Jump to ContentJump to Main Navigation
The Demands of Liberal Education$

Meira Levinson

Print publication date: 2002

Print ISBN-13: 9780199250448

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0199250448.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 29 January 2020

The Development of Autonomy

The Development of Autonomy

(p.36) 2 The Development of Autonomy
The Demands of Liberal Education

Meira Levinson (Contributor Webpage)

Oxford University Press

Abstract and Keywords

Discusses the relationship between the exercise and development of autonomy and analyses their implications for liberalism and liberal education. Section 2.1 proves that since the liberal state values adults’ exercise of autonomy, it must also value children's development of autonomy. Section 2.2 argues that state paternalism towards children, in particular, state efforts to help children develop the capacity for autonomy even against their parents’ wishes, is consistent with liberal principles. Section 2.3 argues that parents have privileges rather than rights to control their children's upbringing, but that these privileges nonetheless offer ample scope for parents to exert independent paternalistic control. Finally, Section 2.4 argues that the liberal ideal of autonomy not merely permits but requires the intrusion of the state into the child's life, specifically in the form of compulsory, ‘detached’, autonomy‐driven schools.

Keywords:   autonomy, children, liberal education, liberal state, parents, paternalism, rights, school

In order for individuals to act autonomously, they must first have developed the capacity for autonomous action. This is a trivial statement; I might as well have pointed out that in order for people to tie their shoes, or to drive to Chicago, they must first have developed the capacity to tie their shoes and to drive. But it is curious that while contemporary liberal theorists often discuss the exercise of autonomy, and debate which rights and freedoms are most conducive to the achievement of autonomous action, they rarely confront in similar detail its empirical precursor—the problem of how to foster the development of autonomy. This problem, and the indispensability of formal, state‐regulated schooling to its solution, is the focal point of this chapter.

2.1 Does Development Follow from Exercise?

The Development of Autonomy Exercising versus Developing Autonomy

In Chapter 1, we saw that the liberal state is committed to establishing and protecting the conditions necessary for citizens to exercise autonomy. This follows from its weakly perfectionist commitment to individual autonomy as a moral good. At first glance, therefore, it may seem fairly obvious that the liberal state will (or should) also attempt to establish the conditions necessary for individuals—and especially children—to develop the capacity for autonomy. The argument for this is straightforward, based on an extension of Ronald Dworkin's argument quoted in section 1.3 about our reasons for exercising certain capacities. My contention there was that the simple fact of possessing the capacity for autonomy (our ‘second moral power’) does not in itself give us any reason to exercise it, i.e. to act autonomously. We all have many personal capacities that we choose to leave undeveloped or even to suppress—capacities for anger, for jealousy, for advancing our social standing at the expense of others, for self‐absorption, and so forth. Our merely possessing these capacities does not give us a reason to exercise them. What does give us reason to exercise a capacity is our belief that its exercise has a certain worth and value—that is, that the capacity represents (p.37) a substantive good. Thus, the argument went, the liberal cannot merely invoke the human possibility for autonomy and argue for liberal freedoms on that basis alone; rather, he must treat the capacity for autonomy as an actual human good that the liberal state must value and foster. The extension of this argument regarding development is quite simple. If the exercise of autonomy is valuable, then its development must also be a good. Mirroring the contrast above, it does not make sense to admit that there is value in exercising autonomy, but to deny that there is value in developing it (which development is, obviously enough, a precondition to its exercise).1 As Amy Gutmann argues, ‘The same principle that requires a state to grant adults personal and political freedom also commits it to assuring children an education that makes those freedoms both possible and meaningful in the future.’2 Thus, if the liberal state so values autonomy that it works to promote individuals' exercise of it, the state should also aid people in developing their capacity for it.

The simplicity of this argument fails to engage, however, with the fundamental differences—even contradictions—between the conditions required for advancing the exercise of autonomy and the conditions demanded for its development. These arise because the conditions required for the acquisition of a skill or disposition are both logically and empirically separate from the conditions required for its exercise. Consider, for example, shoe‐tying. Learning how to tie shoes requires different skills from tying them. Some characteristics, such as having control over one's hands and fingers, are required in common. Other characteristics, however, may be essential to one but utterly irrelevant to the other. Having the power to concentrate and to assimilate information is necessary if one is to learn how to tie one's shoes, but is then purely academic for those ‘in the know’. Conversely, possessing shoes is a prerequisite to being able to tie them, but one can learn how to tie shoes (as I did under the patient tutelage of my mother) using a shoebox threaded with yarn. Finally, there are even cases in which the demands of learning to tie one's shoes conflict with the demands of tying them. The habit of thinking through every action carefully and methodically might be a tremendous aid to learning how to tie one's shoes; but someone who was never able to get past such methodicalism would inevitably lack a certain fluency in shoe‐tying that more instinctive or intuitive actors achieve.

Thus we see, through the lens of this (admittedly pretty basic) example, that the development of a capacity may pose very different requirements from its exercise.3 This is the case, I suggest, with autonomy—and it is this case that makes justifying even weak state perfectionism concerning the development of autonomy much harder than it first appears. The conditions and institutions required for the exercise of autonomy are ‘built into (p.38) the system’ of liberalism in a way that the conditions required for the development of autonomy are not. Recall from Chapter 1 that contemporary liberalism has three distinct defining elements: (A) an acceptance of the fact of pluralism; (B) a set of formal or procedural conditions demanding free, equal, and public agreement on basic political principles as the precondition for political legitimacy; and (C) a set of substantive institutional conditions, including constitutional democracy, broad individual freedoms and civil liberties, and sometimes some measure of state welfare provision or income redistribution. We saw in section 1.3 that autonomy‐based weak perfectionism best accommodates and justifies these commitments by providing a principled basis for the procedural conditions specified by (B) and generating the substantive claims regarding freedom and democracy laid out in (C). In addition, weak perfectionism responds to (A), the fact of pluralism, both by upholding the value of autonomy (and thus of certain forms of pluralism) and by showing equal respect and consideration (via the guarantee of equal citizenship) to heteronomous and autonomous adults alike. When we turn our attention to the development of individual autonomy, however, and especially to the notion that the state ought to promote children's development of autonomy, the liberal legitimation conditions seem to fit together much less well.

To begin with, for the state to foster children's development of autonomy requires coercion—i.e., it requires measures that prima facie violate the principles of freedom and choice affirmed by (C). For example, children are granted none of the rights and freedoms that I mentioned in section 1.1 as centrepieces or hallmarks of liberalism, such as democratic participation, the right to own and exchange private property, freedoms of speech, conscience, religion, and association, and the right to privacy. If children are permitted to vote, for example, it is feared that they may be taken in by politicians who promise them enticing freedoms and services now but who in the long run will (intentionally or simply out of incompetence) dangerously reduce their and our opportunities for choice. Unable to judge competently the claims made by competing politicians, children are thought to be made more free or autonomous in the long run by being denied the vote until their capacity for judgement is more fully developed. Likewise, full freedom of association, rights to private property, and even freedom of movement are denied to children in their own interest. The fear is that children who are permitted these freedoms—to choose their own friends without any guidance, for example, or to ‘hang out’ whenever and wherever they wish—run the risk of making choices that in the end reduce their capacity for autonomy and stunt its development.4 Even the twin virtues of transparency and publicity, so vaunted by liberals at every stage of government legitimation and action, are thought (p.39) to be justifiably modified at times in relation to childhood education for autonomy.5 Furthermore, not only are children denied liberties routinely granted to adults; they are also subjected by the state to regulations—such as compulsory schooling (state or private)—from which adults are utterly exempt. While such institutions ideally aim in the long term at freeing children (by helping them to develop a capacity for autonomy), they still involve massive interference with and restrictions over children in the meantime—and as one author reminds us, ‘We think that the purpose of the child is to grow up, because it does grow up. But its purpose is to play, to enjoy itself, to be a child. If we merely look to the end of the process, the purpose of all life is death.’6

The coercive nature of state promotion of the development of autonomy also means that children do not have the luxury of ‘opting out’ of public autonomy‐advancing opportunities in the same way that adults do. This is true in three ways. First, insofar as adults are confronted by political freedoms and opportunities that advance the exercise of individual autonomy, they can simply decide not to take advantage of them, with no loss to or change of behavior required from either themselves or the state. The state can continue to make freedoms available to all adults, and adults can continue to make individual choices about making use of the opportunities—choices, too, which they are free to change at any future time.7 But if children must be actively coerced by the state (through compulsory schooling, denial of liberties, and so forth), then there is no possibility of retreat. Given that children are viewed as not having the capacity adequately to judge their own interests, it would not make sense for the state to decide that children's interests lie in the development of a capacity for autonomy yet allow them to refuse to develop this capacity. The state may value the exercise of autonomy yet respect those adults who choose not to be autonomous; it cannot treat children the same way. Thus, in the end, children are far more profoundly affected than adults by the state's weakly perfectionist valuation of autonomy.

Second, children's co‐optation is compounded by the fact that insofar as the state decides for children that they should develop the capacity for autonomy, it excludes them from the procedural participation that is deemed essential for political legitimacy. The liberal state values the exercise of autonomy for adults, in a certain sense, because adults agree (in theory) that it is in their interest for the state to do so. They have chosen, through the procedures specified in (B), for the state to promote their exercise of autonomy in the public sphere. And, as just discussed above, adults who prefer to live heteronomous lives can do so within (albeit in spite of) the autonomy‐advancing political culture. Children, in contrast, have made and can make no such decision about the state's weakly perfectionist (p.40) commitment to the development of autonomy. Unable freely to participate in the public political dialogue, they are necessarily excluded from the legitimation of autonomy (or of anything else) as a political goal. Trapped within the confines of profound political inequality, children are subject to state paternalism despite their never having participated in a procedure legitimating this relationship.

Third, the compulsory development of a capacity for autonomy—even if the child is never forced to exercise his autonomy as an adult—entails other consequences that may fall outside the state's legitimate sphere of influence. Once a child develops the capacity for autonomy, it becomes impossible for him to hold certain conceptions of the good in the same way that would previously have been available to him. As we saw in relation to Rawls' ‘burdens of judgement’ in Chapter 1, even passively possessing the knowledge and dispositions that underlie the capacity for autonomy—where such dispositions include toleration of others, values pluralism, openness to criticisms from others, and the recognition that reasonable people can hold incommensurable and even incompatible conceptions of the good—radically alters the way that one approaches certain professions of value and belief. Nomi Stolzenberg expresses this point powerfully in discussing fundamentalist Christian parents' concerns about ‘value‐neutral’ state education:

fundamentalists are not concerned only with the case in which their children unequivocally reject their values; they are also concerned with the case in which their children remain attached to their parents' views, but only after coming to see those views as such—as subjective, contestable matters of opinion. There is a subtle but important difference between the faith that is innocent of alternatives and that which is not.8

She expands this point later:

The point is not simply that the objective mode of exposure exhibits options, or even that it encourages rational selection from them. It is that even if the children [continue to] adhere to their parents' beliefs, they do so knowing that those beliefs are matters of opinion. This knowledge enhances the likelihood that children will form their own opinions and deviate from at least some of their parents' beliefs. It also transforms the meaning of remaining (or in the case of children, becoming) attached to them. It is one thing for beliefs to be transmitted from one generation to another. It is another to hold beliefs, knowing that those beliefs are transmitted, that they vary, and that their truth is contested.9

These same children and parents might highlight a final conflict in the liberal state's commitment to helping children develop autonomy by questioning the justification for the state's interference in the essentially private matter of raising children in the first place. The family is a pre‐political (p.41) institution, they might suggest, whose integrity cannot legitimately be violated by the state. The family cannot be made subject to political and state control without doing violence to our understanding both of the separation between public and private—which separation is essential to liberal theory in general, and to an autonomy‐based liberal weak perfectionism in particular—and of the primacy of the family over the state. In addition, it might be argued, children's interests (and human interests in general) are inseparable from the institution of the family; thus, it is nonsensical and misguided to try to assert a child's right to autonomy from outside the family sphere. These are powerful and to a large extent intuitively plausible claims in favor of family independence from political pressure and control; any political theory which calls for state interference in children's (and therefore families') lives, as liberalism does, must therefore take them very seriously indeed.

In conclusion, some of the most attractive elements of liberalism's substantive conditions—that they are compatible with heteronomous agents and heteronomous ways of life; that they permit individuals a private sphere of life into which the state does not intrude; and, most fundamentally, that they are principles of liberty and not of coercion or paternalism—are lost when liberalism declares the state's obligation to promote not just the exercise of autonomy but its development as well. State involvement in fostering children's development of autonomy violates pluralism, imposes restrictions on people (children) who were not participants in the legitimation process, denies them basic liberties, violates the terms of weak perfectionism by forcing children to develop the capacity for autonomy, and infringes the public/private divide by interfering with the family. In all of these respects, liberalism seems incompatible with a political commitment to children's development of autonomy.

2.2 Paternalistic Control and the ‘Fundamental Interests’ Of the Child

The Development of Autonomy Paternalism and Children' Interests

The case for the compatibility of liberalism with state perfectionism concerning the development of autonomy looks pretty bleak, given the latter's multiple violations of both the procedural and the substantive tenets of liberalism. From all appearances so far, liberalism will at a minimum have to divorce its commitment to promoting the exercise of autonomy in adults from any efforts to help children develop autonomy—although as the Dworkinian argument that opened the chapter pointed out, it is probably conceptually incoherent to value the former as a good in itself (as the exercise of autonomy is taken to be) without thereby (p.42) valuing the latter as well. Thus, if the criticisms articulated in section 1 are right, then it may be that liberalism's entire commitment to autonomy should be re‐evaluated. In this section, however, I shall argue that the above criticisms are misleading, and that liberal theory is compatible with obligating the state to promote in children the development of the capacity for autonomy. My response will come in a slightly odd form, as I believe that the above problems are actually best confronted by responding to a further set of criticisms, this time from the children's liberationist camp. I hope that by working through the arguments set forth by children's liberationists, we will come to understand the relationships among the autonomy‐oriented liberal state, adults, and children much better. This analysis will also illuminate how and why liberalism is compatible with the paternalistic violation of conditions (A), (B), and (C) mandated by the development of autonomy.

Before we step into the paternalism minefield, let me introduce one piece of symbolic notation and clarify one definitional point. First, because it is verbally cumbersome constantly to distinguish between paternalism in general and paternalism specifically toward children, I will use ‘paternalismc’ to mean ‘paternalism toward children’, and ‘paternalisma’ to mean ‘paternalism toward adults’. This will make the discussion cleaner and less convoluted. Second, paternalism in general needs to be defined. At this point, I will take it loosely to mean the restriction of someone's freedom for their own good. Insofar as Chapter 1 is right that it is autonomy that the liberal ultimately views as valuable in individuals' lives, ‘their own good’ must therefore be understood, at least within the context of liberal theory, as relating to the capacity for autonomy. Utilitarianism, desire‐satisfaction theory, and so forth are thus irrelevant to a liberal conception of paternalism, and to liberal paternalists' conceptions of others' good. There are a number of unacknowledged complexities lurking within this characterization, and I myself will suggest a modified understanding of paternalism below, but this definition should suffice for present purposes.

The problems with involving the state in fostering the development of autonomy, as discussed in the last section, can in part be summed up by the statement that, as Howard Cohen puts it, ‘There is one set of rights for adults and another for children. Adults' rights mostly provide them with opportunities to exercise their powers; children's rights mostly provide them with protection and keep them under adult control.’10 It is children's subjection to others that seems to violate the procedural and substantive conditions of liberalism. Of all children's liberationists, Cohen offers one of the most cogent articulations of this problem in his Equal Rights for Children. He starts from the widely recognized universalizability principle of social justice, that we must treat people equally unless there exists a (p.43) morally relevant distinction among them that justifies unequal treatment. Thus, all people deserve equal rights (whatever those rights may be) unless there is a difference between them that is relevant to possessing the right in question. When we consider differential treatment toward adults and children, therefore,

If there are differences between adults and children which are relevant to granting rights to one group but not the other, then it [differential treatment] is not unjust. But if the differences are not relevant, then the two groups are, and should be, treated alike. . . . When we call for equal rights for children—for an end to the double standard—we want the elimination of ‘child’ as a separate category in all aspects of life where the distinction is not relevant.11

Applying this to the problem of paternalism, we might ask what it is about children that makes paternalismc justified; or, conversely, we might raise the question of why adults are assumed to have rights against paternalistica coercion. In other words, on what grounds does the state base the power of self‐governance in adults? When the question is approached from this angle, the answer becomes obvious. From the standpoint of liberalism, paternalisma is wrong because adults' exercise of autonomy (which would be violated by the state's exercise of paternalisma) is valuable and politically relevant. This was one of the important conclusions of Chapter 1: liberal rights and institutions—including adults' state‐protected liberties and rights against undue coercion—are justified because adults cannot exercise autonomy in the absence of such liberties, and the exercise of autonomy is a valuable enterprise. Paternalisma by the state is outlawed where it would violate citizens' autonomy.12

The presupposition of this chapter, on the other hand, is that children are not autonomous—if they were, then our attention to developing autonomy in children would be both misplaced and unnecessary. If children do not possess autonomy, and the source of liberal rights against paternalistic interference is the value of the exercise of autonomy, then children seem not to merit rights against paternalism equal to those merited by adults. The distinction between children and adults is relevant to the rights in question; to impose paternalistic restrictions on children, therefore, seems to be compatible with social justice. As Locke adds in the Second Treatise, ‘Thus we are born free as we are born rational; not that we have actually the exercise of either: age that brings one, brings with it the other too. And thus we see how natural freedom and subjection to parents may consist together, and are both founded on the same principle.’13

Once the liberal admits that adults are protected against paternalisma only because of the value attached to their exercise of autonomy, however, he opens himself up to the most probing criticism of paternalismc: namely, (p.44) that children are actually just as autonomous as adults are—or (on the more pessimistic and equally damning view), that many adults have just as little capacity for autonomous thought and action as children have.14 Another prominent children's liberationist, John Harris, puts the point bluntly:

So if it is supposed that it is the comprehensive possession by adults of capacities lacked equally comprehensively by children that sustains and justifies the political disabling of children and the rule of adults, the supposition is false. False because there are numerous children whom it would be implausible to regard as incompetent and numerous adults whom it would be implausible to regard as anything else. Whatever we may finally decide about the status of children it is unlikely that the ‘facts of human development’, in so far as they are ascertainable, will support the rigid ‘age of majority’ division between adults and children.15

It is important to note that this argument against differential treatment of children and adults can succeed even if the previous criticisms fail. It might be right that the development of autonomy requires paternalism, this argument goes; and it might even be right that the relevant criterion for having certain rights is having the capacity for autonomy. But, Harris's quotation suggests, neither of these admissions gives us any reason to believe that paternalismc is justified when paternalisma is not. Autonomous people simply cannot be sorted out on the basis of age; the two characteristics are logically independent and empirically separate. Thus, if justifications for paternalism are linked to autonomy, then paternalism cannot be linked to age. If paternalism is linked to age, then it cannot be justified on grounds of helping the heteronomous. And so, unless we are willing to accept widespread paternalisma, abandon democracy, and submit ourselves to a ruling elite with the power to decide about every individual whether she is autonomous and therefore free or heteronomous and therefore subject to paternalistic control, we must reject both paternalisma and paternalismc at once as being two sides of the same coin.16

This is an important argument. Insofar as Chapter 1 is right, that the institutional conditions of the liberal state are dependent on the state's recognizing the value of the exercise of individual autonomy, then the fact that many adults do not exhibit the capacity for autonomy does threaten its foundations. The liberal cannot permit the state to decide on a case‐by‐case basis who is autonomous and who is not, for tyranny would rapidly result. Massive numbers of adults would end up being denied fundamental liberal rights—or would be if we treated them like the children to whom they turn out to be similar.17 Also, there is a real fear that if the state were given the power to decide who possesses autonomy and who does not, it could establish arbitrary and coercive requirements that would in essence give it total (tyrannical) power over all individuals' lives. At least under systems current in Western liberal democracies, people have full freedom from age 18 or so (p.45) onwards, even if they may have been unjustly restricted on paternalisticc grounds before then. In the state currently under discussion, people could conceivably be restricted for their whole lives, simply because they fail to satisfy the state that they deserve freedom (because they voted for the wrong political party? because they are members of a despised religion?).

To avoid such collapse into tyranny, liberalism could abandon its links between rights and autonomy—although as Chapter 1 demonstrated, liberalism would lose much of its plausibility and normative weight if it did so. Or, more successfully, the liberal state could operate according to a presumption of autonomy, as opposed to a presumption of its lack; i.e., the state could assume that all citizens have the capacity for autonomous action unless there were good reason to think otherwise. As John Kleinig suggests, ‘competence is presumed until disproved’.18 The advantage of this move is that it has strong empirical as well as theoretical justification. From the perspective of liberal theory, the presumption of autonomy sets the universal provision to adults of substantive rights and freedoms back on a firm foundation, thus pulling liberalism back from the risk of dissolving into a blueprint for tyranny. And from an empirical perspective, there is a group about which there is every ‘reason to think otherwise’ about their possessing the capacity for autonomy—namely, infants and young children. It would be ludicrous to argue that four‐year‐olds—or even most ten‐yearolds—have the capacity for autonomy in the complex sense given in section 1.3. This is not to say that ten‐year‐olds could not take care of themselves, or do not have articulated conceptions of the good. But it is to say that because of a combination of lack of experience, underdeveloped critical reasoning skills, insufficient emotional independence, and so forth, they do not yet have the capacity for full autonomy. Thus, while the liberal starts out by crediting all individuals with the presumption of autonomy, she does have a good reason to withdraw that presumption from young children. In other words, the presumption of autonomy is justifiably age‐sensitive.19

Once this step is admitted, and its full significance recognized, then the rest of the argument for (some as yet‐undetermined measure of) paternalismc falls into place. Given that no four‐year‐old possesses autonomy, we have the unproblematic right paternalistically to aid the development of autonomy in the young; at the same time, we are forbidden from imposing similar compulsory measures on adults (who are assumed to possess the capacity for autonomy). The problem then becomes one of deciding where to mark the transition between the heteronomous young and the presumptively autonomous not‐young—a task which, while difficult and highly contentious, is thoroughly different from trying to justify paternalismc altogether.

Here instrumental considerations can legitimately be brought to bear. As (p.46) was discussed above, it is both too impractical and too dangerous (in terms of giving the state too much power) to distinguish youth from adults by any other means but establishing a standard age (from which exemptions could possibly be sought).20 Even if we had no fear of tyranny, it would be a tremendous burden on the state, and undoubtedly a disservice to the young themselves, to impose autonomy tests on all children every year. Thus, we must simply, and to a certain extent arbitrarily, find some standard age(s) to mark the transition from youth to non‐youth, and therefore from (some measure of) justifiable paternalismc to protection from paternalisma. To argue for standardization is not, however, to argue for a single age of majority. Given that the development of autonomy is itself a gradual process, there can and probably should be a gradual transition from youth to adulthood in which different rights and privileges are assumed at different times. People might gain the right to choose their doctor at age 12, to consent to major surgery at age 14, to withdraw from school at 15 or 16, to run for the US Senate at 30, and so forth. These ages are obviously contestable, malleable, and at least locally arbitrary21—but some set of standard ages must be chosen. So, even though it is true that ‘Bold, quick, ingenious, forward and capable young people are by no means a rarity, [and] neither, unfortunately, are dull‐witted, incompetent adults’,22 we have no reason to reject paternalism toward children tout court, nor are we obligated for the same reason to abandon liberalism in favour of strict paternalism toward adults.

While this local victory may come as something of a relief to autonomy‐valuing liberals, it would be a mistake to think that in answering the criticisms posed by the above children's liberationists, liberalism's task of justification is complete. Our association of paternalism with autonomy (and its violation) left unaddressed three criticisms from section 1: first, that children are wrongly excluded from the procedural legitimation of the state (including the legitimation of the state's guiding aims, of which the development of autonomy is one); second, that because of the conceptual shifts forced by gaining the publicly promoted capacity for autonomy, children lose the ability privately to hold certain valued conceptions of the good—and that this loss can never be recovered, even in adulthood, where the public/private distinction is meant to be respected by the state; and last, following on from these public/private concerns, that paternalistic control over children should be left to the family instead of being assumed by the state, in acknowledgement of the family's integrity as a private institution.

These criticisms raise two fundamental questions that the liberal must address. First, does the child have any substantive interests (e.g. in developing the capacity for autonomy) independent of those acknowledged by (p.47) the individuals who take responsibility for her—or even by herself? In other words, do liberal principles about adults tell us anything about hypothetical interests held by the child that limit what can be done in the name of her own good? Second, who has the right more generally to determine the child's best interest, especially in cases of conflict? Does the identity of the paternalistic agents matter? Should we take the claims of parents, for example, more seriously than we take the claims of other people or institutions who are interested in the child's welfare?23

In regard to these questions, there are two points which deserve to be made immediately. First, we must realize that the child is going to be the subject of paternalistic coercion regardless of who is determining her best interest or what that interest is judged to be. Paternalism and limitations on the child's freedom are required to foster the development of her autonomy, it is true; they are also required to inculcate in her a restricted, heteronomous conception of the good. The child is inevitably treated paternalistically, by her parents, the state, concerned adult friends and relatives, teachers, doctors, and so forth. What is at issue, therefore, is not whether paternalism in general is justified, but who has the right to determine its nature and aim. Second, it is simply false to suggest that as a private institution, the family is or should be completely independent of the state. None of us believes that parents have the right to do anything they wish to their children. The state has to have some notion of the child's fundamental interests, including her future interests, and has some right to interfere in the family based on this account. The task then is to spell out what that conception of the ‘child's fundamental interests’ is; it is not to justify the state's holding such a notion in the first place.

Given these two facts, the state's fostering the development of autonomy in children becomes easier to justify than it seemed at first blush. Consider a case in which an ‘intolerant’ parent raises his child to be heteronomous, believing that to be most conducive to the child's good, whereas the state believes that it is in the child's best interest to develop the capacity for autonomy. Both agents are paternalistically motivated, interested solely in the child's own benefit, and both parties assume (rightly, let us suppose) that the child does not already have an authentic developed conception of the good. Which—or whose—perception of the child's good has the right to trump? Well, as I just noted, both parties (the state and the parents) are justified in taking an interest in the child, and in positing and defending a substantive notion of the child's future interests. In addition, both parties will enact their notions through the use of paternalistic coercion. The child has no more control under the parent than she does under the state; in this sense, both of the first two criticisms raised above concerning the child's lack of choice are equally applicable to both prospective paternalistic (p.48) agents. Thus from the child's (hypothetical) point of view, it is morally arbitrary who—the state or the parent—exerts coercive control over her. Both are in equal need of justification. As Gutmann comments, ‘We have no a priori reason to favor one paternalistic agent over another.’24

Given that it is morally arbitrary from the child's perspective who the paternalistic agent is, the conflict must be resolved in terms of what paternalistic doctrine is more in the child's interest.25 Here again the notion of choice becomes relevant. As we saw in Chapter 1, one of the primary sources of legitimacy in liberal theory is individual agreement or consent; this reliance on consent is symbolized by the legitimation conditions introduced in section 1.1 and cited throughout that chapter. Children have no opportunity for choice when young. Their lives and their values are inevitably (and rightfully) shaped by others. But the most legitimate basis for this coercion, on a liberal reading, would seem to be that which gives children the capacity for choice—the capacity to overcome the bounds of coerciveness—later on in life. The way to transfer power over their own lives back to individuals, of course, is to help them to develop a capacity for autonomy. Thus, a paternalistic doctrine that favors the development of autonomy must be judged superior (continuing with the child's hypothetical perspective) to one which inculcates heteronomy. The state is justified, therefore, in helping children to develop the capacity for autonomy, even against parents' and children's expressed wishes. In promoting the development of autonomy, the state is not claiming that the autonomous way of life is the only legitimate way of life, or that autonomy is a prerequisite for citizenship. It is simply trying to right the balance of power by giving individuals the ability in their adult lives to do what they could not do as children—specifically, to determine their own values and to adopt a conception of the good with which they identify (as opposed to those with which they happened as children to be identified).26

The autonomy‐fearing parent introduced above may still protest that his child's choice is stunted because she will no longer be able to adopt her parent's conception of the good in the same way she might have before she developed her capacity for autonomy. This was the thrust of the second objection above. It is likely that there is no liberal answer that will fully satisfy this parent—for which reason I said in Chapter 1 that autonomy‐based weak perfectionism provides the best, but not a perfect, means of simultaneously satisfying (A), (B), and (C). There is nonetheless a response that satisfies the demands of weakly perfectionistic liberal theory, which is what matters for my purposes. The parent's criticism, I suggest, is effectively countered from a liberal perspective by Joseph Raz's insight that ‘while autonomy requires the availability of an adequate range of options it does not require the presence of any particular option among them’.27 The (p.49) liberal state must permit individuals to lead heteronomous lives because ‘denying a person the possibility of carrying on with his projects, commitments and relationships is preventing him from having the life he has chosen’. But this toleration of heteronomy need not carry over to children: ‘A person who may but has not yet chosen the eliminated option is much less seriously affected. Since all he is entitled to is an adequate range of options the eliminated option can, from his point of view, be replaced by another without loss of autonomy.’28

Again, this response will presumably not satisfy the autonomy‐fearing parent, who is concerned that his child lead the right way of life, not that his child be able to choose among a number of what are by his definition bad lives. The point of this chapter, however—and largely of the book as a whole, as I discussed in the Introduction—is not to justify liberalism to all illiberals, which I do not think can effectively be done. In a deeply plural society no set of liberal principles and institutions will gain free, unanimous consent; not all individuals will become liberals. Insofar as autonomy‐based weak perfectionism provides the strongest and most widely acceptable justification for liberal principles, it also provides a solid justification for protecting children's range of options without necessarily protecting any particular option—even if some parents do not agree. Furthermore, I suggest that just as the liberal state need not adjust its behavior when certain cultures find it hard to flourish as a result of the atmosphere created by liberal rights and freedoms, neither need the state be overly concerned about cultural attrition caused by its advancement of the development of autonomy in children. If people who have gained the capacity for autonomy choose to leave the communities in which they were raised and to develop their lives in different directions, then while this outcome might be unfortunate, it is not a violation of the state's legitimacy or of any of the principles for which it stands.29

It is worth making explicit at this point an argument that has thus far been implicitly developed: namely, that liberalism is not, and cannot consistently be, a neutrality‐driven theory. Chapter 1 made this argument in relation to adults, suggesting that neither acceptance of the fact of pluralism nor commitment to the legitimation conditions can effectively justify traditional liberal freedoms and institutions without a non‐neutral, weakly perfectionist embrace of the value of individual autonomy. Even if one were to reject that claim, however, the fact that children are inevitably subject to paternalistic coercion also deals a blow to liberalism as a neutralist theory. For insofar as a child's best interests must be determined by actors other than the child herself, the liberal state may either participate in determining the content of children's interests, or abdicate responsibility by allowing parents, family members, communities, secondary associations, (p.50) etc. to make all the decisions about the child's upbringing. If it does the former, the state is obviously not acting neutrally in the sense favored by political (neutrality‐based) liberals.30 In the latter case, the state might plausibly be described as neutral, but this form of neutrality can hardly be desirable insofar as it amounts to tacit consent to the coercion of the weaker (child) by the stronger (parent/family/community).31 The idea that abdication of this kind is a desirable form of neutrality can be traced, I believe, to the misapprehension that liberalism requires the state to be neutral among competing conceptions of the good. But if anything can be said to be entitled to neutral treatment, it is not the conceptions of the good, but rather the people who hold these conceptions. Moral doctrines or value systems have no independent political standing; they are not agents or actors or citizens. What liberals really mean to say is that the state should be neutral among individuals holding competing conceptions of the good. Children, however, do not have developed conceptions of the good. Thus, neutrality seems not to apply to the relationship between the state and children, except insofar as the state's involvement with children affects parents' abilities to pursue their own conceptions of the good. And, as I argue in the following section, parents' inclusion of children in the realization of their own conceptions of the good is again unjustified to the extent that it amounts to the coercion of the weaker by the stronger. If the principle of neutrality does not justifiably govern the relationship between the state and children, of course, then this has implications for the status of neutrality between the state and adults as well.

2.3 Justifications for Parental Control

The Development of Autonomy Justifications for Parental Control

It may seem that I have now moved from one unattractive philosophical extreme to the other. Having started out by examining suggestions about the inseparability of the interests of the child from the interests of the family, I have ended up with what appears to be a radical articulation of the child's independence from the family and of the state's overriding interest in forcibly developing children's capacities for autonomy.32 In this section, I argue that although it is true that from the child's perspective there exists ‘no a priori reason to favor one paternalistic agent over another’,33 there are nonetheless compelling, child‐centred reasons to grant primary control and child‐rearing responsibility to the parent(s) as opposed to other agents.34 I thus argue in favor of a presumption of the parents' privilege rather than right to govern their children's upbringing. As James Dwyer succinctly suggests, ‘From a moral [and political] perspective, a parental privilege would not convey or reflect a sense of entitlement to direct a child's life, but (p.51) instead would bear the aspect of a benefit contingent upon the fulfillment of attendant responsibilities, much in the nature of a trust.’35 I suggest in this section that establishing parental privilege over the care and governance of children best respects children's needs and interests in developing their capacities for autonomy without unjustifiably infringing upon parents' autonomy.

First, there is a strong argument in favor of straightforward parental control deriving from what Judith Shklar has termed the ‘liberalism of fear’:36 namely, that a state with the strength to enforce any other means of distributing control over children would inevitably be tyrannical in all aspects of governance. As Gutmann comments, ‘liberals justifiably fear a state so powerful that it could, as a matter of routine practice, take all children away from their biological parents’.37 The desire of parents to raise and care for their children is probably one of the most innate and universal of all human desires. It would take extraordinary physical strength and/or massive psychological manipulation in order to dispossess parents wholesale of their children and assign caring responsibility to someone else. All people, adults, parents, and children alike, would suffer in a state possessed of such power, since it is inconceivable that such a state would limit its actions to justified, non‐tyrannical behavior. Following on from this recognition of parents' psychological link with their children, a secondary argument might also be made that the incentive to produce children is drastically reduced if biological parents run the risk of being deprived of their children at birth. While this suggestion, if true, gives extra weight in favor of biological parents' rights to keep their children, I believe that the first argument alone is enough to shift the balance of power from the state (or a similar redistributive agency) to biological parents.

That biological parents (or their surrogates) normally have the right to keep their children in the family, so to speak, does not necessarily give them rights, however, to determine the terms of paternalistic control. This in part is a restatement of the conclusion of the previous section. Drawing upon the recognition in that section that the identity of the paternalistic agent is morally arbitrary from the point of view of the child, there seems to be no reason in principle that a committee consisting of the state and other self‐appointed interested adults might not demand to make collective judgements in tandem with the parents about the child's best interests. Locke reminds us of the ephemeral source of parental authority: ‘The power, then, that parents have over their children arises from that duty which is incumbent on them, to take care of their offspring during the imperfect state of childhood. To inform the mind, and govern the actions of their yet ignorant nonage, till reason shall take its place and ease them of that trouble, is what the children want, and the parents are bound to.’38 (p.52) There is nothing in this passage to suggest why it is only the parents who may judge how to ‘inform [their child's] mind’, or to ‘govern the actions of their yet ignorant nonage’. If others can govern better, why should they not be granted partial control? Parents would still be allowed to keep their children and act as primary caretakers, but decisions about child‐rearing would be subject to collective deliberation. A second set of arguments, therefore, is needed to justify parents' exerting primary control over the child's upbringing.

These arguments divide into three types: parent‐autonomy‐regarding, pragmatic or psychological, and child‐autonomy‐regarding. The latter two provide strong liberal justifications for parent privilege; parent‐autonomy‐regarding arguments, on the other hand, make uncompelling and illiberal cases for parents' rights over their children, and I address and dispense with these first.

From the perspective of parental autonomy, three arguments support the assumption of parental rights to paternalistic control over their children. The first of these is fairly simple: to rear a child may be a joy, but it is also a tremendous burden. From the day‐to‐day worries about day care and baby‐sitters to the soul‐rending crises of wondering ‘Am I doing the right thing?’ or ‘Where did I go wrong?’, child‐rearing is not an easy business. Thus it is suggested that in exchange for taking on such a severe burden, parents have a right to raise their children in a way that will least disturb their own lives or cause conflicts with their own visions of the good.39 Given that somebody will have to shape the child's moral outlook and mores for living, parents deserve to take primary control so that the child's mores do not cause unnecessary rifts or tears in the fabric of daily family life. This argument is parent‐regarding in the weak sense that it appeals to a notion of just compensation, reimbursing parents in the form of rights to paternalistic control for what they lose in personal freedom and independence. Although this argument does give some weight to the notion of parental privilege to control their children's upbringing, it does not, however, justify the assignment to parents of a right to do so. Especially insofar as child‐rearing is a burden (activity, joy, etc.) that adults in Western, industrialized countries can choose to take on or reject (via birth control before conception, and abortion or adoption after the fact), parents do not merit the right to control another person's future as compensation for a burden voluntarily shouldered.

A more interesting and controversial parent‐regarding argument depends on a much more robust understanding of parental rights and parental autonomy. It is suggested that an essential part of one's autonomy is the right to pass one's values onto one's children. We cannot fully and autonomously enact our conception of the good life, under this reading, (p.53) unless we have the right to share our conception with others, and specifically to initiate our children into this same vision. As Charles Fried argues, ‘the right to form one's child's values, one's child's life plan and the right to lavish attention on the child are extensions of the basic right not to be interfered with in doing these things for oneself’.40 A striking example of this claimed inseparability between one's own conception of the good and that of one's children is raised in an American case, Mozert v . Hawkins County Public Schools.41 In 1983, a group of fundamentalist Christian parents in Tennessee tried to get permission to remove their primary‐school‐aged children from a reading class in which, they claimed, the mandatory reading text: taught ‘that some values are relative and vary from situation to situation’; included ‘various humanistic values’ that were ‘evil’, ‘polluted’, and ‘heathen’; and induced ‘children to stray from the way of God’.42 When permission to remove their children to the library during the reading period was denied by the school board, the parents sued the Hawkins County Public Schools for violation of the clause in the First Amendment of the US Constitution that guarantees free religious exercise. Although the case as a whole is fascinating (the parents eventually lost43), what is relevant in this context is the parents' argument that independent of the evil brought upon the children by reading the text, the parents themselves faced ‘eternal damnation’44 for permitting their children to participate in the reading program. This is because by doing so they were breaking the biblical commands to ‘learn not the ways of the heathen’ and to ‘have no fellowship with the unfruitful works of darkness, but rather reprove them, for it is shameful even to mention what the disobedient do in secret’.45 In essence, living out their own conception of the good was inseparable from inculcating that conception in their children; to fail to do the latter would be a gross violation of the former, and of everything the Bible and their lives stood for.

While the sincerity of the parents' complaint is compelling, and the conflict not an easy one to resolve, this argument seems highly suspect from the perspective of liberalism. Liberals unanimously agree that the fact that my conception of the good might include your becoming my personal slave, or your affirming my conception of the good, gives neither me nor anybody else any legitimate reason whatsoever to command your participation in or affirmation of my project. Why, then, should we consider this demand to be any more justified when it is parents making the claim in relation to their children? Fried's suggestion notwithstanding, children are widely recognized to be individuals separate from others, and not mere extensions of their parents or their parents' interests. As Kleinig drily notes, ‘Although there is historical support for the view that children are the property of their parents, there is not much else to be said in its favor.’46 (p.54) It is implausible, therefore, that the Mozert parents' interpretation of parental‐autonomy reasons alone could make a sufficient case for parent‐dominated paternalistic control over children—let alone for the stronger concept of parents' rights.

A final, related parent‐regarding argument in favor of biological parents' rights to control their children's lives has been put forward by Ferdinand Schoeman. It rests on the value of intimacy in human beings' lives. Parents have a ‘moral claim to raise their biological offspring in a context of privacy, autonomy [meaning that parents have wide latitude in determining the character and range of their children's experiences as well as their approach to discipline], and responsibility’, not because that is necessarily always in the best interests of the child, but because child‐rearing under these conditions enables parents to experience intimate relationships with their children, which experience of intimacy is itself a basic human good or right.47 Intimacy, Schoeman argues, is a large part of what makes life worth living; it is the opportunity to ‘share our selves’ with others and have them share themselves with us that gives our lives meaning (or, less strongly, that many people see as giving their lives meaning). Thus, Schoeman concludes, ‘the parent's right to raise her children in a family [including privacy, autonomy, and responsibility rights] stems naturally from the right to engage in intimate relationships, even when recognition of this right involves some comparative cost to the child’.48 Eamonn Callan builds on this point, arguing that ‘the freedom to rear our children according to the dictates of conscience is for most of us as important as any other expression of conscience, and the freedom to organize and sustain the life of the family in keeping with our own values is as significant as our liberty to associate with others outside the family for any purpose whatsoever’.49 In other words, the liberal freedoms (C) that I said in Chapter 1 were partially constitutive of the liberal state, and without a commitment to which no theory can be considered liberal, necessarily include the freedoms to follow our conscience and associate intimately with our families in raising our children.

This is an intriguing interpretation of parent‐regarding rights to control their children insofar as it binds children to their parents by blurring the distinctions between the parties' identities (because they ‘share’ themselves with each other) without thereby necessarily asserting that children are extensions or property of their parents. Insofar as we find this conception of intimacy and the role it plays in human lives compelling, it may provide the strongest reason yet to recognize parental privilege to control their children's upbringing. Nevertheless, it is unsuccessful in justifying a parental right to such control. For although liberals (and the liberal state) can coherently recognize the good of intimacy as giving meaning to many people's (p.55) lives, they also recognize that the conditions under which intimacy is established are significant both morally and politically. Intimacy may possibly be conceived as a ‘right’ (which is how Schoeman must conceive of it in order to support his claim of parental rights to an intimate relationship with their children) when it applies to a relationship that two or more parties enter willingly and autonomously. But we do not and should not recognize intimacy as a ‘right’ to force, manipulate, or prey upon the weaknesses of a person in order to establish an intimate relationship. Rape (forced physical intimacy), unprofessional teacher–student relationships (emotional intimacy established by manipulating an extreme difference in power, maturity, and autonomy), and even extreme dependency between two adults (consider Harry from Chapter 1) are all examples of intimate relationships that are rightly legally prohibited and/or morally condemned. To the extent that a person must rely upon some form of manipulation to establish an intimate relationship with someone, therefore, it is simply false that a liberal state would recognize that person's ‘right’ to intimacy.

Insofar as children do not enter the intimate relationship freely or autonomously, therefore, it is difficult to see how parents' production of children in order to have the opportunity to establish an intimate relationship with them gives them a right to such a relationship. It might give them a privilege—based on the parent‐ and child‐regarding assumptions that such relationships virtually always give deep meaning to both parties' lives (and, as I discuss below, that they are essential to children's well‐being and healthy development)—but not a right. This conclusion is especially justified since neither the parents' nor the child's legitimate interests in establishing intimate relationships with each other seem to be harmed if one replaces ‘right’ with ‘privilege’ regarding parents' ability to control the upbringing of their children without undue interference. Thus, while parent‐regarding arguments may support the notion of parents' privileges to control their children's upbringing, they do not lend support to parents' rights to the same.

Child‐regarding considerations also support the concept of parent privilege, although not of parental rights. As I mentioned above, two related arguments about children's psychological and autonomy‐based interests in stability and cultural coherence provide the most plausible justifications for parental control. First, the empirical argument is often made that parents are better able to advance their children's interests in general than other parties are. There exists a natural, even biological, bond of affection between children and their parents that cannot be broken without tremendous psychological cost to both sides, as the advocates of intimacy note. This affective bond not only offers children essential emotional support and sustenance, it is argued, but it also inspires parents (p.56) to further their children's interests more than other agents would, by giving them greater insight into their children's needs and greater forbearance and understanding when their children fail to live up to certain ideals. As Kleinig argues,

there is some reason to think that the developmental needs of children are likely to be met most successfully in an environment in which primary responsibility, and the authority that is derivative of that, lies with the parents. There are grounds for believing that parents, more than anyone else, will have the kind of commitment to their offspring that will safeguard and promote their welfare interests, and encourage the formation of an identity and life‐plans compatible with their individual character, abilities, and talents. Parents, therefore, are permitted a wide range of discretion so far as the treatment of their children is concerned.50

In addition, insofar as such a bond can be established only with real people with whom one lives in close contact (as opposed to with abstract institutions such as the state, or with governing committees such as the paternalistic overseeing board suggested above), there is good reason to think that a relatively small number of adults will qualify for such mutual affection.

Finally, the most overwhelming reason to grant parents the privilege of primary paternalistic control derives from the child's particular need for cultural coherence and a well‐developed and culturally embedded personality. As section 1.3 argued, membership in a community and embeddedness within a cultural and normative framework is a primary need of individuals—and an essential prerequisite for autonomy. One cannot act autonomously if one has no firm structure of beliefs on which to act. Although secondary associations such as churches, Scouts, and community groups often play an important part in teaching children values and giving them a sense of identity, families are probably most central to this process. Not only do families provide this cultural coherence and structure for their children, but they also require freedom from undue interference from outsiders in order to do it successfully. As Kleinig argues, ‘If children are to develop into beings with stable and cohesive interests, it is important that they be nurtured in a secure and integrated environment, where they are loved and cared for as intimates and not aliens. To interfere too much with parents' upbringing of their children is likely to destabilize the environment whose stability is of such importance to the children.’51 Thus, it would be counterproductive to interfere with or remove primary paternalistic control from the family on the grounds of advancing children's autonomy, because (except in cases of abuse, neglect, and the like) only the family whose integrity is respected is able (in conjunction with other secondary institutions) to provide children the cultural coherence that constitutes one of the essential preconditions for autonomy.

(p.57) It is important to note that in terms of establishing cultural coherence and nourishing the various elements of the child's personality, even autonomy‐fearing or ‐hating families can provide an adequate grounding for children's future autonomy. Families can provide a great deal of the groundwork for autonomy without intending to, which gives a further reason not to interfere in their internal dynamics. For example, parents might raise the child within their community because they believe that is the way to salvation, or because they think the community embodies the only true way of life. So long as children are able to acquire from elsewhere the further capacities and habits of thought required for autonomy, their development will actually have been advanced by being initiated into the norms and practices of a coherent community. Furthermore, the emotional stability and support offered by many families, in conjunction with basic but important lessons about trust, love, and mutual respect that children often learn in the family context, crucially help children's development of autonomy regardless of parents' intentions in this regard.

We see, therefore, that parents and the liberal state can occupy complementary roles in relation to children's development of autonomy. Liberal principles commit them both to acting in accordance with children's interests, but as very different institutions, they generally meet this obligation in different, complementary ways. The state is responsible for ensuring that children are given the means to develop their capacities for autonomy (whatever those means may be). Because of children's deep interests in having a stable home life, part of its execution of this duty lies in respecting parents' privilege to govern their children's upbringing, and thus in not violating the family's integrity unless it reasonably believes that parents are not adequately discharging their obligations (exercising their privilege) in the child's interest. Parents, on the other hand, are responsible for their children's physical and psychological well‐being, development of identity, and sense of cultural coherence—in addition to many other components of their children's well‐being. Because their control over their children is justified via children's interests in developing autonomy rather than via parental‐autonomy‐regarding considerations, parents are considered to have the privilege to control their children's upbringing rather than the right to do so. Their exercise of this privilege, however, offers them ample scope for independent paternalistic control, regardless of their stance on the moral value of individual autonomy.

2.4 The Role of Education

The Development of Autonomy The Role of Education

Thus far I have established that children have a fundamental interest in developing the capacity for autonomy, and that the liberal state has a role in (p.58) advancing this interest. I have also argued that state involvement is compatible with a strong affirmation of the integrity of the family in raising children and inculcating in them a vision of the good life. Thus far, therefore, my analysis of liberalism, families, and the development of autonomy seems fairly benign, demanding little if any radical social or political change. There is much more to developing autonomy, however, than possessing cultural coherence and an adequately developed personality (the two elements I focused on in section 3). In this section, I argue that the liberal ideal of autonomy not merely permits but requires the intrusion of the state into the child's life, specifically in the form of compulsory liberal schooling. I suggest that it is difficult for children to achieve autonomy solely within the bounds of their families and home communities—or even within the bounds of schools whose norms are constituted by those held by the child's home community. If we take the requirements of autonomy seriously, we see the need for a place separate from the environment in which children are raised, for a community that is defined not by the values and commitments of the child's home, whatever they happen to be, but by the norms of critical inquiry, reason, and sympathetic reflection. This community is the liberal school, and its achievement and provision to all children is the duty of the liberal state.52 In order to see where this vision of the ideal liberal education comes from, let us first remind ourselves of the characterization of autonomy developed in section 1.3.

The liberal ideal of personal autonomy, I concluded in Chapter 1, is a substantive (as opposed to formal or procedural) notion of higher‐order preference formation within a context of plural constitutive values and beliefs, openness to others' evaluations of oneself, and a broadly developed moral, spiritual or aesthetic, intellectual, and emotional personality. It is substantive in that it denies that individuals whose volitions involve completely subsuming their will to another can be autonomous. Independent of the formal aspects of such individuals' actions and decision procedures, the substance of their volition to give up their will to the absolute control of another is itself sufficient to establish them as heteronomous. I assume that most children who are raised in a non‐abusive or psychologically neglectful household will rise above this level of subservience and achieve basic substantive independence.

This conception of autonomy also demanded that individuals have a plurality of constitutive commitments. As we saw in section 1.3, plurality serves three purposes: (1) It enables the person to question any particular value without suffering a wholesale loss of identity, insofar as her identity is not dependent on the affirmation of a single value, desire, or belief. (2) Similarly, it allows the individual to criticize some subset of her beliefs from the standpoint of the others; she can revise her assessment of love, (p.59) say, or of her career, in the context of other normative beliefs and values, as opposed to searching for a mythical objective space from which to criticize everything. (This point relates also to the need for cultural coherence: i.e., the need for a pre‐established but malleable normative structure in which to evaluate options.) And (3) an individual with plural constitutive values and commitments will likely be more receptive to other viewpoints and to other people's criticisms of her own views. All of these characteristics are necessary to the process of deliberating about the values one holds, and thus eventually to basing one's life on higher‐order values and desires with which one identifies.

These three elements of pluralism themselves imply other useful conditions for autonomy, the most significant of which are: the existence of a thoughtful community which offers valuable evaluations and criticisms of one's own beliefs, a lack of false consciousness on one's own part and on others', and mutual respect and toleration. As social beings, we are always comparing ourselves to, contrasting ourselves with, adjusting ourselves to, and trying to live up to the ideals of other people; this is especially true of children. To achieve autonomy, therefore, it is useful to be surrounded by other people whose opinions, values, and activities are worth taking seriously. Further, this worth is in part premised on the assumption that what they have to say, and how they lead their lives, is in some way accurate and authentic. Just as true reflection and self‐definition are difficult in a social vacuum, neither are they easily achieved in a context of falsehood or ignorance; thus the achievement of autonomy by individuals is aided by the presence of an educated community free from false consciousness. Finally, values pluralism and a disposition for self‐reflection are meaningful only in the context of respect and toleration for differing viewpoints. An individual must realize at the least that reasonable people can hold views about the good that are different from his own—i.e., accept the burdens of judgement—for otherwise he cannot take criticisms of his own views seriously, and thus would have difficulty engaging in critical reflection properly understood.

These are demanding conditions. What implications do they have for the development of autonomy? Truthfully, if we take the conditions seriously, their implications are almost overwhelming. The following list sets forth just some of the personal qualities that seem necessary for individuals to achieve full autonomy. (It should be noted that the length and demanding nature of the list should be taken to indicate not only how much children have to learn in this regard, but also how much we as adults have yet to learn. In this sense, the list might best be described as setting forth a set of ideals to be striven for by all individuals who value autonomy, rather than as a practical guide to be realized in twelve short years of schooling.) I (p.60) suggest that in order to develop and achieve autonomy, children should: gain sufficient self‐esteem and confidence to feel comfortable articulating their views in public and laying themselves and their views open to challenge—but also possess enough humility to take challenges to their positions seriously; learn to express themselves in terms others will understand, and to listen to others' responses; be imaginative, possessing the ability to step into other people's shoes and to see perspectives other than their own; be creative, observant, and sensitive to subtlety; learn to think critically and to use reason effectively and judiciously; be willing to subject their own arguments and intuitions to the demand for proof; gain the skills and knowledge to put their beliefs and values into practice, including vocational and personal skills, manual and technical competency, and social skills, among others; be exposed to and interact with people from different backgrounds, and in different contexts; learn how to read and write, and to do basic mathematics; learn about their political rights and obligations, and learn as well the history of their locality, nation, and the world; and be initiated into the social meanings of the community in order to participate as members in the cultural conversation. (The list could clearly go on.)53

These skills, areas of knowledge, competencies, psychological states, habits, and structures of belief cannot be acquired from only one source. They are achieved by children over time and in a multitude of contexts, of which the most important is certainly the family. But it is hard for the family to help the child achieve autonomy completely on its own. This is obviously true in the case of ‘intolerant’ families whose conceptions of the good preclude autonomy in the way described above. The parents in the Mozert case discussed above specifically objected that the reading program had ‘the potential to make Christian children to be other than who they are; to create in them a desire to change from their heritage’ and to cause ‘the children to become more rebellious, to believe that they are their own authority’.54 Insofar as these objections boil down to the accusation that the reading program was advancing children's autonomy, it is clear that such parents and families cannot be counted on to help their children develop a capacity for autonomy themselves. The state must take an active role, in the form of compulsory education that is regulated, if not provided, by the state.

It is not only children of ‘intolerant’ parents, however, who need aid from the state to develop their autonomy. Because of lack of knowledge, resources, time, interest, or some combination of these, the vast majority of parents are unable to provide a learning environment in which all of the various competencies listed above can be acquired. Most people reading this, for example, would be overwhelmed if, while trying to lead their own lives, they had also to teach a child even the essential knowledge, skills, (p.61) habits, and beliefs from the list above. For these reasons alone, then, the liberal state is justified in making some sort of formal education compulsory.

Beyond the school's capacity to teach basic knowledge and skills, however, there is an additional and even more pressing reason to view the state‐regulated school as intrinsic to the development of autonomy in children. This has to do with the unique institutional role the school is able to play in fostering the development of autonomy—a role that cannot be replicated by even the most well‐intentioned and resource‐laden parents. I argued above and in section 1.3 that value pluralism and critical reflectiveness can best be achieved in an environment which itself fosters pluralism and reflection. This means that children are most likely to develop the capacity for autonomy in a community whose normative structure is itself autonomy‐driven—i.e. in an environment that is explicitly committed to and structured by the norms of critical inquiry and reflection, evidential justification, and mutual respect and toleration. Such a community would ideally bring together, in an atmosphere of toleration and respect, both children and adults who come from a variety of backgrounds and who hold a range of potentially conflicting beliefs and values. It would also privilege critical inquiry over indoctrination; it would value beliefs and commitments which are held evidentially and authentically, and disvalue those held as a product of false consciousness; and it would foster an atmosphere of reflection detached from the constitutive commitments of the other arenas of the child's life. Finally, such a community would provide a group of people whose evaluations and judgements of one's choices are worth taking seriously; and, more generally, establish an environment in which the various aspects and competencies of autonomy can be freely practiced and improved.

This community, I suggest, best describes the ideal liberal school. More than any other institution in the child's life, the school can—and should—establish the sort of artificial environment in which all of these norms together can be achieved. Why cannot the family, or other social institutions, achieve these norms? It is because of the very artificiality of such a community that the family is unsuitable for the role. First, even the most liberal or autonomy‐loving family cannot escape the conceptual and emotional bounds of its own commitments; nor can it ensure the child's honest exposure to that which the family would find utterly foreign or repugnant. Parents who believe that religion is the opiate of the masses, for example, might be unlikely to encourage their children to find friends who are strongly religious; likewise, parents who believe in the centrality of the church in guiding their and their own children's lives might not seek out for their children friends who are avowed non‐believers. Second, even though (p.62) any one of the above aims could be achieved at home—many families, for example, attempt to foster critical thinking skills and thoughtfulness in their children—such aims are also sometimes sacrificed (for good reasons) in contexts other than the school. While many liberal parents may in general favor their children's thinking for themselves, interest in family harmony or the need to make an important decision might (properly) trump the child's opportunity to practice autonomous decision‐making. Thus, a space separate from that constituted by the child's family and home community is needed in which to pursue these norms without compromise.

Further, I would suggest that no institution other than the school could achieve all of these norms, or create a community displaying all of these characteristics at once, because no other institution has the same singularity of purpose—to educate children for autonomy—as the school is able to have. Nor do other institutions reach people at the same critical age that schools do. Although intermediate associations such as the Scouts or the YMCA may also help children develop certain capacities for autonomy, their voluntary and often socially and culturally homogeneous nature makes them unfit for fulfilling all of the conditions listed above. As an institution uniquely devoted to constructing an environment in which children's interests in particular are furthered and developed, on the other hand, the school can (and should) take on the demands imposed by the liberal ideal. Detached from the inevitably partial values, beliefs, and commitments of children's families and home communities, the liberal school makes available an essential space in which children are enabled to start defining themselves on their own terms, encouraged—as well as repeatedly challenged—by an educational community in which norms of autonomy have a central place.

Children's membership in communities of this sort, however, has meaning only if they learn to use and to value the capacities and skills associated with autonomy in the first place. It might be admirable that a child's teachers try to get her to reconsider her commitments in light of evidence that is worth taking seriously, but if she does not value evidential arguments or simply has little experience in responding to such kinds of arguments, her autonomy will not thereby be advanced.55 The conception of the ideal liberal school as an autonomy‐driven community is also important, therefore, because membership and participation in such a community will help children to come to value and to practice the capacity for autonomy—both of which are essential to developing autonomy. By living and working (at least for the six to eight hours per day of school) in a community that self‐consciously values and shapes itself according to the attributes of tolerance and reflectiveness, the child herself will hopefully come to adopt and to value these dispositions as well; these are crucial first steps in attaining (p.63) autonomy. ‘Autonomy is an excellence of character—a virtue in the Aristotelian sense. . . . If the development and exercise of autonomy are connected in this manner, then a schooling system devoted to the end of cultivating autonomy must be concerned with ensuring, among other things, ample scope for its exercise.’56 The liberal school under this ideal, therefore, establishes a singular environment in which children's capacity for autonomy is constantly reinforced and developed through classroom discussions, written work, projects and simulations, other pedagogical exercises, curriculum design, and even the structure of the school. It provides a safe and supportive environment in which children can take the intellectual and psychological risks that are essential to becoming autonomous, and in which their interest in taking these risks is rarely outweighed by other, non‐autonomy‐advancing interests, as might occur in the other arenas of the child's life.

Properly understood, then, the ideal liberal school establishes a plural community whose structure and content are dictated by the overriding goal of fostering the development of children's autonomy—a community instantiated by the norms of critical inquiry, toleration, and reflectiveness. Such schools, we have seen, help foster the acquisition of essential knowledge and skills, help children develop attitudes of toleration and sympathy toward other people, and provide a reflective environment detached from children's home communities in which their capacities for critical thinking and self‐reflection can develop and flourish. It may be objected, however, that they are too detached. In attempting to foster children's development of autonomy by establishing separate, artificially constructed educational communities, the liberal educational ideal may have set up a dangerous—and pedagogically backwards—opposition between state‐regulated schools, on the one hand, and parents, families, and communities on the other. I examine four problems associated with this objection in the next chapter.


(1.) This might not be the case if the exercise of autonomy were merely an instrumental good needed to satisfy an end or need that was generated only by the possession of autonomy itself. For example, it could be argued that if one has the capacity for autonomy, one will be unhappy if the capacity is frustrated and unhappiness is intrinsically bad, so people with the capacity for autonomy should be able to exercise it. But the argument for autonomy was never made in this way—nor should it have been. It is autonomy itself that is viewed as a good, not its exercise as a means to satisfying the burden of its possession.

(2.) Amy Gutmann, Democratic Education (Princeton: Princeton University Press, 1987), 30. See also Eamonn Callan, ‘Tradition and Integrity in Moral Education’, American Journal of Education 101 (1992), 23.

(3.) This is not to ignore Aristotle's important insight about habits—that often the best way to develop a capacity (for courage, temperance, or virtue, say) is to be forced to exercise it, awkwardly and with much coaching at first, and then more confidently as time goes on. But even Aristotle reminds us that instruction, which is separate from the act of being courageous or temperate, is also demanded for the successful acquisition of (a) virtue (Aristotle, Nicomachean Ethics, x. 9). For a further discussion of habits, see sects. 2–4 and 4.2.

(4.) For further justification of this point see Diana T. Meyers, Self, Society, and Personal Choice (New York: Columbia University Press, 1989), 193: ‘Although the availability of a variety of options provides an incentive to master autonomy skills, children cannot gain autonomy competency simply by being flooded with options. A more direct attack on the problem is needed. The competency of autonomy must be handled in the same way that other desirable competencies are. It must be deliberately taught. Parents and teachers must encourage their charges to attend to their self‐referential responses and must help them to find acceptable ways to act on these insights.’

(5.) This is not to say that no elements are shared between adults exercising and children developing their autonomy. Some provisions, such as adequate shelter, food, health care, and security, in addition to less tangible goods such as self‐esteem, the capacity for reason, and imagination, are equally essential to the development as to the exercise of autonomy. There are many points of overlap. But it remains true that a much larger set of the conditions for the exercise of autonomy, as with shoe‐tying, either are irrelevant to or actually conflict with its development.

(6.) A. I. Herzen, quoted in Anthony O'Hear, Education, Society, and Human Nature (London: Routledge and Kegan Paul, 1981), 46.

(7.) This description is obviously unrealistic insofar as the choices a person makes at one stage of his adult life inevitably shape and limit his future possible choices, even if his opportunities are in theory the same as they were before. Furthermore, the choices that other adults make about their exercise of freedom and autonomy will, in many cases, alter the choices available in practice to oneself. Autonomy‐advancing cultures and communities do pose a threat to autonomy‐fearing individuals, and especially to the maintenance of certain autonomy‐fearing cultures and communities. Significantly, the threat can go the other way as well. Even if a state guarantees all the liberal freedoms discussed in Ch. 1, if a substantial segment of the population engages in self‐censorship, choosing not to avail themselves of the freedoms of speech, press, or association, then all citizens will suffer reduced autonomy as a result of the desiccated political culture. Thus, my claims about the costlessness of adults' choices regarding the exercise of autonomy would need modification in practice. The point remains the same, however: there is a fundamental difference (p.180) in kind between guaranteeing freedoms for adults (to exercise autonomy) and coercing children (to develop their autonomy).

(8.) Nomi Maya Stolzenberg, ‘ “He Drew a Circle that Shut Me Out”: Assimilation, Indoctrination, and the Paradox of a Liberal Education’, Harvard Law Review, 106/3 (1993), 597 n. 26.

(9.) Stolzenberg, ‘ “He Drew a Circle that Shut Me Out” ’, 633.

(10.) Howard Cohen, Equal Rights for Children (Totowa, NJ: Rowman and Littlefield, 1980), 43.

(11.) Cohen, Equal Rights for Children, 45 (italics his).

(12.) This is not to claim that paternalisma is always impermissible, or to deny that it poses controversial problems for the modern liberal state. There is a tremendous amount of debate about whether the state is justified in paternalisticallya violating citizens' autonomy in order ultimately to protect it, and about what one does in clashes between individuals' short‐term goals and their longer‐term preservation of autonomy. It is certainly arguable that paternalistic mandatory seat belt or crash helmet laws, for instance, violate some individuals' autonomous pursuit of their conceptions of the good; whether these measures are unjustified as a result is a much more controversial question. But in any case, the legitimacy of paternalistica legislation is irrelevant to the argument at hand. The issue is one of equal rights of children and adults against paternalism—not the legitimacy of paternalism altogether. If adults can rightfully be subject to paternalistic seat belt laws, then so can children; if seat belt laws are invalid for adults, the argument goes, then neither should they be imposed upon children. We can remain agnostic on the justifiability of specific examples of paternalisma, yet still acknowledge that adults are granted rights and protected from most paternalism (such as being forced to live according to the ruler's vision of the moral life, or to go to church in order to gain salvation) on the basis of their capacity to exercise autonomy.

(13.) John Locke, ‘An Essay Concerning the True Original, Extent and End of Civil Government’ (hereinafter referred to as Second Treatise), in Two Treatises of Government (London: Everyman's Library, 1990; 1st pub. 1690), sect. 61. For an analysis of the scope of parental versus state authority, see the second half of this section and sect. 3.

(14.) J. S. Mill expresses a variation of this fear in On Liberty: ‘If protection against themselves is confessedly due to children and persons under age, is not society equally bound to afford it to persons of mature years who are equally incapable of self‐government?’ (John Stuart Mill, On Liberty (Harmondsworth: Penguin Books, 1974; 1st pub. 1859), 147).

(15.) John Harris, ‘The Political Status of Children’, in Keith Graham, (ed.), Contemporary Political Philosophy (Cambridge: Cambridge University Press, 1982), 37.

(16.) For an interesting summary of and objection to this critique based on consequentialist grounds, see Laura M. Purdy, In Their Best Interest? The Case against Equal Rights for Children (Ithaca, NY: Cornell University Press, 1992), 21–54.

(17.) David Miller has pointed out to me that this might be overstating the case a bit, (p.181) since as I noted in sects. 1.2–1.3, state recognition of autarchy—which almost all adults possess—would guarantee the presence of a number of basic freedoms, including those of speech, movement, association, etc. If it is true that respect for autarchy adequately generates these freedoms, however, then liberals find themselves confronted by an important dilemma. Most children over the age of 5 or 6 years might be said to possess autarchy—yet children are routinely denied the sorts of freedoms I've suggested political regard for autarchy might generate. This might imply that children should be accorded many more rights and freedoms than they currently possess. The fear remains, however, that political recognition of children's autarchy (through granting political freedoms) would stunt the long‐term goal of helping children develop their autonomy. We are back at square one, in a certain way, again grappling with the question: does any moral justification exist for treating children and adults differently in the political sphere?

(18.) John Kleinig, Paternalism (Manchester: Manchester University Press, 1983), 156.

(19.) Again, this is not to say that the possession of autonomy is age‐sensitive, insofar as it is admitted that many adults currently do not possess autonomy. But given that we assume everyone possesses autonomy unless proven otherwise, and that 100% of young children (say, under 5 years old) have been proven otherwise, then the presumption of autonomy distribution by the liberal state is legitimately age‐dependent.

(20.) See Cohen, Equal Rights for Children, 52–5 for a convincing argument against establishing a state board of competence to examine the young for autonomy competency on a case‐by‐case basis. See also Francis Schrag, ‘From Childhood to Adulthood: Assigning Rights and Responsibilities’, in Kenneth Strike and Keiran Egan, (eds.), Ethics and Educational Policy (London: Routledge and Kegan Paul, 1978), 61–78, for an interesting argument against evaluating children's ‘maturity’ as a means for giving them equal citizenship rights.

(21.) As Cohen points out in an application of the sorites paradox, if the average child aged 16 has sufficient autonomy to decide to leave school, then so must the average child aged 15 years and 11 months—and one can always keep subtracting a month. See Cohen, Equal Rights for Children, 49–51.

(22.) Harris, ‘The Political Status of Children’, 37, paraphrasing Shakespeare's Richard III.

(23.) Shelley Burtt asks virtually the same question—and ultimately arrives at virtually the opposite answer—in her thoughtful and provocative article, ‘In Defense of Yoder: Parental Authority and the Public Schools’, in Ian Shapiro and Russell Hardin, (eds.), Nomos XXXVIII: Political Order (New York: New York University Press, 1996), 412–37.

(24.) Amy Gutmann, ‘Children, Paternalism, and Education’, Philosophy and Public Affairs, 9/4 (1980), 341. This is not to deny that there are many consequentialist moral reasons, as well as parent‐regarding reasons, for favoring parental control; these are addressed in the next section. It is rather that parents have no intrinsic pre‐legitimation claims to control over the child. Insofar as Locke is right to reject the notion that children are the property of their parents (see (p.182) Locke, Second Treatise, sects. 54–8), and insofar as children are recognized as separate individuals who are more than mere extensions of their parents, there exists no obvious a priori reason to free parents from the burden of legitimation (of their right to coercive control over their children). This may prove to be an easy burden to fulfill, on grounds that will be discussed in section 3, but it is not trivial or unnecessary.

(25.) As Gutmann agrees: ‘Nor is it a valid objection to say that the democratic state . . . is simply imposing its values on children. Some values must be imposed in any case. What is at issue here is not whose values but what values ought to be imposed upon children’ (Gutmann, ‘Children, Paternalism, and Education’, 351).

(26.) Gutmann again sums up this conclusion well: ‘We can justify limitations upon parents’ rights [to limit ‘a child's future ability to exercise meaningful choice’ (p. 350)] because our valuation of liberal freedom to pursue differing conceptions of the good is dependent upon that freedom being exercised by beings who have been raised under conditions conducive to choice' (Gutmann, ‘Children, Paternalism, and Education’, 351).

(27.) Joseph Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986), 410.

(28.) Raz, The Morality of Freedom, 411.

(29.) I must append one caveat to this rosy picture of the trouble‐free advancement of autonomy. There is a chance that the individual's newfound capacity for autonomy will lead her to reject the community in which she was raised, but will not give her the means to put a revised conception of the good in its place. Rather, she becomes directionless, overwhelmed by a seeming infinitude of choice and having no principles or settled values by which to sort through and choose one set of beliefs over another. Deprived both of a determinate set of values and of the linchpins of her former personality, such an individual ends up being stripped of two of the essential prerequisites of autonomous action. She falls into anomie instead of attaining autonomy. In such a case, the state's effort at developing her autonomy actually leaves her worse off than when she started. I address this problem in sects. 3.4 and 5.1.

(30.) See John Rawls, Political Liberalism (New York: Columbia University Press, 1993), 192‐4, on the desirability of ‘neutrality of aim’ (although not, notably, neutrality of effect or influence, nor procedural neutrality).

(31.) This has parallels, I suggest, with Alan Montefiore's example of the father who remains ‘neutral’ in a dispute between his two children when he knows that the older, stronger child will win if he doesn't intervene (quoted in Raz, The Morality of Freedom, 113‐14). It also has parallels with modern outrage over Bosnia—the belief that abdication of responsibility by Western countries under the guise of ‘neutrality’ during the early 1990s led to the otherwise potentially avoidable torture, death, and loss of territory of hundreds of thousands of Bosnian Muslims.

(32.) One author who feels no compunction about asserting radical state control over parents and their right to bear and rear children is Hugh LaFollette: see his ‘Licensing Parents’, Philosophy and Public Affairs, 9/2 (1980), 182–97. See (p.183) also Lawrence E. Frisch, ‘On Licentious Licensing: A Reply to Hugh LaFollette’, Philosophy and Public Affairs, 11/2 (1981), 173–80; and LaFollette's response in the same issue, ‘A Response to Frisch’, 181–3.

(33.) Gutmann, ‘Children, Paternalism, and Education’, 341.

(34.) My use of the terms ‘parents’ (in the plural) and ‘family’ in this section are not meant to indicate bias against single‐parent‐families or other non‐nuclear family child‐raising conditions. I adopt the term ‘parents’ solely for ease of exposition.

(35.) James G. Dwyer, ‘Parents’ Religion and Children's Welfare: Debunking the Doctrine of Parents' Rights', California Law Review, 82/6 (1994), 1375. Dwyer's article is an excellent and extremely tightly argued rejection of parents' rights; I borrow my distinction between parents' rights and privileges from him. Other theorists who take a similar approach include Yael Tamir, ‘Whose Education Is it Anyway?’, Journal of Philosophy of Education, 24/2 (1990), 161–70; John Wilson, ‘Indoctrination and Rationality’, in I. A. Snook, (ed.), Concepts of Indoctrination (London: Routledge and Kegan Paul, 1972), 17–24; Barbara Bennett Woodhouse, ‘ “Who Owns the Child?”: Meyer and Pierce and the Child as Property’, William and Mary Law Review, 33/4 (1992), 995‐1122; and Richard Arneson and Ian Shapiro, ‘Democratic Autonomy and Religious Freedom: A Critique of Wisconsin v. Yoder’, in Shapiro and Hardin, (eds.), Political Order, 365–411.

(36.) Judith N. Shklar, ‘The Liberalism of Fear’, in Nancy Rosenblum, (ed.), Liberalism and the Moral Life (Cambridge, Mass.: Harvard University Press, 1989), 21‐38.

(37.) Gutmann, ‘Children, Paternalism, and Education’, 344.

(38.) Locke, Second Treatise, sect. 58.

(39.) I would like to thank Daniel Markovits for bringing this argument to my attention. See also Eamonn Callan, Creating Citizens (Oxford: Oxford University Press, 1997), 142–3.

(40.) Charles Fried, Right and Wrong (Cambridge, Mass.: Harvard University Press, 1978), 152; quoted in Gutmann, Democratic Education, 29 n. 20. Gutmann also cites another proponent of this view, namely the Irish Constitution, Article 41, which ‘recognizes the Family as the natural primary and fundamental unit group of Society, and as a moral institution, possessing inalienable and imprescriptible rights, antecedent and superior to all positive law’ (Gutmann, Democratic Education, 29).

(41.) Stolzenberg cites five published opinions (Stolzenberg, ‘ “He Drew a Circle that Shut Me Out,” ’, 584): Mozert v. Hawkins County Public Schools, 579 F. Supp. 1051 (E. D. Tenn. 1984); Mozert v. Hawkins County Public Schools, 582 F. Supp. 201 (E. D. Tenn. 1984); Mozert v. Hawkins County Public Schools, 765 F. 2d 75 (6th Cir. 1985); Mozert v. Hawkins County Public Schools, 647 F. Supp. 1194 (E. D. Tenn. 1986); and Mozert v. Hawkins County Board of Education, 827 F. 2d 1058 (6th Cir. 1987), certiorari denied, 484 US 1066 (1988).

(42.) Various sources quoted in Stolzenberg, ‘ “He Drew a Circle that Shut Me Out,” ’, 595–6.

(43.) ‘Lost’ is a term of art in this case, as Harry Brighouse has pointed out, since in response to the court case: some parents removed their children from the state school and placed them in fundamentalist Christian schools instead; Hawkins County Public Schools dropped the offending curriculum; and the textbook publishers excised the disputed passages from future editions. See Harry Brighouse, ‘Egalitarian Liberals and School Choice’, Politics and Society, 24/4 (1996), 467.

(44.) Stolzenberg, ‘ “He Drew a Circle that Shut Me Out,” ’, 596.

(45.) Quoted in Stolzenberg, ‘ “He Drew a Circle that Shut Me Out,” ’, 631.

(46.) Kleinig, Paternalism, 145. One argument in support of parents' ownership of children is the bizarre but comparatively recent lawyer's brief to the US Supreme Court in 1925: ‘In this day and under our civilization, the child of man is his parent's child and not the state's. Gone would be the most potent reason for women to be chaste and men to be continent, if it were otherwise’ (brief by William D. Guthrie for appellee, Pierce v. Society of Sisters, 268 US 510 (1925) (no. 583), 66–7; cited in Woodhouse, ‘ “Who Owns the Child?” ’, 1102). Aristotle provides another unambiguous note of historical support: ‘There cannot be injustice in an unqualified sense towards that which is one's own; and a chattel, or a child until it is of a certain age and has attained independence, is as it were a part of oneself; and nobody chooses to injure himself (hence there can be no injustice toward oneself); and so neither can there be any conduct towards them that is politically just or unjust’ (Aristotle, Nicomachean Ethics, 1134b; quoted in Woodhouse, ‘ “Who Owns the Child?” ’, 1044).

(47.) Ferdinand Schoeman, ‘Rights of Children, Rights of Parents, and the Moral Basis of the Family’, Ethics, 91 (1980), 6.

(48.) Schoeman, ‘Rights of Children’, 8, 14.

(49.) Callan, Creating Citizens, 143; see also 142–5.

(50.) Kleinig, Paternalism, 145.

(51.) Kleinig, Paternalism, 147.

(52.) I take no stand at this point on how the provision by the state of liberal education is to be achieved—either directly through state schools or indirectly through state‐regulated private schools. I address the specifics of school provision, administration, funding, and so forth in Ch. 5.

(53.) To those who object that this conception of autonomy is overly demanding and unrealistic, I would respond that while it is demanding, so that probably no individual ever does satisfy all of the demands of full autonomy, that should not prevent us from holding autonomy up as a (distant, utopian) goal. There is no point in weakening the ideal of autonomy simply because a thinner version would be easier to achieve, any more than we should accommodate human cowardice in assessing the meaning of courage. Gutmann makes a similar point regarding education for democracy and citizenship: ‘Were the only goal of a democratic state to prepare its members for citizenship, its maxim would be, ‘Mandate the maximum education.’ Citizens would be forced to spend most of their lives preparing for citizenship rather than exercising it. . . . That is why (p.185) we encounter no paradox, only a serious problem, when we acknowledge that democratic states have the authority to make schooling compulsory for children but not for adults who fall below the democratic threshold of education. Since the threshold defines not a fully but an adequately educated citizen, this constraint on democratic authority may leave many adults less than adequately educated. A stricter constraint—mandating the maximum education—is ruled out by our recognition of the primacy of treating adults as sovereign citizens’ (Gutmann, Democratic Education, 278–9).

(54.) Stolzenberg, ‘ “He Drew a Circle That Shut Me Out” ’, 597.

(55.) This point is parallel to my argument for children's development of autonomy (and thus for education) in the first place; liberal freedoms have meaning and value only in a context in which people have the capacities actually to avail themselves of these freedoms—likewise with autonomy.

(56.) Eamonn Callan, Autonomy and Schooling (Montreal: McGill‐Queen's University Press, 1988), 146.