on privately chosen, but abusive or disabling, relationships, particularly if this enables the individual to develop and exercise her autonomous agency more fully.
Thus far I have suggested that an analogy could be drawn between the promotion of an individual's legal capacity and the way that Article 8 has been interpreted. But one might object that this analogy fails, precisely because, unlike Article 8, the right of legal capacity is envisaged as a universal, unqualified right. So long as an individual can express their rights, will and preferences, with or without support and assistance, this takes precedence over other considerations. But this response is unconvincing. In the first instance, the notion that the right of autonomy is absolute in this way is implausible. Even the most extreme defenders of libertarian rights see the need for the right of autonomy to be balanced with other conflicting rights, interests, or public goods [27,28] 4 . How Article 8 is interpreted and applied illustrates well how rights function, not as "trumps" in Dworkin's sense [29], but require balancing continually, moving between fluid boundaries of what is deemed private and public. Arbitration between these conflicting goods will inevitably demand deliberation about the values these rights are trying to protect. Even as it seems uncontroversial to suggest that the individual's right of autonomy is paramount in interpretations of Article 12, how we actually realise this right is far from straightforward: it may necessitate further consideration of how relationships actively impede or disable the development and exercise of autonomy. I will have more to say on this point in the next section. But suffice to say here, if Article 12 demands the promotion as well as respect of an individual's autonomy, this may well sanction third-party interventions in disabling relational circumstances.
Feminist analyses of legal agency in cases of violence against women provide even stronger challenges to the public/private dichotomy implicit in interpretations of Article 12. According to Jennifer Nedelsky, the law faces a strange conundrum in cases where women kill their battering partners: on one hand, feminist defense lawyers emphasise expert testimony which provides evidence of abuse, violence, and its resultant psychological damage-what is commonly called "battered well-being of the country, for the prevention of disorder or crime, for the protection of health or morals, or for the protection of the rights and freedoms of others" ( [26], Art. 8, para. 2). 4 See for example the discussion of compossibility in [27].
women's syndrome" ( [30], pp. 175-83). This syndrome helps capture ways that relational abuse fundamentally inhibits and damages the skills and sense of self which are required to exercise individual autonomy. It also mitigates the assumption that the abused could leave if the abuse was so severe. On the other hand, battered women's syndrome reinforces simplistic, harmful gender stereotypes-such as the helplessness, passive victimhood, and non-agency of women-which contribute to ongoing inequality and patriarchal oppression. Thus, women who kill abusive partners out of self-defense are characterised as simultaneously reasonable and incapacitous. But whereas the law currently focuses more on what this contradiction means in terms of the culpability of the individual agent, Nedelsky shows that a relational analysis of a battered woman's agency would concentrate on how her relationships and social environment impacts on her agency. A woman who repeatedly returns to a battering partner can be a reasonable agent, yet particular relationships compromise and damage her autonomy ( [30], p. 176). Going further, Nedlesky highlights that battered women are kept in abusive situations, not just due to their private relationships. Societal failures to protect women also function as coercive forces that keep battered women in these situations, since the absence of societal options make exit unavailable even when she faces death threats ( [30], p. 181). In Nedelsky's words: It would seem a full understanding of the impact of battering on a woman's autonomy requires attention both to deep psychological impairment-feelings of worthlessness, dependence on the batterer, difficulty in seeing a way out, a profound sense of helplessness-and the sorts of coercive force that, absent societal protection, even a fairly convention of autonomy would see as a serious constraint ( [30], p. 181).
In other words, societal, legal, public institutions that adhere to the private/public dichotomy are often disinclined to protect and intervene in privately chosen relationships. It is deemed a private matter: the battered woman will make her own choices, even as this raises contradictions in how we characterise her agency (as reasonable-but-not-autonomous). Nedelsky argues that, to capture the full scope and depth of harm to women requires a fundamental shift in understanding, away from the private/public dichotomy. Battering is, not just a private affair that happens between individuals, but a social phenomenon that persists through the systemic patterns of behaviour by various levels of public officials (police, prosecutors, judges) and society (neighbours, friends, and family), all of which collectively fail to protect women from intimate abuse ( [30], p. 183). Conventional views about the private and public domains fall away as a result, as societal structures' failure to protect women are themselves implicated in domestic abuse. Failure to protect can occur in two ways: through the absence of external support, positive options, and assistance, as well as the failure to sanction external interventions within certain limits. We can and should debate about what limits are appropriate. But the important point is that, regardless of whether or not interventions are made in situations of domestic abuse, the law is structuring relationships in crucial ways between intimates, between the individual and what is typically deemed the "public sphere". Crucially, this structuring occurs, even when private decisions are deemed legally sacrosanct. Acceptance of conventional boundaries between the private and public, non-regulated and regulated, is essentially to take a value-laden stance of what power relations are legitimate and valid, and-by implication-sanctions violence against women so long as it remains cosseted away in the private sphere.
This example helps probe the implications of interpretations of Article 12 5 . In the first instance, both strict and rigid readings would likely concur with feminist analyses' emphasis on the need to reform societal structures to improve provisions of support, exit options, and resources. However, where they part ways is perhaps in Nedelsky's claim that protective mechanisms need to be more robust. This conclusion is consistent with feminist arguments that undercut the private/public distinction-the private is political essentially. By contrast, interpretations of the "will and preferences" paradigm tend to gloss over the need for protective safeguards which may sanction legitimate interventions in an individual's abusive relationships. Resistance to this conclusion is understandable: after such historical, paternalistic mistreatment by public institutions, individuals with impairments do and should have the "dignity of risk", to make mistakes and unwise decisions about their health and care treatments, living arrangements, and relationships. In no way am I denying that this is a valid proposition. But present interpretations of Article 12 have tended to minimise how avenues to protect individuals likewise need to accompany supportive structures. The Committee on the Rights of Persons with Disabilities recognises that these disabling relationships "may be exacerbated for those who rely on the support of others to make decisions". Whilst it is stressed that safeguards should be in place, particularly in cases where "the interaction between the support person and the person being supported includes signs of fear, aggression, threat, deception or manipulation", they continue that such protection nonetheless "must respect the rights, will and preferences of the person, including the right to take risks and make mistakes" ( [13], para. 22). However, this response fails to see that protection is not always contrary to respect for the autonomy of individuals. Private relationships of abuse, manipulation, coercion do inhibit an individual's potential to develop autonomy skills, and the failure to intervene can itself signal the neglect of positive duties of support to encourage the autonomy of those with impairments.
One could argue that the flexible interpretation of Article 12 mitigates this concern. Bach and Kerzner acknowledge explicitly the widespread abuse and coercion of individuals with impairments and recommend safeguarding mechanisms to protect individuals from "serious adverse effects", particularly those who are isolated, with limited financial resources. "Serious adverse effects" occur when individuals experience (a) loss of property or necessities for themselves and their dependents; (b) serious illness, injury, or is deprived of liberty and personal security; (c) threats or attempted threats to cause physical/psychological harm to oneself; violent or threatening behaviour that generates cause for others to fear physical/psychological harm from oneself ( [2], pp. 174-75). Various representatives, facilitators, and monitors will evaluate what is required to protect the individual's legal capacity in cases of serious adverse effects and provide supports to the individual in question. For example, a Legal Capacity and Support Officer will be assigned to "arrange supports as needed to address situations where serious adverse effects are occurring or may occur and there is reason to believe that a person's 5 Dempsey [31] suggests that the state's duty to intervene in domestic abuse is partly aimed at challenging patriarchy. In drawing analogies with feminist arguments, I am not suggesting in the first instance that the state's obligation to intervene in cases where individuals with impairments are abused is also part of a broader strategy to combat systemic disablism and discrimination. Though intuitively plausible, further argumentation would be required to make this stronger point and would extend beyond the scope of this paper. My thanks to an anonymous reviewer for raising this interesting point and drawing my attention to the reference. ability to make and/or act on their decisions will be enhanced by such supports" ( [2], p. 120). An Administrative Tribunal would adjudicate disputes around appropriate supports for the individual.
On one hand, the establishment of these various monitors (the Legal Capacity and Support Officer, the Administrative Tribunal) suggests that in certain situations of severe adverse effects, and depending on the decision-making status of the individual, there may be a positive duty to intervene even if the individual in question opposes it. As they state: The Legal Capacity and Support office would have the authority to investigate concerns, complaints, and allegations of serious adverse effects in situations where individuals are in a supported or facilitated decision-making status, or where there are reasonable grounds to indicate that a person is unable to act legally independently ( [2], p. 141).
In other words, public bodies have an obligation to investigate suspicions of abuse involving individuals with a particular decision making status, even as he or she might insist on staying with their abuser. It would follow that the safeguarding powers of monitors may well recommend overruling the individual's preference to remain in an abusive situation in order to best support her legal capacity. But Bach and Kerzner eventually reject this conclusion, arguing alongside the strict interpretation instead that the rights, will and preferences of the individual has priority if we are to protect disabled persons from an overly paternalistic approach to safeguarding. Positive obligations to protect individuals from abuse, if finely balanced, should defer to respect for one's autonomy so that "people have the legal capacity to say 'no' to others who would impose treatment or confinement, or a particular service upon them in the name of protection" ([2], p. 95). In contrast to a protectionist approach, respect for individuals' legal capacity will outweigh future interests in physical, emotional, psychological safety or long-term autonomy. From the standpoint of both strict and flexible interpretations of Article 12, the possibility that the safety and long-term autonomous agency of individuals may not always be secured may be a price worth paying in order to respect the expressed choices of individuals (which is coextensive with respect for their legal capacity).
But whether interpretations of the "will and preferences" paradigm address the nub of the problem remains doubtful. Where I agree with these views is that individuals with impairments do experience unacceptable levels of abuse, neglect, coercion, and manipulation [32]. Safeguarding is relatively straightforward if the individual in question is receptive or amenable to outside support. But legal cases evidence time and time again that individuals who are in these relational circumstances frequently resist outside intervention and demand to remain with their abusers who continue to disable their agency [18][19][20]33,34]. Bach and Kerzner's approach fails to properly address the issue of what is justifiable to do (or not do) in these scenarios. It is a relatively optimistic view that individuals and those with whom they are in relationship will welcome supports offered by monitors like the Legal Capacity and Support Officer, or would be willing to attend an Administrative Tribunal; indeed, in situations of abuse, coercion, and exploitation, hostility towards, and suspicion of, outsiders remains a real possibility. Would this support then be imposed on the individual? Moreover, comparisons with the psychology behind the battered woman who refuses to leave are apt here-she is, on one hand, reasonable (or capacitious), but on the other hand, external and internal consequences of her surrounding intimate relationships severely compromise her autonomous agency; they skew her perspective of what choices and options are available to her. And sometimes staying in an abusive but familiar situation will seem the more natural option as opposed to leaving to go into a wholly unfamiliar, frightening, (albeit) supportive context. If this is true, then we need to question the ways in which adherence to the private/public dichotomy reinforce those relationships which compromise the autonomous agency of individuals with impairments. The inquiry into what our obligations and duties are in terms of intervention does not end when individuals actively resist support and choose to remain within abusive, disabling situations. It may be that feminist philosophers are more open to the prospect of outside interventions to protect, given how oppression of women has frequently occurred in what has been deemed the "private sphere", where law and public policy has implicitly sanctioned practices of gender inequality. By contrast, the state's disrespect for the private person, through forced treatment and care, characterises the historical mistreatment of those with impairments. The move to shore up the "private" is understandable given this context. Nonetheless, careful recognition of the historical misuse of public power is a necessary, not sufficient, condition to mitigate the abuse and mistreatment of individuals with impairments which often goes undetected in the private sphere.

Relational Autonomy
The public/private dichotomy may persist in interpretations of Article 12 due to their commitment to an underlying picture of the autonomous rights-bearer that is oddly individualistic, where the expression of liberty revolves around the protection of one's choices from outside incursions. For example, Bach and Kerzner state: Quinn provides an apt metaphor for the positive and negative freedoms that legal capacity is meant to protect. For Quinn, legal capacity is both a "sword" to advance positive freedom and make one's way through the world in "un-coerced" relations with others; and a "shield" protecting against others who would impose decisions upon you ( [2], p. 43).
If I understand this correctly, their argument (alongside Gerald Quinn) is that legal capacity functions as a protective sphere against outside interference into one's private choices, whilst expressing one's entitlement to certain goods and resources. In this way, Bach and Kerzner argue for the compatibility of positive and negative freedoms in legal capacity.
However, this is a rather odd formulation that rests on some conceptual confusion about the precise relationship between positive and negative freedom. Interpretations of Article 12 seek to stress the illegitimacy of third-party intrusions, to emphasise how respect for an individual's choices should outweigh welfarist, paternalistic concerns. But this articulation sits uneasily with the concept of positive liberty. In its most basic form, positive liberty emphasises that freedom is an expression of how the will is structured and motivated; it requires value judgements about one's values, desires, and preferences. On this formulation, interventions by others can be justifiable. Or to put it differently, the premises of positive liberty do not immediately rule out others imposing decisions upon you. Positive and negative freedom is linked, but not in the form Bach and Kerzner suggest. Charles Taylor has convincingly shown the incoherence of negative liberty, conceptualised purely as freedom from restraint [35]. Rather, it has to embed the evaluative stance inherent to the concept of positive liberty. To speak of freedom from restraint purely in the sense of external restraints, of protection against outside intrusions and impositions, does nothing to address the internal restraints and barriers that can restrict one's freedom. Prioritising the internal domain of freedom ignores how, even there, unfreedom can occur. And if unfreedom occurs internally, then it is question-begging as to why we would say external barriers somehow violate our freedom more than those internal barriers which directly impede the functioning of our motivations or knowledge of our authentic wishes. Moving beyond this impasse of which restraints are acceptable or not then requires some comment, discussion, judgement about the type or source of motivation that one claims as genuinely free or unfree. We have already moved beyond the scope of negative liberty at this level of analysis. Contra Bach and Kerzner, the interplay between negative and positive liberty fundamentally questions, rather than asserts, the priority of the internal, subjective domain. Taylor's view suggests why the concepts of liberty and freedom are not necessarily coextensive with the concept of autonomy. Liberty articulates the conditions of action, whereas autonomy articulates the conditions of willing. I don't wish to get too bogged down by the philosophical argument. But this slippage between concepts-whilst implicit within the General Comment-occurs explicitly within Bach and Kerzner's analysis ( [2], pp. [38][39][40][41][42][43][44]. If we accept that the primary consideration in Article 12 surrounds the right of autonomy, then it seems to me both imprecise and misguided to appeal to the positive/negative liberty distinction as justificatory grounding for the "will and preferences" interpretation. For the sake of argument, however, let me take as my starting point the "will and preferences" assumption that liberty and autonomy are intrinsically connected (or possibly equivalent). Even then, their contradictory use of the positive/negative liberty distinction is symptomatic of a rather strange commitment to an insular picture of autonomous individuals. Similar to the traditional liberal view of the decision-making self, freedom and autonomy expresses the assertion of one's will so that others defer to one's subjective choice. Interpretations of Article 12 do claim that autonomy is a relational concept ( [2], pp. 40, 84), however, assumptions of relationality are used mainly to minimise differences between disabled and able-bodied individuals-to assert that all are embodied individuals, dependent on others at some point in our lives, who require the care and assistance provided by relationships and social, public goods. It is to make space for the requisite positive obligations of support needed for individuals to realise their legal capacity. At the same time, however, the locus of decision-makingthe source of our motivation, preferences, values-remains subjective and individual. The expressed "will and preferences" of the individual is presumed authentic; it is the ultimate trump to safeguard against outside incursions into one's private life.
Autonomy as a negative liberty, shield concept exercises a powerful influence in the liberal understanding of individual rights. But this reductive account fails to capture the full complexity of how the social and individual combine in the formation of our values and practical orientation. Relational models of autonomy attempt to adopt a more holistic picture of selfhood, critically scrutinising how pernicious socialisation and norms impact internal motivating sources and values, and together subsequently, impair the skills necessary for authentic self-determination 6 . Socially acquired skills-or what Meyers calls "autonomy competencies" [37]-enable individuals to inquire about their values, needs, to answer these inquiries with a sense of their authentic self, with a willingness to address 6 I discuss the importance of relational autonomy in the context of mental capacity law in my forthcoming book [36]. There is significant debate about the extent to which whether the socialisation process is constitutive or causally implicated in autonomous selves, but I set this aside. For more on this, see [37][38][39][40][41]. discord internally and externally. Internal characteristics, such as self-confidence, self-esteem, self-trust provide the groundwork for such skills, but are vulnerable to the socialisation process. In abusive situations, these internal dispositional traits are uncultivated, and narratives reinforce opposite traitssuch as the psychological disposition to distrust ourselves and others, a tendency towards crippling self-doubt, self-loathing, a sense of diminished (if any) options. Resistance to these autonomy-damaging norms demands an analysis that probes the interpenetration between the relational support, care, social expectations surrounding us, as well as our unique motivational, perspectival lens; together they either disable or enable us to choose and act authentically. In other words, autonomy skills don't just originate within ourselves, but involve the complex interplay between individual, dispositional traits and social environment ( [37], p. 53). When our relationships discourage, suppress, or damage these skills, the authenticity of our choices may be questionable. By making respect for autonomy coextensive with deference to subjective choice, interpretations of Article 12 hive off areas of the socialised self which demand scrutiny, making us ill-equipped to probe the deeper, oppressive power structures that these preferences might reflect.
A relational analysis of autonomy therefore challenges the idea that support and respect for an individual's capacity is coextensive with deference to an individual's subjective choices [42]. Relational autonomy suggests that respect for an individual's autonomy is only part of the picture; we also need to consider ways that autonomy can be promoted. Whereas respect reinforces autonomy as a shield and boundary from external intrusions, promoting autonomy denotes positive action which may permit intervention into the conventional private sphere, particularly if this secures the individual's agency in the long-term. Indeed, other parts of the CRPD could be interpreted as supporting this point: Article 24, for example, focuses on the right to education so as to promote disabled individuals' "human potential and sense of dignity and self-worth", along with their "personality, talents, and creativity, as well as their mental and physical abilities" [14]. On one reading, these positive duties of promoting education could overturn an individual's expressed rights, will and preferences to remain in an inherently disabling relationship, particularly if we view such interventions as providing the basic skills and education necessary for autonomy, such as self-esteem and self-trust. It is not immediately clear to me how, on balance, deference to the rights, will and preferences of individuals could sanction the neglect of these positive duties in Article 24.
Thus, if we were to return to the case study of P above, a relational analysis of the situation would likely offer a different conclusion to the "will and preferences" paradigm. The authenticity of her expressed wish to remain living with her father is something to probe and consider, particularly as the discourse of helplessness has clearly infiltrated her sense of herself and what is within her power. Though she can communicate her decision and reasons clearly, a relational analysis would question how her father's imposed restrictions (physical and social restrictions in terms of contact, psychological restrictions through narratives of disablement and extreme dependency) may compromise her autonomy and long-term agency. This type of analysis needs to be undertaken if support is paramount to the exercise of legal capacity. Positive duties of support may well recommend safeguarding removal from her father in such circumstances [36]. This is not to say that P is not owed deliberative respect: indeed, the participation of P is central to any sort of engagement, whether P's wishes are ultimately implemented or overturned [42]. Ultimately, my point here is that the promotion of an individual's autonomy and long-term agency requires analysis of both the external and internal reasons behind one's choices, particularly as oppressive socialisation can entrench disablement, disempowerment, and inequality.

The Charge of Disablism
One might object that my argument thus far is inherently disablist: why should individuals with impairments be subject to more robust safeguarding interventions compared to others? Further, if the promotion of autonomy justifies interventions into the abusive relationships of such individuals, surely this is a slippery slope towards increasing public control over the private lives of individuals? The worry of disablism informs one direction of this critique, whilst a liberal worry about illegitimate state encroachment informs the other. My threefold response cannot deal with these worries in sufficient depth, but I do wish to tackle them briefly.
First, the worry about disablism can be summarised as follows: We don't intervene in privately chosen, abusive relationships between able-bodied individuals, why should we treat those with impairments differently? Disabled individuals have been subject to differential treatment on many levels, and authorising safeguarding measures in this way merely perpetuates discrimination. Foregrounding this objection is the social model of disability and its distinction between the disability (as causally rooted to social, structural, and environmental features) and impairment (as neutral features of the individual). Impairment in itself should not be a determinant of paternalistic, protectionist policies, even as it may imply the need for greater social supports. Should impairment be the grounds on which protectionist interventions proceed, this amounts to unequal treatment; intervening in abusive relationships between or involving an individual with impairment would therefore be discriminatory.
But the dynamic between impairment and disability is oversimplified in this picture [36]. As Tom Shakespeare rightly argues, disability involves a complex interaction between intrinsic and extrinsic factors. Reasonable accommodations made at the structural level may not address the residual difficulties that are linked to facts of bodily impairment ( [43], pp. 54-67). Crucially, these residual difficulties can be exploited, heightening an individual's vulnerability to situations of relational abuse. Though questioned by some [44], the concept of vulnerability arguably helps capture how, at one level, we are all inherently vulnerable through the reality of our embodiment, interests, and needs. But such inherent vulnerability can be exploited more or less, in certain situations and with certain practices, depending on the extent to which we rely on others to cope with the daily realities of our embodiment ( [45], p. 47). Appreciating the context of an individual's particular impairments and the ways it may make one susceptible to harm from others is not disablist-indeed, one could argue that such contextual sensitivity is necessary to fulfill the positive obligations owed to the individual, to secure and promote their own potential for autonomous decision-making and action in light of their unique, bodily reality. Indeed, closer examination of interpretive, dialogical practices that are grounded in equality and mutuality suggest one must recognise both intrinsic and extrinsic features of disablement to take precautions against the dangers of narrative, dialogical assimilation 7 . Blindness to difference and the 7 In [36], I discuss in depth how closer scrutiny of the dialogical, interpretive practices in the determination of capacity is required. My discussion there is relevant to how an individual's will and preferences is to be determined in marginal cases. Certain interpretive, dialogical mechanisms are necessary to ensure that it is indeed the view of the individual, not the interpreter. This paper has focused primarily on abuse, exploitation, and coercion, but it is also important to note residual difficulties caused by an individual's impairments can exacerbate what Catriona Mackenzie calls pathogenic vulnerability (generated through abusive, prejudicial, or discriminatory actions) ( [45], p. 40). This isn't to deny the fact that some interventions will be more appropriate and enabling than others [36]. But we cannot foreclose the possibility that positive action, not simply the constraint or elimination of certain practices, will sometimes be crucial for the amelioration of such pathogenic vulnerability. These positive actions must incorporate a contextual understanding of the individual, the residual difficulties of her impairment, of her relationships, in order to promote her autonomy [36].
The second reply to the charge of disablism overlaps with liberal worries about state encroachment into the private sphere. At some point, one must be willing to bite the bullet that the private sphere has to become more permeable in order to mitigate oppressive, abusive relational practices. I would argue, alongside other feminist philosophers and jurists, that this logically follows from questioning the private/public dichotomy. A relational account of autonomy works in concert with a relational analysis of the law. A relational perspective on law extends the observations of relational autonomy about socialisation's impact on developing selfhood and agency, and draws attention to ways that the law structures relationships, favours certain values over others, and has a differential impact on individuals and groups. The idea of relational abuse as a private matter is contestable from this standpoint. As Nedelsky states, "many 'interventions' are in fact the removal of unjustified impunity for actions to which the state would otherwise react" ( [30], p. 72). She suggests that, in the case of intimate abuse against women and children, a women's responsibility needs to be contextualised within "the consequences of both her personal relationship and how that relationship is situated in the wider structure of social and governmental relationships that effectively tolerate the abuse of children as well as women" ( [30], p. 302). Questioning the regulatory boundary between private and private could sanction moves similar to post-apartheid South Africa, where constitutional reform was undertaken, transforming the remit of judicial review in order to tackle discrimination in the "private" sphere ( [30], p. 214). Commenting on the fluid private/public nature of sexual practices, Lacey similarly argues: Feminist arguments are perfectly consistent with the idea that sexual practices are among those from which people have the right to exclude others and the state, but point out that the range of seriously autonomy-reducing sexual practices which call for political critique and, sometimes, action go beyond those, such as rape, traditionally acknowledged to be harmful ( [22], p. 96).
In other words, feminists accept the consequences of their critical challenge to conventional boundaries between state and society, public and private, regulated and non-regulated. Broader intervention into traditionally protected spheres is accepted as necessary to directly challenge violence against women. These boundaries are fluid and, as Lacey's example demonstrates, what is thought of as private can easily bleed into the public sphere, depending on how practices impact on the autonomy of an individual. By contrast, counterintuitive consequences result if we apply the logic of the disablist dangers with narrative, dialogical assimilation-where the other is viewed as a mere reflection of oneself. This could happen in a number of ways and need not be pernicious in all circumstances. But unless there is an awareness of how another's communicative impairments are not always value-neutral, one can potentially remain unconscious of the temptation to assert one's own voice in the process of "speaking on behalf of another". This problem is even worse when relationships exploit an individual's impairments for personal gain and power. charge. Safeguarding interventions in abusive situations between individuals with impairments and their carers or family members-particularly if the individual concerned "consents" to this relationship-is deemed discriminatory, evidencing unequal respect for disabled individuals. By the same token, then, interventions in practices of rape or intimate abuse against women who choose to remain with threatening partners would likewise be discriminatory and unequal. This logic entrenches conventional boundaries between public/private. One's resistance to these implications in the case of domestic battering exposes how traditional liberal boundaries tend to reinforce inequality and abuses of power, rather than protect and enable the agency of individuals who are already (or have been) systemically discriminated against.
Finally, I do not want to go into too much depth addressing liberal worries about the encroachment of the state, as others have discussed this more comprehensively ( [22], pp. 82-97; [24], pp. 81, 88-152; [30], pp. 231-76). As mentioned above, the law functions as a mechanism through which relationships are structured; it is not value-neutral but embeds certain values and goods (and interpretations of those values and goods), and it promotes certain privileges and power dynamics. Interpretations of Article 12 tend to use the language of rights as entitlements. This is, of course, part of the intuitive appeal of rights language; it captures the normative force of rights that many believe and accept. Crucially, however, numerous contestable assumptions inform liberal individual rights (which I have discussed above), such as ways that rights as "trumps" or "shields" reinforce the private/public dichotomy. The fact that these are treated as descriptive facts ignores the multiple ways that the legal framework itself dictates how those rights will be protected and implemented. The CRPD forwards a range of rights which suggest a combination of positive and negative duties on the part of the state, but it should be noted how these do not necessarily cohere in an easy way, that to implement the positive support required to promote the right of autonomy does not readily map on to conventional liberal categories and boundaries. To fully capture the range of duties, obligations, and rights of individuals with impairments will therefore require greater critical scrutiny of these liberal assumptions rather than strict adherence to them and-in some cases-acceptance that the consequences of fulfilling positive obligations of support will mean carefully considered interventions into what is traditionally deemed the realm of the private.

Conclusions
My argument above is likely to be controversial, but none of it discounts the profound import of the human rights paradigm in the context of mental health law. The CRPD has been instrumental in articulating crucial rights and duties owed to individuals with impairments. And quite understandably, we should be wary of overly intrusive state institutions, particularly given the history of publicly sanctioned, coercive mistreatment of disabled individuals. But I have tried to shed critical light on interpretations of the "will and preferences" paradigm, particularly in the context of intimate abuse, coercion, and exploitation of those with impairments. Advocates of Article 12 clearly seek to break down conventional liberal dichotomies between private/public through the emphasis on supportive mechanisms that are required for individuals to exercise their legal capacity. Yet, adherence to the "will and preferences" paradigm ends up limiting the normative force of these positive obligations, as seen when applied to issues of abuse and safeguarding. A number of contradictions embedded within these interpretations need to be clarified if its claims are to be fully convincing. Drawing analogies with feminist analyses of violence against women may be a productive way forward in providing a more nuanced interpretation of Article 12 beyond the "will and preferences" paradigm, so that safeguards are more robust and permissible scope is given for the legitimate evaluation and intervention into privately chosen but disabling relationships. A more thoroughgoing relational analysis of the law, selfhood, and autonomy tries to recalibrate the balance between positive obligations of support with the "dignity of risk" that individuals are entitled to. Deference to their personal choices about their relationships-even as they might be detrimental to their health or welfare-may be required. But this is not the whole story of legal capacity: as the CRPD correctly stresses, relational, social, public institutions must be provided so as to genuinely enable individuals, where there is a sense of substantive inclusion. But, importantly, this does not mean that all third-party interventions are strictly unjustified [36].