Political liberalism and autonomy education: Are citizenship-based arguments enough?

Abstract

Several philosophers of education argue that schooling should facilitate students’ development of autonomy. Such arguments fall into two main categories: Student-centered arguments support autonomy education to help enable students to lead good lives; Public-goods-centered arguments support autonomy education to develop students into good citizens. Critics challenge the legitimacy of autonomy education—of the state imposing a schooling curriculum aimed at making children autonomous. In this paper, I offer a unified solution to the challenges of legitimacy that both arguments for autonomy education face. I first defend a particular construal of liberal legitimacy, and then consider each legitimacy challenge in light of that construal. I argue that the legitimacy challenges confronting both types of argument can be overcome. Further, I explain why we should pursue both arguments, rather than resting the entire case for autonomy education on one or the other. I conclude that each argument—if it can justify autonomy education at all—can justify autonomy education consistent with the requirements of liberal democratic legitimacy.

If I am confident that I know what kind of life would be best for my child, I might sensibly think that I should raise her to live it—instill in her the relevant values, orient and encourage her to identify with them. This approach is often associated with a religious outlook on raising and educating children, and a passage from Proverbs illustrates it nicely: “Train up a child in the way he should go and when he is old he will not depart from it.”¹ Dominant mainstream secular culture can also have the effect of “training up” a child. In the normal course of their upbringing, children will be exposed to powerful messages from peer groups, popular culture, and media sources that affirm mainstream—for example, consumerist—values. If I omit to provide a counterpoint to these values, my child may internalize them in much the same way that some religious upbringings aim to inculcate spiritual values.

We need not find fault with the particular way of life into which one is trained up in order to find fault with certain forms of the training. Many liberals think that there is something valuable in people living lives that they themselves authentically endorse, not just because they have been steeped in that life but because they genuinely deem it a good way for them to live. Motivated by such considerations, several philosophers of education argue that schooling should facilitate or promote students’ development of autonomy. What I will call “student-centered” arguments for autonomy education undertake to show that education should help enable students to lead good lives, which goal, on the whole and on average, will best be served by students autonomously choosing a way of life for themselves. We might also think the goods of autonomy extend beyond the autonomous individual herself—that the health and stability of a liberal democratic state requires autonomous citizens. In that spirit, “public-centered” arguments undertake to defend autonomy education as a constituent of citizenship education: Schooling should aim to develop students into good citizens, which development requires some education for autonomy.

Autonomy education is not without its critics. Some argue that it unfairly burdens traditional and religious ways of life, reduces social diversity, or restricts parents’ prerogatives to raise their children as they see fit.² Such objections generate a challenge to the legitimacy of autonomy education: To the state’s authority coercively to impose such an education. Autonomy is a disputed concept among philosophers, but the kind of robust autonomy that I will be concerned with involves critical, reflective, and independent evaluation of one’s deepest ends. In involves “living one’s life according to one’s own best lights because one judges this a good way to live” (Gutmann 1995, p. 563). Many citizens in a pluralist democratic society—in particular, those concerned to “train up” their child to share their values—will reasonably reject the value of robust autonomy. Autonomy education, then, seems to involve promoting the values associated with some citizens’ way of life at a cost to the values associated with other citizens’ way of life.

If—as many think—liberal democratic legitimacy requires that we show mutual respect by intervening only in ways that can be mutually justifiable, then the legitimacy challenge to autonomy education seems genuine and pressing: Public expenditures to promote and coercively impose the values associated with a particular way of life seem to violate the spirit of mutual respect that regulates civic relations in pluralist societies. How can we justify a compulsory public educational regime intended to facilitate students’ development of autonomy to citizens who reject the value of robust autonomy? If autonomy education cannot be justified except by invoking values that many citizens reasonably reject, then imposing it on students and using public revenue to finance it seem to constitute a failure of mutual civic respect.

This way of understanding mutual civic respect is associated with political liberalism, exemplified in Rawls’s later work. In re-orienting his liberal theory of justice as a political liberalism, Rawls aimed to make it consistent with relations of mutual respect in a pluralistic society. Political liberalism specifies that citizens should eschew comprehensive liberal values in justifying the design of political institutions—like schools—and in justifying the coercive exercises of political power those institutions enact—like educational regimes. As a political liberal, Rawls found autonomy education to be inconsistent with mutual respect:

Various religious sects oppose the culture of the modern world and wish to lead their common life apart from its unwanted influences. A problem now arises about their children’s education and the requirements the state can impose. The liberalisms of Kant and Mill may lead to requirements designed to foster the values of autonomy and individuality as ideals to govern much if not all of life. But political liberalism has a different aim and requires far less. It will ask that children’s education include such things as knowledge of their constitutional and civic rights so that, for example, they know that liberty of conscience exists in their society and that apostasy is not a legal crime, all this to insure that their continued membership when they come of age is not based simply on ignorance of their basic rights or fear of punishment for offenses that do not exist (1993, 199).

My project in this paper is to clarify political liberalism’s most pressing legitimacy challenges to the student-centered and public-centered arguments for robust autonomy education, and to defend those arguments against the challenges in question. As we will see, I am hardly the first to argue that Rawls was wrong about what political liberalism licenses with respect to autonomy education. More generally, there is a well-developed literature on political liberalism and autonomy education. I should be very clear, then, about what I take my argument to add, and why I think it constitutes a worthwhile intervention.

Disagreement among political liberals has so far focused on the implications of the public-centered argument, with general acceptance that no student-centered argument can pass muster by the lights of political liberalism. This makes sense. We have already seen that political liberalism is predicated on an ideal of mutual respect by way of mutual justifiability. This apparently leaves the public-centered argument to bear the weight of our case for autonomy education because, while participating in a society of mutual respect by way of mutual justifiability seems to require a certain set of autonomy-related skills and dispositions which are on that basis publicly-valuable, it’s much less clear how a controversial notion of autonomy as good for the autonomous individual could ever comply with requirement of mutual justifiability. In short, political liberals can accept autonomy as a political ideal, but cannot endorse it as a substantive constituent of the good life for any particular person.

Here is where I diverge from previous politically liberal defenders of robust autonomy education: I argue that we cannot rest content with the autonomy education that the public-centered argument provides; we can and should develop a politically liberal rendering of the student-centered argument as well, based on a political ideal of the good of the student herself.

Though spelling it out fully must await further unpacking of some of the theoretical machinery of liberalism, it might be helpful now to foreshadow how my thesis fits into one common construal of the debate I’m entering. This construal centers the debate around a question of convergence: To what extent does a politically liberal citizenship education, which focuses on the development of students into good deliberative citizens, license a regime of autonomy education that converges with the kind of autonomy education that a comprehensive liberalism licenses on the basis of students’ own interest in becoming autonomous? My argument supports a convergence of political and comprehensive liberalism on the question of autonomy education, but with two twists: First, political liberals can endorse a robust education for autonomy only if they embrace a student-centered case in addition to the public-centered, citizenship-based case they typically rely upon. They should want both. That is, to build a satisfactory politically liberal argument for autonomy education, we must embrace not only the considerations at the heart of traditional democratic theory regarding citizenship education; we must also look to the surprising judgments that political liberalism can render about students’ own interest in becoming autonomous. Second, the reason the public-centered case alone isn’t sufficient for convergence is not about (or isn’t only about) the type of autonomy education it justifies; rather, we’ll see, it’s about whether it can justify providing it to all students.³

My defense of autonomy education should not be regarded as a defense of secular liberal values against religious values. As I have said, secular parents and secular mainstream culture may effectively “train up” children to embrace the values implicit in that culture just as religious parents may aim to “train up” children to share their faith. Questions of whether, why, and how we should educate for autonomy are just as much questions about enabling children to defect from the former as from the latter.⁴ And, as much as either of these, they are about enabling them reflectively and authentically to recognize the value of not defecting. While some discussion of autonomy education reads as if the goal is to enable unencumbered choice among a menu of options for how to live,⁵ this is neither a goal that we should aim for nor a goal that a liberal education will serve well. What we’re after is not unencumbered choice based purely on rational deliberation over personal preferences, but a capacity for independent reflection and judgment: Not the capacity “to detach ourselves from all our ends,” but to ask “about the value of any particular end with which we currently identify and…give a thoughtful answer to what we ask” (Callan 1997, p. 54). We want the “cognitive and emotional tools with which to sort out what it means to live well, given who [we] are” (Burtt 2003, p. 192). These tools might include, among other things, the capacities to investigate claims one encounters, to withstand peer pressure, and to defer gratification. Understood this way, autonomy requires a degree of “cultural coherence,” or a kind of embeddedness that “aids individuals’ senses of identity and hence agency” (Levinson 1999, p. 31). And, understood this way, one might autonomously endorse reasons of faith or community or tradition as reasons to embrace received values.

Just as the concept of autonomy is disputed, so too is there disagreement over what an autonomy education would look like or require. I think we must first consider what justifications for autonomy education survive scrutiny before turning to substantive questions about what it should look like in practice. After all, the kind of autonomy education that the state may legitimately pursue depends upon the case that can be made for it. But any autonomy education rightly so called will plausibly require that students from different backgrounds be educated together. We best learn to critically reflect on our own ways of life if we encounter and learn with people whose values and lives are different. This is, in part, why the debate over autonomy education is so important: It is crucially intertwined with the liberal ideal of the common school in which children from different ways of life are educated together.⁶

I do not try to show that either student- or public-centered arguments are sound. Nor do I argue that to be admissible they must comply with the requirements of political liberalism. Rather, I defend the arguments against what I take to be the most daunting challenges they face, assuming that the demanding legitimacy requirements of political liberalism apply. Because the legitimacy of an exercise of coercive political power depends on the justification that can be given for it, student- and public- centered arguments face distinct legitimacy challenges. Still, the challenges can fruitfully be explored in tandem. Defending public-centered arguments against the challenge leveled against them requires unpacking some theoretical machinery that will prove useful when we turn to a politically liberal rendering of the student-centered argument.

In section one, I sketch the two types of argument and discuss the legitimacy challenges each faces. In sections two and four, I defend each argument against its legitimacy challenges. In between, in section three, I explain why we should not be satisfied with the public-centered argument alone. Political liberals should—and can—try to formulate a politically liberal version of the student-centered argument as well. I conclude that, assuming the arguments are otherwise sound, we have two independent cases for imposing a schooling regime that educates students for autonomy. Robust autonomy education is consistent with mutual respect, and can be justified on both public-centered and student-centered grounds: Each can be offered as a justification that should be persuasive to citizens who are otherwise in deep disagreement about matters of justice and the appropriate role of the state in the education of children.

1 Two arguments for autonomy education

Public-centered arguments make no claims about what makes a life go well for the individual living it, but focus instead on the needs of the liberal democratic state.⁷ As such, these arguments seem to be the best hope for justifying autonomy education consistent with the neutrality constraint on liberal legitimacy. The neutrality constraint holds that “political decisions must be, so far as possible, independent of any particular conception of the good life or what gives value to life” (Dworkin 1978, p. 127). For political liberals, this constraint codifies the fundamental commitment of mutual respect by way of mutual justifiability: The commitment to arrange social institutions in a way that is acceptable to all reasonable citizens within a diverse society. If we justify a coercive exercise of political power using an argument whose premises are neutral among reasonable value systems, then adherents to those value systems can all accept that exercise, and it is consistent with mutual respect within an ideologically diverse society.⁸

Public-centered arguments seem like the most promising route to a neutral justification for financing and imposing autonomy education because, in theory at least, we can all endorse the value of citizenship education: We must equip children to be good citizens if we are to have any hope of living in a just and stable society. Because we all have an interest in living in such a society, no matter what else we value, citizenship education is valuable “independent of any particular conception of the good life or what gives value to life.” Citizenship educational therefore appears consistent with the neutrality constraint; and if autonomy education is an essential component of citizenship education, then it looks to be legitimate by extension.

Disagreement abounds both about what kind of citizenship education is compliant with the neutrality constraint and about the implications that regime of citizenship education will have for autonomy education in particular.⁹ Amy Gutmann is perhaps the best-known defender of robust civic education that includes education for autonomy. Gutmann argues that the virtues of citizenship include mutual respect among citizens, or “a reciprocal positive regard” across deep social divisions, which requires that children be educated to examine and evaluate diverse conceptions of the good: To take seriously ideas different than their own, rationally considering and evaluating the value judgments implicit in those ideas (1995, p. 561).¹⁰ According to Gutmann, mutual respect requires the intellectual skills with which to evaluate and take seriously others’ ways of life, but these same skills happen to enable critical reflection on our own values, too: “Most (if not all) of the same skills and virtues that are necessary and sufficient for educating children for citizenship in a liberal democracy are…[also] necessary and sufficient for educating children to deliberate about their own ways of life, more generally (and less politically) speaking” (1995, p. 573). So while Gutmann’s public-centered argument justifies only facilitating the skills and dispositions necessary for students to become good deliberative democratic citizens, this preparation unavoidably leads to students applying those skills and dispositions to their personal lives as well. In short, robust autonomy is an unavoidable by-product of teaching the mutual respect component of good citizenship.¹¹

Assume Gutmann is right that autonomy education is part of the civic education necessary for maintaining a liberal democratic state. If public education is the best available means of providing civic education, then we have some positive case for autonomy education in schools. The liberal commitment to neutrality, remember, requires that political decisions be justifiable without recourse to any particular conception of the good life. The health and stability of the liberal democratic state is a goal that citizens can agree to pursue, whatever else they think gives value to life, because the liberal democratic state protects and promotes their very capacity to pursue whatever (reasonable) good life they aim for. Insofar as the health and stability of the liberal democratic state requires that students be educated for citizenship, and insofar as education for autonomy is an unavoidable component of citizenship education, we have a justification for autonomy education that apparently complies with the neutrality constraint.

But a different sort of legitimacy challenge lurks. In a liberal democracy, where social institutions garner legitimacy from the consent of the governed, civic education appears to be “stacking the deck” (Brighouse 1998): In order to prepare students for citizenship in a liberal democracy, civic education aims to produce citizens who are “committed at least partly through inculcation of habit, to living up to the routine demands of democratic life” (Gutmann 1987, p. 52; italics added). It aims “to foster the wide acceptance of common norms, principles and procedures that provide a certain coherence and viability to communal life” (Costa 2004, p. 6). Accordingly, civic education privileges certain liberal democratic values: Public reasonableness, respect, and anti-repression. These values are endorsed to students pre-reflectively and thereafter subjected to less rigorous critical scrutiny than other values. By molding students to endorse liberal democratic values, civic education effectively conditions prospective citizens to consent to liberal institutions. But because consent is conditioned, its capacity to confer legitimacy appears undermined.¹² Since the public-good justification defends autonomy education as a component of civic education, this conditioning concern threatens to undermine the public-centered case for autonomy education right along with Gutmann’s argument for civic education.

A second type of argument for autonomy education is based on the interests of students themselves. Harry Brighouse argues that, as a requirement of justice, all children should enjoy roughly equal prospects to “live a life which is good for them” (1998, p. 729). To make good on this requirement, we must educate children to have the capacity to choose and act autonomously: Each child should have access to a range of good lives, and each should develop the skills necessary to choose from among the available options a life that they can authentically endorse and live well “from the inside” (Brighouse 1998, p. 730).”¹³ If schooling is the best means available for developing this capacity, then we have a case for autonomy education, independent of autonomy’s role in citizenship. An autonomy education justified on the basis of students’ own interests would equip students to reflect carefully on their beliefs and values, to reject those they find alienating, and to re-affirm those that survive scrutiny. To be capable of autonomy in this sense, most students will need knowledge of a range of alternatives to their received values, and exposure to sincere and enthusiastic advocacy of some among that range (1998, p. 733).¹⁴

Student-centered arguments might appear to rely on a perfectionistic judgment of autonomy as an intrinsically valuable constituent of the good life—on something like the Socratic judgment that the unexamined life is not worth living. If so, such arguments are in clear violation of the neutrality constraint. But Brighouse’s argument relies only on the instrumental value of autonomy: Autonomy is valuable not because an autonomous life is intrinsically better, but because autonomy equips students to evaluate the various life courses available and choose well for themselves. Similarly, any student-centered argument that aims to abide by the neutrality constraint will regard autonomy not as an intrinsic contributor to a good life for everyone, but as a valuable tool that will enable many students to live a life that is good for them.

The student-centered argument is not out of the woods, however. For one thing, an autonomy education justified on the basis of the instrumental value of autonomy may be effectively indistinguishable from one justified on the (highly controversial) intrinsic value of autonomy. Brighouse emphasizes that his instrumental argument justifies not autonomy promotion but only autonomy facilitation: Not teaching children to be autonomous; merely equipping them to make autonomous judgments should they choose to do so. Still, while autonomy facilitation does seem less partisan than autonomy promotion, the projects may be indistinguishable in effect.¹⁵ Enabling autonomy raises the likelihood that children will grow to value autonomous reflection and autonomous choice, which for some threatens alienation from the traditional values that their parents aspire for them to share. Such a project imposes an asymmetric burden on faith-based ways of life and threatens to diminish social diversity.

Moreover, educational resources are scarce. Expending them to teach the skills associated with autonomy at least strongly insinuates that those skills are valuable, and prioritizes the development of autonomy above other educational projects we might pursue with public resources. What neutral reasons could justify that prioritization? William Galston argues that “any liberal argument that invokes autonomy as a general rule of public action in effect takes sides in the ongoing struggle between reason and faith, reflection and tradition” (1995, p. 526). Such an argument appears reliant on prizing autonomous choice over other values that autonomy jeopardizes; indeed, many theorists take for granted that no student-based justification for autonomy education can be made compliant with the neutrality constraint on liberal legitimacy.¹⁶

Two distinct justifications for autonomy education are on the table. The first begins with the needs of the liberal democratic state; the second begins with the interests of students themselves. Each faces a challenge from the perspective of liberal legitimacy: The public-centered case relies on the justifiability of citizenship education, which appears objectionably to manipulate the mechanism by which the consent of the governed confers legitimacy on liberal democratic institutions. The student-centered case appears to rely on non-neutral judgments about the value of autonomy, and the educational regime it justifies threatens to erode social diversity.

I think these challenges can be met, and the rest of this paper undertakes to show how. If my defense is successful, then we will have two promising routes to justifying autonomy education. The defenses are worth exploring in tandem, because each depends on clarifying the conception of neutrality that rightly constrains coercive political action, and because each, we’ll see, is necessary if we are to have a complete justification for the robust autonomy education regime that proponents of both sorts of arguments apparently envision.

2 The public-centered argument defended

The legitimacy worry for the public-centered argument is that it apparently illicitly “stacks the deck”: It conditions consent for liberal democratic governance, thus undermining the capacity of that consent to legitimize such governance. The key to dispelling this worry is to notice that not all liberal democratic governance depends for its legitimacy on the actual consent of the governed, and that a public-centered autonomy education can limit its conditioning only to those aspects of liberal democratic governance that are legitimate irrespective of that consent.

Understanding why this is so requires a brief detour into the liberal democratic theory, and in particular, into the requirements it imposes on legitimate exercises of political power. Questions of legitimacy concern the permissibility of coercive political interventions to bring about a more just arrangement of institutions, or to preserve a just arrangement once established, given deep disagreement about what a just arrangement is.¹⁷ Such questions are related in important ways to the concept of citizenship. In one sense, citizens are actual members of society, on whose assent legitimacy depends. This is the concept of citizenship at play when we regard a suitably-constrained majoritarian process as legitimacy-conferring.

A second concept of citizenship generates the reasons by which actual citizens should be moved, and determines the constraints within which actual citizens’ wishes carry authority. Notice, just intuitively for now, that some coercive interventions seem not to rely for their legitimacy on the actual consent of the governed. Interventions necessary to protect certain basic rights, for example, need not be approved by an actual majority in order to be legitimate. The second conception of citizenship explains why this is so, and why we might properly insulate political protections for those rights from majority rule. This second conception is the idealized conception of political personhood that lies at the heart of liberal theories. In a politically liberal framework, this notion of citizenship (henceforth, “C-citizenship” for “capital-C-citizenship”) is not meant to describe actual citizens. It is, rather, a tool for systematizing the constraints on legitimate exercises of political power that such a framework is committed to imposing.

Consider a passage from Eamonn Callan:

The consensual core of liberal democracy resides in just those matters of institutional design where we can all agree that free and equal citizenship unambiguously requires certain social rules, such as legal protection for rights and free speech. But free and equal citizenship is also about the kind of people we become, and … allow our children to become (1997, 2).

Callan’s second sentence refers to actual individuals who are citizens, whose actual consent expressed through majoritarian processes confers legitimacy on social institutions and their practices. But his first sentence, I take it, refers to C-citizenship—an idealized concept of political personhood that gives substance to our theorizing about legitimacy. Liberalism’s concept of C-citizenship is built upon liberal values: In a Rawlsian framework, C-citizenship embodies two moral powers—the capacity for a conception of the good and the capacity for a sense of justice—and a higher-order interest in the protection of those powers.

C-citizenship gives substance to political liberalism’s basic criterion of legitimacy: That “our exercise of political power is fully proper and hence justifiable only when it is exercised in accordance with a constitution the essentials of which all citizens may reasonably be expected to endorse in light of principles and ideals acceptable to them as reasonable and rational” (Rawls 1993, p. 137).¹⁸ This criterion spells out the neutrality constraint we’ve already encountered. In a free society, citizens will reasonably disagree about matters of personal value—about what kinds of lives are good lives and what kinds of values are worth espousing. They will also reasonably disagree about matters of political value—about what kinds of societies are just, and about the conditions under which the state may restrict our personal judgments for the good of public values. Given this reasonable disagreement, a state that acts to promote and preserve justice inevitably will be promoting a disputed conception of justice, and inevitably will make it costlier for some to live out their values than for others to do so. For example, a policy mandating autonomy education will plausibly make it more difficult for citizens to preserve the intergenerational sharing of values on which traditional and communal ways of life depend. But if such a policy abides by the neutrality constraint, it is consistent with mutual respect. It respects all those in our political community as part of a justificatory community: Though we may disagree about many things, we show mutual respect by exercising the coercive power of the state only when that exercise can be justified on the basis of reasons that we can all recognize as such, because they derive from interests that we share.

If reasonable pluralism is inevitable, what interests could those be? While there may be very few interests that we all actually endorse, certain interests can be derived from the very project of finding fair terms of cooperation for a pluralistic democracy. Insofar as we all stand to benefit from that project, we share the interests that the project implies: What furthers that project is in our interest, whether or not we recognize it as such. The role of C-citizenship is to systematize these interests we share: Because we assume that the project of seeking fair terms of cooperation is not futile, we construe C-citizens as capable of modulating their behavior to comply with principles of justice. Because we assume that we are seeking terms of cooperation for a diverse society, but one wherein individuals can be held accountable for the values they espouse, we construe C-citizens as capable of forming and rationally revising their conceptions of the good life. Because we aim to specify conditions under which a just society can stably persist over time, we attribute to C-citizens an interest in protecting these moral powers of C-citizenship.

From this characterization we can infer still further interests of C-citizenship. Most straightforwardly, there is a C-citizenship interest in protecting a prerogative for each individual to pursue her conception of the good life, free from state intrusion. This interest generates a sort of presumption against coercive political intervention. But the presumption is readily overridden, as we can see by examining other interests of C-citizenship. Recall Callan’s words: “free and equal [C-]citizenship unambiguously requires certain social rules” (1997, p. 2). These include protections for the basic rights that protect fundamental interests of C-citizenship. Freedom of conscience, for example, protects the strong C-citizenship interest, derived from the capacity for a conception of the good, to freely live out one’s deepest values.

In this way, the concept of C-citizenship gives content to the ideal of liberal neutrality. A neutral liberal democracy seeks to act only in ways to which all citizens could assent on the basis of shared interests. Crucially, interventions can be compliant with the constraint of neutrality despite being strongly opposed, because not all actual citizens will actually recognize all C-citizenship interests as reasons-giving.¹⁹ They share those interests nonetheless, because the interests are implied by a conception of C-citizenship that is derived from the parameters of a project that we are all taken to endorse: Finding fair terms on which to justly regulate cooperation and legitimately regulate state coercion in a society marked by profound disagreement.²⁰

We are now well positioned to see why, while citizens’ actual consent expressed through democratic processes can confer legitimacy, some interventions can be legitimate without it. Social policy aimed at protecting basic liberties does not depend for its legitimacy on actual consent expressed through majoritarian processes, because such protections are fully and decisively justifiable on the basis of the C-citizenship interest in securing the capacity for a conception of the good. Plausibly, civic education will condition consent for such protections by encouraging prospective citizens to learn about and to value their basic liberties and the basic liberties of others. That conditioning might undermine the consent’s ability to confer legitimacy on those interventions, but it would not thereby undermine the legitimacy of the interventions themselves; the interventions are legitimate with or without the consent.

If civic education conditions consent only for interventions like protections for the basic liberties—interventions that are fully justified on the basis of shared C-citizenship interests—then the “stacking the deck” challenge is defused. Can civic education proceed by conditioning consent only for institutions and interventions that are independently justified on the basis of C-citizenship interests? I think so. Consider Gutmann’s proposed civic education. By educating students “to accept ways of life that are consistent with sharing the rights and responsibilities in a democratic society,” we condition them to accept coercive political interventions to protect some range of basic political liberties. But such protections would be legitimate in a liberal democracy whether actual citizens consent to them or not. Perhaps other aspects of Gutmann’s proposal involve conditioning consent for protections that do rely for their legitimacy on citizens’ actual consent, but whatever the verdict on Gutmann’s civic education in particular, we can well imagine a civic education program that privileges only values implied by C-citizenship. Such a program might undermine the capacity of subsequent consent to confer legitimacy on those protections, but those protections are legitimate with or without that consent.

Civic education aims to prepare students to deliberate together about matters of public concern on terms of mutual respect, to think autonomously and openly about ways of life different than their own, and to recognize the value of such practices for the liberal democratic state. It might also dispose students to value the liberal democratic state that protects the rights and capacities of individuals to pursue their own conceptions of the good, and that enables collective political solutions to shared problems. It thereby conditions subsequent consent for the associated, basic protections of liberalism, and for liberal democratic institutions generally. But this generates no problem of legitimacy, so long as the liberal policies and protections for which consent is conditioned are justified independently of that consent, on the basis of the very same considerations—C-citizenship interests—that justify civic education in the first place.

The public-centered argument is therefore not undermined by the worry about conditioning consent. To the degree that autonomy education is a constituent of a civic education necessary for the health and stability of the liberal democratic state, it is justified on the basis of C-citizenship interests, which generate justification compliant with the neutrality constraint.

3 Is the public-centered argument enough?

The public-centered argument justifies autonomy education only to the extent that educating students for autonomy is an unavoidable component of a civic education that prepares them to relate to one another on terms of mutual respect, and only insofar as relationships of mutual respect are necessary for the health and stability of a liberal democracy. The extent to which these conditions meaningfully constrain the substance of autonomy education has been a matter of some discussion, unfolding as a debate about the convergence thesis: The thesis that political liberalism can license an autonomy education that is (nearly) as robust as that mandated by comprehensive liberalism.²¹ Because of the strong association of the public-centered argument with political liberalism, the debate over the convergence thesis has centered on questions about how robust an autonomy education the public-centered argument can license.

For example, while the public interest in mutual respect on which the public-centered argument rests requires that students learn to scrutinize and to take seriously others’ values, and while this requires that they be able to consider the possibility that their own values may not suit everyone, it may not require that students carefully consider whether their own values are values that they themselves can authentically endorse. Plausibly, scrutinizing one’s own values requires a different (though overlapping) set of skills, and practicing those skills requires different educational experiences than those associated with taking others’ values seriously.²² I want to set aside this possible limitation of the public-centered argument, in part because it is explored so thoroughly elsewhere, but for a deeper reason as well: I think that the public-centered justification is limited in a way that should worry proponents of autonomy education—and lead us to look for a neutral rendering of the student-centered argument—whether or not the content of the education the arguments mandate diverges.

Let’s begin a few steps back. Gutmann argues that civic education requires education for mutual respect, which means preparing students to be part of the justificatory community whose members offer shareable reasons as justification for the policies they endorse. Gutmann thinks that this ideal demands that students be educated for mutual positive regard—for rationally engaging with one another’s views and taking them seriously. But it is not obvious why this should be so. The ideal of mutual respect in political liberalism is about respecting the political equality of others by limiting ourselves to political reasons for intrusion into one another’s lives—reasons that derive from shared C-citizenship interests—and by recognizing that citizens whose values are very different than ours can likewise live up to this ideal of mutual respect. But C-citizenship interests can be explored and deliberated over whether or not we know the particular ways in which others’ comprehensive values diverge from our own.

Consider the defense just mounted against the “stacking the deck” objection: I explored how C-citizenship interests justify political protections for the basic liberties without at any point canvassing any or all of the particular comprehensive doctrines that might be espoused in a pluralist society. It was enough to understand the most fundamental basic interests we share in a liberal democratic society. We may recognize interests of our own that are inadmissible in public deliberation by noticing one-by-one which interests of ours are unshared by others whose values diverge from ours. But shared reasons are not just those that happen to be endorsed by all those in our diverse community; they are reasons that we share in principle because they derive from our shared status as C-citizens. And so we can also learn to unearth and deliberate over those reasons directly, by understanding that status and the interests that it generates.

If this is what citizens must be educated to do in order to sustain mutual respect and in order to recognize one another as equals, then mutual positive regard—the kind of rational engagement with others’ values that brings autonomy education into the picture—begins to look less crucial than Gutmann takes it to be. Children can learn mutual respect not only by engaging with conceptions of the good that diverge from their own and working backward by recognizing what interests of theirs do not constitute political reasons; they can learn it more directly by understanding what makes something an interests that we all share, whatever else divides us.

We should not move too quickly here. Citizens need more than just the cognitive skills for discerning and deliberating over shared reasons; they need the dispositions associated with doing so in the spirit of mutual respect. Plausibly, then, civic education includes an affective component. Still, we might think that this requires something less than mutual positive regard. We might think it’s enough that children learn that others have values different from their own, that they can engage in deliberation over shared interests nonetheless, and that the health of our liberal democracy depends on their doing so. Plausibly, they might develop the relevant dispositions just by seeing how important public deliberation is.

Gutmann and her allies have plenty they can say in reply. As a matter of political sociology, perhaps, some degree of mutual positive regard is necessary to sustain a commitment to mutual justifiability; enough of us must be disposed to care about preserving the justificatory community in order to sustain it. Perhaps it is best preserved if students practice actually engaging with others’ values. Plausibly, then, maintaining the justificatory community requires that some considerable portion of citizens exhibit the mutual positive regard that Gutmann envisions—that they take seriously and rationally engage with one another’s values and ways of life. So far, so good. But the limitation in the public-centered argument that I think has not yet been recognized concerns how broadly this defense on Gutmann’s behalf can extend. Perhaps a society of mutual respect requires some degree of mutual positive regard like that which I’ve just imagined sustaining the justificatory community. But it is highly implausible, I think, that a society of mutual respect requires that everyone exhibit mutual positive regard.

My suspicion is that the arguments Gutmann and her allies might offer to bridge the distance between preparing students for participation in the justificatory community and preparing them for mutual positive regard and reciprocal engagement will leave some students behind. Plausibly, a liberal democracy needs everyone to be able to engage in the exchange and evaluation of shared public reasons, to recognize the basic equality of those very different than themselves, and to recognize that those very different can also engage in justification by way of shared reasons. This, I have argued, falls short of requiring engagement with others’ comprehensive values and ways of life, since shared reasons can be discerned directly from C-citizenship interests, and deliberated over without comparing diverse ways of life. Plausibly, a liberal democracy needs some number of citizens to actually engage with diverse ways of life. But does mutual respect require that everyone develop the stronger set of capacities and inclinations?²³

Plausibly, it is enough to sustain the justificatory community that all students be educated to understand the requirement of mutual justifiability and that enough students be educated additionally for mutual positive regard. But only the latter aspect of civic education carries with it education for autonomy. If this is right, then the public-centered argument justifies only ensuring that a critical mass of students receive such an education for autonomy, not that all students do. And it justifies preserving the common school ideal of a diverse array of students being educated together only to the extent necessary to maintain that critical mass.

That this difference makes a difference can be seen by examining two distinct ends to which we might put an argument for autonomy education. First, we need an argument to justify autonomy education as a public expenditure. Shared social projects require justification because their pursuit requires social investment. Social resources are scarce, and using them in one way rather than another requires some argument that there is value to be gained by that investment. Neutrality constrains the kinds of considerations we can invoke to demonstrate that value; in short, the value must be political. Second, we need an argument to justify autonomy education as a compulsory social project. What is to be done when particular individuals want to exempt themselves and their children as subjects of the coercion the social project involves? We need some justification for mandating that children attend school and that they be subject to the educational projects of schooling during that time. We want an argument for autonomy education, then, both to justify a public expenditure and to justify an exercise of coercive power to compel students to consume the product of that expenditure. These justifications will be related, but we cannot assume that they are equivalent. Public subsidies for art or for educational television programming might be justified as public expenditures, but we cannot from that conclude that we would be justified in compelling any particular person to consume them.

In one sense, the public-centered justification for autonomy education accomplishes both ends. The public good that justifies the expenditure requires that autonomy education actually be consumed, even if students—or their parents—prefer to opt out. But need it be consumed universally? Plausibly, the justificatory community of the liberal democratic state can be preserved so long as we educate all students to recognize and be conversant in justification by way of shared political reasons, and only some to rationally engage with others’ private unshared reasons. Because rationally engaging with others’ values and ideas is the vehicle for autonomy education, the public-centered argument seems to provide only a “herd immunity” case for autonomy education: The public good is realizable so long as enough students are educated for autonomy. Perhaps universal provision is the most efficient or effective way to secure that good. But efficiency seems to weigh weakly against the values that autonomy education apparently threatens, especially if we take seriously the reasons many parents have to resist it. It seems weaker still when the threshold of autonomy education necessary to secure the good can be preserved by way of compliance from parents who don’t object to autonomy education. Efficiency is a reason to make autonomy education compulsory, but it is surely defeasible.

Plausibly, the public-centered argument can justify making autonomy education and the common schooling ideal a strong public funding priority. This is important in an era of budget cutting. But we also live in an era of increased distrust of compulsory educational projects. The public-centered argument seems not to provide a justification for making autonomy education universally compulsory. The student-centered argument, in contrast, could generate strong reasons to disallow opt-outs, because it grounds the case for autonomy education in each students’ own interests. It remains to be seen whether this argument can be made compliant with liberal legitimacy, but the foregoing discussion shows why it is worth finding out. While the public-centered argument generates strong reasons of the public good for a robust education for autonomy, the student-centered argument would, if successful, generate reasons that apply to each particular student, even when their being opted out is consistent with preserving the threshold of autonomy education necessary for the functioning of the liberal democratic state.

4 The student-centered argument defended

Student-centered arguments for autonomy education begin with the interest of students themselves in developing the capacities to exercise autonomy. But by the lights of political liberalism, these arguments appear illegitimate on their face. The value of robust autonomy seems embedded in a conviction that lives go better insofar as their guiding values are reflectively endorsed by the agents in question. The arguments appear non-compliant with the neutrality constraint, then, because there are no neutral grounds on which to establish that autonomy is good—even instrumentally good—for students. Presently, I will offer an argument for the value of robust autonomy that rests entirely C-citizenship interests. Before I do, I want to address a legitimacy concern more easily dispensed with: That autonomy education discourages social diversity.²⁴ If education for autonomy restricts the ability of traditional and fundamentalist ways of life to retain adherents over time, it constitutes a hardship for those ways of life. Can such an education comply with neutrality even if political reasons favor it?

No political regime is neutral in its consequences, and liberals have long recognized that they cannot impose a standard of neutrality that forbids all political interventions that impede or promote certain ways of life.²⁵ Inevitably, political institutions will influence the values of those living under them. What matters is that we be able to justify political interventions using reasons that are neutral among conceptions of the good. This notion of neutrality—justificatory neutrality—prohibits the state from acting on the basis of reasons that assume the truth of some particular conception of the good. The state cannot, for example, invoke the good of avoiding eternal damnation as a justification for encouraging atheists to embrace the tenets of Protestantism; nor can it invoke the alleged value of monogamy to justify policies promoting monogamous families. But the state may, consistent with neutrality, interfere in ways that have non-neutral consequences.²⁶ Even interventions to protect political liberties have non-neutral consequences, after all, making it likelier that citizens will exercise or value those liberties.

Diversity might better equip us to educate students for mutual respect, and it may help secure for each citizen an array of options for meaningful lives. But while diversity can help promote these C-citizenship interests, C-citizenship interests do not favor diversity as such, and liberal neutrality does not aim maximally to accommodate it. Rather, liberal neutrality imposes principled restrictions on which limitations of social diversity we tolerate. Will our intervention diminish diversity to such an extent that citizens lack an adequate range of options for good lives? Will it diminish diversity so much as to lessen our capacity to educate students for mutual respect? Only at the extreme in which these questions are answered affirmatively do C-citizenship interests favor diversity. And even when C-citizenship interests do favor protections for diversity, countervailing C-citizenship interests might justify diversity-limiting civic education, all-things-considered. A state-mandated autonomy education may present hardships for citizens who celebrate certain fundamentalist or communal ways of life, but insofar as it can be justified on the basis of C-citizenship interests, it can be legitimate despite unequally burdening different conceptions of the good and despite limiting social diversity.

In short, liberal neutrality construes legitimacy not as a futile attempt to eliminate social costs, but as a commitment to ensuring that those costs be mutually justifiable. In arguing against robust autonomy education, Galston concedes that “the scope of permissible diversity is constrained by the imperatives of citizenship” (1995, fn. 29, p. 528). I agree. But I have been arguing that the constraining imperatives are the imperatives of C-citizenship, and that the imperatives of C-citizenship can call for political interventions that unequally burden different groups, and thus fail maximally to promote social diversity.

But can autonomy education for students’ own sake be justified on the basis of C-citizenship interests? Initially, the prospects for such a neutral rendering of the student-centered argument look bleak. The concept of C-citizenship seems not to generate claims about what kinds of life are valuable, apart from the thin starting assumption that liberal democracy itself is good. We could develop a student-centered argument parasitic on the public-centered argument: Because autonomy is good for the liberal democratic state, and because liberal democracy is good for individuals, it is in the interests of citizens to be educated for autonomy. But such an argument will inherit the limitations of the public-centered argument we just explored. Can we do better? Can we develop a neutral conception of wellbeing such that autonomy really is good for students? Neutrality prohibits the state from promoting any controversial notion of the human good for its own sake. What view of the human good could be other than controversial?

The C-citizenship interest in protecting each actual citizen’s capacity for a conception of the good might look to be of help. The capacity for a conception of the good is “the capacity to have, to revise, and rationally to pursue” one’s conception of what is valuable in life (Rawls 2001, p. 19). Some degree of self-directedness is surely required for this. But protecting the capacity for a conception of the good does not require that prospective citizens be educated for the robust kind of autonomy that involves scrutinizing one’s deepest ends, asking oneself whether they are worthy ends, and living according to one’s own best lights; indeed, as we have seen, some protected conceptions of value find such critical scrutiny positively inimical to leading a good life. If the capacity for a conception of the good does not require robust autonomy, what C-citizenship interests could fuel a student-centered case for autonomy education?

Any theory of distributive justice must incorporate some metric by which to make interpersonal comparisons of wellbeing. Politically liberal metrics will not purport to constitute actual wellbeing, because C-citizenship interests do not settle what wellbeing is. Still, some metric must be used to assess citizens’ relative prospects for attaining good lives, along a dimension of good lives that can be recognized as such compliant with the neutrality constraint. As we have seen, robust autonomy will not be among the constituents of wellbeing on such a metric. I will argue, however, that the basic capacity for robust autonomy is a good whose value is derivable from C-citizenship interests, because it is an indispensable precondition for citizens to derive the value claimed for any other goods that comprise a neutral conception of wellbeing.

We can illustrate using Rawlsian social primary goods as an example. Rawls asks “how, given the fact of reasonable pluralism, a public understanding is possible concerning what is to be counted as advantageous in matters of political justice” (1993, p. 187). His answer is that advantage consists in possessing certain all-purpose goods that enable individuals to pursue whatever conception of the good they happen to have. These “social primary goods” include basic liberties, freedom of movement and choice of occupation, income and wealth broadly construed, and the social bases of self-respect (1993, pp. 308–309). Crucially for our purposes, some of these enumerated goods imply still further goods, which are necessary to secure the value of the enumerated goods: With respect to the political liberties, what matters is not just the formal protection of the liberties, but their fair value, including a threshold of income and wealth sufficient to make full use of them (2001, p. 148). Rawls’s argument for including the fair value of the political liberties in his metric of justice invokes the positionality of political influence²⁷: “The limited space of the public political forum…allows the usefulness of the political liberties to be far more subject to citizens’ social position and economic means than the usefulness of other basic liberties. Therefore, we add the requirement of fair value for the political liberties” (2001, p. 150). All basic liberties are meant to have value for their holders; it’s in virtue of the fact that this is all purpose value that they are included in the political metric of justice. But securing the value of political liberties requires something more: Because of their positionality, securing the full value of the political liberties requires securing their fair value—a distributive requirement that arises even before distributive principles arrive on the Rawlsian scene.

Because the value of the political liberties for the individual in question can be secured only if that individual’s share of those liberties compares favorably enough with others’ shares, the fair value of political liberties can itself be recognized as advantageous consistent with neutrality: There is a C-citizenship interest in securing the political liberties, and those liberties have the value that justifies regarding them as neutrally advantageous only insofar as they come along with adequate material resources to influence political decisions. The argument I want to develop for autonomy as a component of a neutral metric of justice parallels Rawls’s argument for the fair value of the political liberties. The material preconditions for exercising the liberties—their fair value—is necessary to securing the value of those liberties that justifies their inclusion in the metric of justice; similarly, I want to argue, a basic capacity for robust autonomy is necessary to secure the full value of any goods included in a neutral metric of justice.

The social primary goods constitute a neutral metric of welfare because those goods enable each of us to pursue whatever conception of the good we happen to have. But for these all-purpose goods actually to be valuable in the way they are claimed to be, the conception of the good each person uses them to pursue must be a good one for that person. Fit matters: There are perfectly good lives that some can live well while others cannot. The state has no business deciding which lives have objective value or who can live them well, but it can recognize the simple fact that all-purpose goods are valuable for each person only if the fit condition is met—only if the life that person uses the goods to pursue is a life that that can be valuable for her. The capacity for autonomy is neutrally advantageous as a sort of safeguard: To ensure that the fit condition can be met, whatever an individual’s fortune in being born into a life she can live well.

None of this is to say that autonomy itself is a good for any particular individual. Many will flourish living out the conception of the good they inherit from their families or from mainstream culture. And, for all the politically liberal state need say, many may flourish pursuing ways of life that discourage substantive enactments of autonomy. For these, robust autonomy is not a good. Just as some will choose not to exercise political liberties even if the fair value of those liberties is secure, some will not need to exercise autonomy in order to flourish. Others will inherit values that they cannot live will. For them, autonomous reflection on and revision of their inherited values will be necessary to living well. Without the capacity to recognize a conception of the good that we cannot live well and rationally revise it, the value of social primary goods is contingent on our fortune in being born into values that are a good fit. But social primary goods should be valuable for their possessors irrespective of that fortune. For those born into a life they cannot live well, robust autonomy is a pre-condition for the social primary goods to have the value claimed for them. We need not and cannot make judgments about who these people are, any more than we judge which citizens will need to exercise their basic liberties in order to flourish. As with universal protections for the basic liberties that only some will choose to exercise, C-citizenship interests favor ensuring that all can develop the basic capacity for autonomy as a necessary condition for some to secure the value of the social primary goods.

Rawlsian social primary goods are meant only to illustrate a more general point. Whatever goods comprise a neutral metric of advantage, the value of those goods will depend on their instrumentality for pursuing a wide variety of good lives, and on each individual’s capacity to direct them toward a life that is good for her. The basic capacity for robust autonomy, then, is a precondition for the goods to have the kind of all-purpose value that justifies their inclusion in the metric. If schooling is the best way to educate students for that capacity, then we have a student-centered justification for autonomy education. This is a neutral justification, because it is based on the same political values that justify invoking a set of all-purpose goods as a metric of advantage and disadvantage within a political conception of justice. While political liberals generally accept that their theories cannot accommodate a public educational regime that fosters students’ development of the capacities for autonomy for the students’ own sake, I argue that they have missed a crucial role for autonomy in political liberalism: As a precondition for political, all-purpose goods to have the value that political liberalism claims for them.

Two distinct kinds of justification are available for autonomy education. The first starts with the public interest in the health and stability of the liberal democratic state, and derives the justification for civic education—including autonomy education—from that interest. The second starts with the interests of students themselves, and derives the justification for autonomy education from the instrumental value of autonomy to securing those interests. The task of this paper has been to defend both types of argument against the legitimacy challenges they face.

Important work remains to be done. In light of their divergent ranges of application, we should revisit the question of whether the arguments diverge in the substance of the autonomy education they support. Even if the supported regimes of autonomy education overlap, their priorities may differ. The public-centered argument justifies autonomy education that aims to inculcate the skills necessary for critical scrutiny of public political institutions, and spillover into scrutiny of personal values will be incidental. The student-centered argument justifies an autonomy education that aims to enable autonomous reflection on students’ own values, with a focus on the sources likeliest to undermine autonomy in their particular circumstances: For example, materialist values of mainstream culture and the values of the religious tradition in which they are raised. The public-centered argument might call for civic engagement opportunities, whereas the student-centered argument might tell in favor of exposure of students to sincere and enthusiastic proponents of a diverse range of religious and moral outlooks. While both can justify autonomy education as a clear public funding priority, only the latter argument seems well-placed to demand it on behalf of each particular child. Educational resources are scarce, and tradeoffs will need to be made. How are we to weigh the educational mandates that derive from the two sources of justification for autonomy education?

This questions and others warrant further exploration. A first step, though, is to evaluate perennial legitimacy challenges to both public- and student-centered arguments. I hope to have shown that, notwithstanding the pull of those challenges, both types of autonomy education can be justified without violating the neutrality constraint on liberal legitimacy.

I have not argued that we must abide by the constraints of political liberalism, or that we must formulate neutral versions of our arguments for educational policy. The point, rather, is that if those requirements apply, we can meet them. Teaching students to be autonomous is controversial, but is nonetheless consistent with the liberal demands of mutual respect: Autonomy education secures important public goods, and we owe it to students themselves.