The Principle of Sufficient Autonomy and Mandatory Autonomy Education

This essay discusses two contributions of the principle of sufficient autonomy to educational justice. In Just Enough , Liam Shields criticizes instrumental accounts of autonomy. According to these accounts, autonomy is valuable insofar as it contributes to well-being. Shields argues that instrumental arguments fail to support mandatory autonomy education in all cases, while his non-instrumental principle of sufficient autonomy does support this. This essay develops a version of the instrumental argument and argues this version can do the work of supporting mandatory autonomy education. Another contribution of the principle of sufficient autonomy is the requirement of talents discovery. According to Shields, the requirement of talents discovery renders Rawls’s principle of fair equality of opportunity more plausible, since one’s chances of accessing a given economic position depend on one’s opportunities to discover one’s innate talents. This essay argues that Rawlsian fair equality of opportunity does not have the same implications as the principle of sufficient autonomy as to which types of talents should be discovered and to what extent.


INTRODUCTION
Sufficientarianism is a doctrine that affirms that what matters is whether individuals have enough of the relevant goods. In his book Just Enough: Sufficiency as a Demand for Justice, Liam Shields develops an alternative original account of sufficientarianism. According to this account, once the threshold is secured, there could be further moral requirements. However, the nature and weight of the reasons to secure and reallocate the relevant goods changes after individuals have reached the threshold (Shields, 2016: 30).
176 Danielle Zwarthoed LEAP 5 (2017) In Chapter 3, Shields states and defends the principle of sufficient autonomy: "Principle of sufficient autonomy: We have weighty, non-instrumental, non-egalitarian, satiable reasons to secure enough autonomy for everyone to enjoy the social conditions of freedom, the conditions under which we freely form and revise our conception of the good life." (Shields, 2016, 53) A person has sufficient autonomy if (1) she is well-informed, meaning that she can establish third-person assurance of the freedom (not the truth) of her beliefs; (2) she is capable of giving reasons for her views; (3) she is disposed to exchange reasons and to participate in public reasoning activities with others (Shields, 2016: 53, 84).
The fourth chapter of Just Enough is devoted to showing the contributions of the principle of sufficient autonomy to debates about education. This essay discusses two of these contributions. First, Shields argues that instrumental arguments for autonomy-supporting education fail to support mandatory autonomy education in all cases; his non-instrumental argument does support this. This essay develops a version of the instrumental argument and argues this version can do the work of supporting mandatory autonomy education as well as the principle of sufficient autonomy, and perhaps even better (Section 2). Second, Shields argues that the principle of sufficient autonomy implies a requirement of talents discovery. According to Shields, the requirement of talents discovery renders Rawls's principle of fair equality of opportunity more plausible, since one's chances of accessing a given economic position depend on one's opportunities to discover one's innate talents. In Section 3, I shall argue that Rawlsian fair equality of opportunity does not have exactly the same implications as the principle of sufficient autonomy as to which types of talents should be discovered and to what extent.

INSTRUMENTAL AND NON-INSTRUMENTAL ARGUMENTS FOR MANDATORY AUTONOMY EDUCATION
The justification of the principle of sufficient autonomy appeals to noninstrumental reasons to promote autonomy. The chapter entitled "Sufficiency and Education" argues that the principle of sufficient autonomy provides a more decisive reason to support mandatory autonomy education than instrumental arguments for autonomy. A concrete issue at stake is that parents of conservative religious communities may oppose certain forms of autonomy-promoting education. They fear that mandatory The Principle of Sufficient Autonomy and Mandatory Autonomy Education autonomy education will turn their children away from the core beliefs, values and behaviors endorsed by their communities. They could (and do) appeal to religious freedom or parental rights to justify their position. From a perspective centered on children's interests, the main worry is that mandatory autonomy education could jeopardize the long-term well-being of child-members of conservative communities. Autonomy education may estrange these children from their family and community. This would make it impossible for them to meaningfully sustain important familial and social relationships. They would also be deprived of the important contribution of cultural affiliation to one's sense of identity and capacity for wholehearted commitments.
According to Shields, the principle of sufficient autonomy succeeds in showing that securing a certain level of autonomy outweighs these countervailing considerations, while instrumental accounts of the value of autonomy do not (Shields, 2016: 90). The instrumental argument for autonomy education affirms that autonomy is good because it leads to something else, namely well-being or flourishing. It derives the value of autonomy education from the good of well-being. Worries with the instrumental argument arise from the contingent character of the connection between autonomy and well-being (Shields, 2016: 72). Such worries need not arise if the value of autonomy is not derivative. Note this does not necessarily mean concerns with the well-being of children should disappear. Valuing autonomy non-instrumentally does not preclude Shields from valuing well-being non-instrumentally. If so, the case of child-members of conservative communities will require him to balance autonomy against well-being. This balancing reasoning might lead to practical conclusions similar to those reached by "instrumentalists". To strengthen the case for the non-instrumental argument, it seems we need to know why the intrinsic value of autonomy is superior to the value of well-being.
While the case for the non-instrumental account might not be as strong as expected, the case for the instrumental one could be stronger than Shields assumes. A closer examination of the connection between autonomy and well-being shows that the instrumental argument provides little support to those who want to withdraw child-members of conservative communities from autonomy education. How detrimental we think autonomy education is to these children depends on the nature and the importance of the connection between autonomy and well-being, on one hand, and (as Shields himself puts it) on the conception of well-being we assume, on the other hand (Shields, 2016: 93). The instrumental argument affirms autonomy leads causally to wellbeing. The causal connection between autonomy and well-being can be understood in at least two ways (Schinkel, 2010: 100): (a) Autonomy may/is likely contribute to individual well-being. (b) Autonomy is necessary for individual well-being.
Shields's argument must assume version (a) of the instrumental argument since version (b) would also succeed in convincing those who are concerned with children's well-being that autonomy education should be mandatory. Could it be the case that autonomy is necessary for individual well-being? Drawing on Raz's reflections, as well as on the philosophical works they have influenced (e.g. Brighouse, 2005;Raz, 1986;Wall, 1998;White, 2006), I would like to examine two ways in which an adequate level of autonomy might be necessary to live well.
Autonomy is unlikely to be necessary to achieve some conceptions of well-being such as those based on hedonistic and actual preference satisfaction. It is possible to experience pleasure or to satisfy one's actual preferences without being autonomous. This is emphasized by Shields's discussions of happiness pills and cheap tastes inculcation (Shields, 2016: 74-76). 1 What is wrong with these examples is that people by-pass the autonomous deliberative process involved in forming and realizing their conception of the good life.
Raz's partly subjective conception of well-being (Raz, 1986: 288-312) is not vulnerable to counterexamples like the happiness pill. According to Raz, a person's well-being depends, first, on her capacity to meet basic biological needs and, second, on the successful pursuit of her current and future goals. The content of these goals does not matter so long as they are independently valued by the person herself. This means attempts to improve the life of someone else by making her achieve a good she does not and will not see as her goal will fail. Suppose Mary's mother tries to secure her daughter's future well-being by preventing her from studying history, a subject Mary is passionate about. Mary's mother believes history is a poor choice of major because she does not see the point of spending one's life neck-deep in dusty archives to write unreadable books. She pushes Mary to study communication instead, a seemingly more fun major. Unless Mary revises her judgment on the merits of a history major, her mother's attempt to make her happy will be unsuccessful.
However, success in pursuing a goal, regardless of its objective value, does not suffice to secure a person's well-being. A person's well-being also 179 LEAP 5 (2017) The Principle of Sufficient Autonomy and Mandatory Autonomy Education depends on the value of the goals she pursues. We evaluate goals, we have reasons to pursue them, and some reasons are better than others. A person has a goal, properly speaking, only if her reasons for having it are valid. Failing to achieve a goal which is actually supported by no valid reason is a "blessing in disguise" (Raz, 1986: 301). Suppose Mary is genuinely interested in history, but decides to study philosophy instead because she believes a philosophy degree is more likely to improve her job prospects. Now, suppose also that Mary's belief turns out to be false. History graduates are actually more popular with employers. If this is the case, Mary does not have, in a normatively relevant sense, the goal to become a philosophy graduate. Of course, it is a psychological fact that she has the desire to study philosophy (since she ignores her reason for having this goal is not valid). But, properly speaking, she does not have the goal to study philosophy because studying philosophy will not contribute to the success of her life. If she is not admitted to a philosophy program, this failure might contribute to her well-being unbeknownst to her.
Since there are no reasons to value worthless cheap tastes or a life determined by a happiness pill, Raz's conception of well-being avoids Shields's objections to welfarism. Now, having goals does not suffice to live well; one must also succeed in pursuing them. A person's goals provide her with action reasons, reasons that speak in favor of performing certain actions. Others cannot reach a person's goals for her: actively pursuing the goals that constitute one's life is constitutive of living well. This does not mean the good life must be athletic or hyperactive. A flourishing life can certainly consist of modest pursuits. What matters is that the person achieves these goals herself, lives her life herself and from the inside.
According to Raz, these goals need not be acquired in an autonomous way (Raz, 1986: 290-291) and one can live well without being autonomous. Yet I submit that a minimal degree of autonomy is, in fact, necessary to secure the dependency of a person's goals on reasons. It is true that a person may acquire goals she has valid reasons to value through nonautonomous processes such as habituation, early socialization, and so on. In fact, most of our goals are acquired this way. Furthermore, the successful pursuit of some possibly attractive goals, such as ballet performance, requires the child to commit to cultivate her talents from an early age, before she is fully autonomous (Arneson and Shapiro, 1996: 392).
However, being equipped with the deliberative capacities that partly constitute autonomy enables people to avoid at least two potential obstacles in the process of assessing the validity of the reasons one has to have certain goals. The first obstacle is: I cannot make sure the reasons I have to pursue goals are independently valid if I do not possess the skills and 180 Danielle Zwarthoed LEAP 5 (2017) knowledge needed to critically assess, or reassess, their validity. Without a minimal degree of autonomy, I cannot make sure the goals I pursue are based on valid reasons, and therefore I cannot make sure these goals will contribute to my well-being. As Arneson and Shapiro put it, we do not want to choose life plans we just believe are valuable, but we do want to choose those which truly are valuable. Insofar as truly valuable life plans are those which resist critical reflection, autonomy is a good instrument to verify whether our life plans are truly valuable (Arneson and Shapiro, 1996: 399).
The second obstacle is: even if I am well-equipped with the cognitive abilities needed to assess the independent values of my goals, I might not be able to want my reasons to be true reasons. I might be rationalizing my choices rather than honestly reflecting on them. I might be deceiving myself about my real reasons. Self-deception is seriously damaging to wellbeing because it breaks the connection between goals and reasons. Rationality and intelligence do not protect us from self-deception. But the capacity for autonomy does. One of the crucial dimensions of autonomy is authenticity (Shields, 2016: 59). Authenticity involves being able to reflect critically upon one's major goals and to revise them so that they cohere with one's reflectively constituted higher-order commitments and conception of oneself. By definition, authenticity requires being honest with oneself and one's reasons, even when the truth is uncomfortable. The capacity for minimal autonomy is thus necessary to have genuine wholehearted commitments to goals that constitute our well-being. When society has to decide whether to authorize parents to withdraw children from autonomy-promoting subjects or schools, controversial assumptions regarding the superiority of a secular way of life are not necessary (Arneson and Shapiro, 1996: 401). They might even be detrimental to children's wellbeing, since they could amount to unsuccessful attempts to make these children live well by pushing them into ways of life they do not endorse. But society can assume that autonomy-promoting education makes it more likely that future adults will choose the goals that are truly better for them without falling into the trap of self-deception.
The first way in which a minimal degree of autonomy is necessary to live well is by securing the dependency of our goals on valid reasons, insofar as the capacity for autonomy equips us with the skills, knowledge, and disposition to avoid errors and self-deception. The second way appeals to Raz's well-known contextual argument for the special value of autonomy in modern societies. If this argument is valid, a higher degree of autonomy might be needed to achieve well-being in these societies. Our well-being depends on our successful pursuit of goals we have valid reasons to commit to. But we create these goals out of something. Our goals are based on 181 LEAP 5 (2017) The Principle of Sufficient Autonomy and Mandatory Autonomy Education existing social forms. Social forms refer to existing shared beliefs, cultures, imaginations, practices, behaviors, and so on (Raz, 1986: 307-312). Our pursuits and activities are to a large extent socially defined. This does not mean we should align with existing social conventions. It means the meaning, significance and sometimes the very possibility of some comprehensive goals depend on existing social forms.
As Shields puts it, autonomy involves certain social conditions (Shields, 2016: 48). The social conditions, and more broadly, the social forms of modern democratic societies constitute an autonomy-supporting environment. Modern autonomy-supporting environments are characterized by fast-changing technology and economic circumstances, geographical and social mobility, value pluralism, secularization and a commitment to human rights (Wall, 1998: 166-167). In such environments, people need the capacity for autonomy in order to flourish (Raz, 1986: 391). This is not just because autonomy enhances our ability to cope with changes. This is because this environment makes it extremely difficult, requiring almost complete isolation, to lead successful lives non-autonomously.
At this point, one could argue that this argument does not apply to child-members of isolated conservative communities. Their social environment differs from the characteristic circumstances of modern societies. The range of comprehensive goals available to them is not based on autonomy-supportive social forms. Furthermore, insofar as well-being depends on the successful pursuit of socially defined goals and activities, autonomy education might render them ill-equipped to succeed in the pursuits available to them.
The objection would hold if these communities were entirely isolated from the "external world". In those specific circumstances, instrumentalists must grant that a relatively high degree of autonomy is unnecessary to live well (a minimal degree of autonomy might remain necessary to secure the dependency of goals on valid reasons). However, most of the communities which currently want to withdraw their children from mandatory autonomy education are not fully isolated. They interact with non-members at various levels. Existing political and social institutions structure interactions among members and between members and non-members. Members pay taxes and consume public goods. As the sheer existence of the Wisconsin v. Yoder case makes it clear, member of these communities rely on the same judicial system as non-members to protect their rights. Some produce goods they sell to non-members and buy consumer goods produced outside of the community. Some read newspapers and watch LEAP 5 (2017) television programs infused with the background, autonomy-supporting culture. As a result, even when they are able to protect their culture, the presence and influence of the broader autonomy-supporting context unavoidably alter the social forms that prevail in these communities. They also alter the very nature of the opportunities these communities provide to their members. The significance of pursuing the project to live in a traditional community differs greatly in a traditional society from one in which one can freely revise her goals. The very nature and value of these choices depend on whether they exist in an autonomy-supporting environment or not. In concrete terms, the option to stay in the Amish community or to become a nun does not have the same significance in an autonomy-supporting society and in a traditional society. In modern societies, this option unavoidably involves a choice, if only because background institutions provide exit options. 2 And this choice requires exercising deliberative capacities. An autonomy-supporting environment reshapes the very conditions attached to these seemingly non-autonomous pursuits. It transforms them into autonomous choices. Since childmembers of conservative communities will be confronted with these sorts of choices, their future well-being requires the capacity for autonomy too. Therefore, in our modern circumstances, this version of the instrumental argument for autonomy helps us to reach the conclusion Shields wants to reach, that is, that autonomy education should be mandatory.
Before closing the discussion, a few critical remarks on mandatory autonomy education might be helpful to refine the debate. Liam Shields does not only argue that autonomy education should be mandatory. He also suggests it should be delivered by the state: "The state cannot refuse to get involved with education and simply allow private individuals to provide for it. To do so would be to allow educational provision to be distributed in a particular way that may fail to recognise citizens' rightful claims." (Shields, 2016, 85) Additional philosophical work might be needed to make the move from the claim that autonomy education should be mandatory to the following claims: first, the state should be responsible for delivering autonomy education; second, the state should be authorized to use its coercive power to make sure all children are enrolled in state-provided autonomy education. I have no space to discuss these issues in detail, but I would like to point out a couple of questions. If Shields thinks states should deliver 183 LEAP 5 (2017) The Principle of Sufficient Autonomy and Mandatory Autonomy Education autonomy education because they are the most able agent for this purpose, something could be said about why other educational agents, and especially parents, are more likely than the state to fail to render children sufficiently autonomous (especially in less than ideal states). If Shields thinks states should provide autonomy education because they are the only agent which has the legitimate power to "force" children to get such education, the theory of legitimate authority with which his sufficientarianism needs to coordinate should be developed further.

SUFFICIENT AUTONOMY, FAIR EQUALITY OF OPPORTUNITY, AND THE REQUIREMENT OF TALENTS DISCOVERY
Let us now move to the requirement of talents discovery. Sufficient autonomy is related to talents discovery in the following way. Educating for autonomy requires agents to be well informed about the options available to them. Being informed about options involves being informed about one's interests and talents. Therefore, according to Shields, "everyone should be given opportunities sufficient to discover their talents and interests insofar as this constitutes our freedom as sufficiently autonomous agents" (Shields, 2016, 100). The array of opportunities to discover one's talents should be sufficiently broad and varied.
Shields argues that, by requiring talents discovery, the principle of sufficient autonomy renders Rawls's principle of fair equality of opportunity more plausible and should thus supplement it (Shields, 2016: 100-105). According to the Rawlsian principle, social and economic positions should be opened to all under fair equality of opportunity, meaning that those with equivalent talents and the same degree of willingness to use these talents should have equal chances of access to the same offices and positions, regardless of gender, race or social background. The principle must include undeveloped talents and not just to the subset of talents that have been actually developed. Otherwise, the principle would validate background unjust inequalities (Shields, 2016: 102). But giving productive jobs to those who couldn't have developed the appropriate skills due to unjust circumstances wouldn't benefit society in general, and the least well off in particular. Therefore, justice requires the educational system to provide prospective citizens with the opportunity to discover and develop their talents. But it would be excessively costly to attempt to discover all possible talents. Hence the requirement of talents discovery usefully supplements the Rawlsian principle by providing a criterion to define the 184 Danielle Zwarthoed LEAP 5 (2017) extent to which opportunities for talent discovery should be broad and varied.
It is true that the principle of fair equality of opportunity would be implausible if it did not require the educational system to help children to identify and develop the relevant talents. However, the requirement of talents discovery fits into the principle of equality of opportunity only if their implications regarding the kinds of talents and the extent to which they must be developed converge. This doesn't seem to be the case. The two principles are not necessarily concerned with the same talents. The principle of sufficient autonomy pertains to the talents one needs to adequately develop a conception of the good life (Shields, 2016: 99) and to participate in collective deliberations (Shields, 2016: 98). Fair equality of opportunity pertains to the talents which enable people to be economically and socially productive in a way that can be beneficial to the least fortunate (Rawls, 1999: 87). Of course, some talents, such as good verbal skills, have polyvalent functions. And, to some extent, marketable talents are instrumental to secure the capacity to adequately develop a conception of the good life. But others, such as the capacity for spiritual experiences, are less likely to be valuable in the job market. And talents that are valuable on the job market, such as combativeness, are not particularly well-suited to developing a conception of the good life or participating in collective deliberations.
Of course, the fact that the implications of sufficient autonomy and equality of opportunities are not co-extensive does not undermine the inherent plausibility of Shields's principle of talents discovery. But it puts into question his claim that the requirement of talents discovery implied by sufficient autonomy fits in well with other demands, such as the demands of fair equality of opportunities. The problem is not just that the range of talents each principle is concerned with is different. The problem is that educational resources are finite. When a society decides which talents the educational system should attempt to reveal in priority, it has to adjudicate between the demands of sufficient autonomy and the demands of equality of opportunity.
The implications of sufficient autonomy in terms of talents discovery may conflict with equality of opportunities at another level. In the sufficientarian educational system Shields envisions, the least advantaged children would only enjoy a sufficiently varied array of opportunities for talents discovery, while their more advantaged counterparts could, in addition, benefit from exposure to a much broader set of disciplines, experiences, and activities. Their chances to discover a talent that matches job market demands well are therefore higher. Or suppose the requirement The Principle of Sufficient Autonomy and Mandatory Autonomy Education of talents discovery is not only sufficientarian in terms of the variety of talents children should be able to try to develop, but also in terms of the extent to which they could develop these talents. Then, the sufficientarian version of the requirement of talents discovery runs the risk of putting some children at a disadvantage in another way. Imagine the sufficientarian version of the requirement of talents discovery requires schools to provide those who are naturally good at mathematics with the opportunity to develop the level of mathematical skills corresponding to a secondary school degree. The students who would be granted this opportunity and no more will be unable to compete to become actuaries, accountants or financial analysts (which are well-paid jobs). This is not compatible with fair equality of opportunity. The policy of talents discovery required by fair equality of opportunities must take into account the effects of competitive and comparative contexts on children's economic and social prospects. The source of the problem is that talents are goods with positional aspects. Their value in competitive contexts depends on how much of the same goods other competitors have (Brighouse and Swift, 2006). Despite its inherent plausibility, Shields' sufficientarian view of talents discovery might actually hinder fair equality of opportunity rather than reinforce it.
Shields addresses the problem of positional disadvantages in his discussion of Anderson's adequacy principle of educational justice (Shields, 2016: 110-114). He suggests that the shift-based approach of sufficientarianism he advocates is better equipped than Anderson's own version of upper limit sufficientarianism. To recall, upper limit sufficientarianism states that, once people have enough, there is no further reason to benefit them. Shift-based sufficientarianism states that, once people have enough, the nature and weight of reasons to benefit them change. Anderson's view entails that, once educational adequacy is secured, there is no further reason to redistribute educational opportunities. Shields's view entails that, once the principle of sufficient autonomy is secured, there may be further reasons to redistribute educational opportunities. Shields could thus respond to the egalitarian critic that, once sufficient opportunities for talents discovery are secured, his theory of justice can recognize there are additional valid moral reasons to limit rich children's opportunities for talents discovery in competitive contexts.
If my understanding of the implications of shift-based sufficientarianism for educational justice is correct, Shields's view of educational justice may plausibly conciliate two conflicting considerations that structure the debate about educational justice, namely, positional disadvantages and leveling down. Still, this does not show the principle of sufficient autonomy 186 Danielle Zwarthoed LEAP 5 (2017) itself adequately defines the requirements of fair equality of opportunities in terms of talents discovery. It only shows that a shift-based understanding of sufficientarianism can supplement the principle of sufficient autonomy with an egalitarian principle of fair equality of opportunities. In terms of talents discovery, this means once opportunities for the discovery of a sufficiently broad and varied array of talents have been secured, educational justice can seek to achieve an equal distribution of remaining opportunities for talents discovery. Then, a worry remains. If the demands of sufficient autonomy require a lot of educational resources, little will be left over to enable schools to equalize the economic and social opportunities of children. Recall that, according to Shields, the ingredients of sufficient autonomy are: (1) being well-informed, that is, being able to establish third-person assurance of the freedom of one's beliefs; (2) being able to give reasons for one's views; (3) being disposed to exchange reasons and to participate in public reasoning activities with others. It seems to me the educational policies needed to secure sufficient autonomy as Shields conceives it would especially focus on helping children to reach a high level of cognitive and critical thinking skills, a level most of the people shaped by our educational systems do not have reached. In the just educational system Shields envisions, important investments in the cultivation of critical thinking skills would have priority over investments in policies aiming at securing equality of opportunity such as the implementation of a school map (when and where it works) or reforms aiming at helping disadvantaged students to access to and succeed in higher education. The influence of parental background on children's future opportunities would remain decisive. Therefore, the extent to which the principle of sufficient autonomy is compatible with fair equality of opportunities seems limited.
One might think the conclusion of this discussion is that egalitarian views of educational justice are superior to sufficientarian ones. But this needs not be true, even for those who share the intuition that an educational system that fails to mitigate the effects of social background on children's social and economic prospects is problematic. The problem does not necessarily originate from sufficiency in itself. It originates from the fact that Shields' account of autonomy is not rich enough. The ideal of autonomy is not limited to critical thinking skills and deliberative capacities. It also concerns social and economic conditions as well as the kind of relationships people have with each other. A richer account of autonomy may account for the problem of positional goods by enabling us to stress how people's relative position in the distribution of certain goods may affect important dimensions of freedom and autonomy. Some capability-based and freedom-based understandings of the sufficiency threshold can address