One of the most troubling aspects of education is the prevalence of blame, scolding, and punishment in public schools. Why do we subject children to that ordeal?

We know the kinds of answers usually given to questions about why a child is scolded or punished: "He deserved it"; "She knew the rules"; "How else do I get through the lesson?" But these are elliptical answers. Where the student has not harmed other students or obstructed their efforts to learn, answers of this sort avoid the real issues. Why did the child deserve scolding or punishment? Why are there these rules and policies? Why should children be subjected to this sort of regimen?

In this essay I try to reconstruct the account that lies behind these elliptical answers, and ask whether it works as intended, and, if so, how much it justifies. I present what I call the responsibility account: a very traditional justificatory argument based on students having a duty to learn and teachers holding them responsible. One main challenge is to figure out the source of this duty. The second is to examine how the account handles excusing conditions.

I argue that the account I have proposed can handle both challenges. Provided certain conditions are met and premises accepted, the regimen of blame and scolding and punishment - what I call the ordeal - is justified. I conclude by suggesting that there is more at stake in this inquiry than just the defense of teachers' conduct. The legitimacy of public education is also at issue. If students do not have the kinds of duties teachers appeal to in their defense of punishment, it is difficult to see how truancy laws and school tax assessments can be justified.

How does the responsibility account work, and what does it justify?

The teacher must link classroom rules and procedures with some broader pre-institutional duty. Before there were classrooms, children had no moral obligation to sit in particular seats or have a pencil within reach. Presumably the teacher's decree cannot generate a duty ex nihilo. If not, where does it come from?

To answer that, we need to know why children are in school. To keep it simple, suppose they are there to learn long division. The teacher will then have to presume that it is a moral obligation for children of a certain age to be able to divide by a two-digit number. If it were not, her justificatory account would never get off the ground.

Even if we concede this duty, the teacher still has more work to do. There is no necessary connection between her seating plan and the algorithm for dividing 300 by 40. Sitting in a certain seat for a certain length of time is not a prerequisite for learning the algorithm. To avoid the obvious objection, she will say that seat-sitting, hand-raising, pencil-holding and the rest are necessary elements of a certain routine that makes it possible for students to learn the algorithm. Compliance is a necessary part of a routine and going through the routine is a sufficient condition for learning long division.

Notice that this justifies the demand for compliance only if there is no other way in which students can learn long division. But that is not of concern here. Few children who daydream during math or neglect homework are proposing a new way to learn long division. Of course, they could learn through a different routine or from a different teacher. But this is not one of their options: they have this routine and this teacher, or none at all. Their choice is to comply or to learn on their own. If they can learn on their own, they should be excused from the lesson. If not, then provided that the two conditions are satisfied and children do have the duty, it looks as if a certain significant part of "the ordeal" is justified.

How much of a difficulty do these three requirements pose?

The sufficiency condition can be met simply by showing that students who go through the routine really do learn how to divide. We can find that out by testing them. Here is an important policy consequence of the responsibility account. We need something like a standards-and-testing program in order to justify the whole complex of legal compulsion that supports public schools. We cannot demand compliance if we cannot show that students are learning what they should learn.

The necessity condition is also easily defended. Teachers get asked all the time whether the rules are really necessary to the routine. Why would the math lesson not work just as well if children were allowed to talk quietly while they did seatwork? Can they at least get up and sharpen their pencils without asking permission? A teacher usually has fairly wide latitude about how she answers these questions. As author of the routine, she is better placed than others to know whether each element is needed or not. All we can say here is that she should have a plausible answer and be open to counter-arguments.

The premise about duty, however, is more interesting. Why is it a duty for a 10-year-old to be able to divide 300 by 40? Answers to this question are explored in the next two sections.

Why do students have a duty to learn what public schools teach? I shall consider three possible answers: first, that they have a duty to themselves to get a good education; second, that the duty to learn is founded on a more general social obligation, for example to contribute to society or to be a good citizen; third, that it is derived from their parents' duty to ensure that they get a good start in life. All of these answers are superficially plausible; all may be genuine duties. But as we shall see, they do not fit teachers' practices equally well. Only one is relevant to those features of the ordeal that most urgently require justification.

Our first candidate is self-regarding duty: for example, the duty of prudence. It is, of course, controversial whether there are duties of prudence, but suppose there are. Students may have a duty to do what needs to be done now in order to live well in the future. Could such a duty justify scolding and punishment?

It seems plausible that children do have prudential duties, if anyone does. Eight-year-old Jason ought to brush his teeth, and sixteen-year-old Erica ought not to go to wild parties and drink beer to excess. These duties are owed to the self, not to others. What makes the conduct obligatory is its effect on the agent's future well-being. If Jason says, "I brush my teeth because my parents make me," he is not acting from a duty of prudence. True, he exhibits prudence if he avoids punishment, but a prudential duty to avoid punishment cannot justify punishment. If it could, then "she knew the rules" would justify the ordeal.

Can any other prudential duty justify punishment? To hold someone responsible for default of a prudential duty is ordinarily to say that the outcome affords her no cause for complaint. She foresaw or should have foreseen the consequences of her conduct. In acting as she did, she can even be said to have chosen the consequences. A person who bears and accepts these consequences acknowledges responsibility for his breach of duty. "He never paid for his mistake" indicates that the agent has managed to avoid responsibility. Someone else may have shielded him from the results of his conduct, for example by making good his losses. "I've paid the price," by contrast, is a defense against further blame from other people. If I have borne all the harm caused by my action, what reason is there for anyone else to blame me?

We are reluctant to hold children responsible in this way. Often they cannot foresee or appreciate the long-term effects of their actions. Eight-year-old Jason does not choose a lifetime of dental work in exchange for the few minutes extra play he gains each night by not brushing his teeth. He does not choose not to be an accountant or engineer in order to avoid doing arithmetic.

If it is not fair to say that Jason has chosen these consequences, he cannot be held responsible for breach of a prudential duty in the way people ordinarily are. Indeed, if he lacks the capacity to recognize the prudential duty involved, how can he be blamed at all? Punishment under these circumstances would seem premature. Jason's parents are more likely to deal with the problem with a direct command. In effect, they resort to coercion. Once the command has been issued, this will no longer be a case of holding Jason responsible for prudential duties. His choice whether or not to act as prudence dictates has been overridden.

How would this reasoning apply to school? Children do have a prudential duty to learn, and therefore also to comply with the teaching routine. If they comply, perhaps they accept the duty. The teacher could hold them responsible by reminding them of the consequences of not complying. For example, if Erica falters in trigonometry class, her teacher might ask her if she still hopes for an engineering career, and remind her of the importance of a strong math background to success in that field.

The difficulty is that this is not the kind of scenario that makes the ordeal an ordeal. What we need is to justify coercion, paternalistic control, blame, scolding, and punishment - all of which are less evident in trigonometry class than in a fourth grade learning long division. Younger children are subject to more stringent paternalistic controls precisely because they are not as likely to be held responsible for duties of prudence.

This explanation of paternalistic intervention, if accepted, closes off one line of justification, but also suggests two other possible approaches. First, some might argue, following Peter Strawson, that punishment has a dual purpose when applied to children.¹ Not only do we thereby hold them responsible for their conduct, but we also teach them the nature of the responsibility and how it is communicated. But the demand for compliance in school can hardly be said to teach children prudence; indeed, it is one of the most common complaints about school discipline that it aims at compliance rather than moral development. Being scolded or punished is unlikely to help children refine their moral judgment and develop capacities needed for self-regulation. Students are not, in John Covaleskie's phrase, "being helped to become moral."²

The alleged pedagogical failure of disciplinary practices - the indifference to children's moral development that they are supposed to reflect - opens up another possible line of justification. The ordeal can be viewed as conditioning. Children, in effect, are being offered inducements to do what will ultimately benefit them, even though they cannot understand those benefits now. On this view, teachers' coercive practices could be justified by something like the legal theory of deterrence as justification for punishment.

No doubt teachers do hope for a deterrent effect, but this cannot bear the weight of justification. In punishing or scolding a child, the teacher is not merely offering inducements to avoid future infractions; he is also saying something about the child's present conduct. He blames the child, holds her morally responsible for compliance with rules and routines, asserts that she has done what she ought not to have done. How else could he claim that the child deserved the punishment, as teachers typically do claim in their defense of these practices?

It seems very unlikely, then, that duties of prudence justify the ordeal. We must look elsewhere for the source of a student's duty to learn.

One possibility is that the duty to learn might be based on a more general social obligation, such as citizenship, self-sufficiency, or social contribution.

Consider citizenship first. A number of political theorists, including Benjamin Barber, Robert Dahl, and Amy Gutmann and Thompson, urge that we take a robust view of citizenship. To be a good citizen, they argue, one should not only vote and obey laws, but also participate actively in public life at various levels, deliberate with other citizens, and otherwise contribute to community political life.

Activity of this sort undoubtedly does require a good education. But robust citizenship is not the norm in the United States. Advocates of this view complain that our political system induces passivity. It seems highly implausible that students can fairly be held to a duty not recognized by most adults, and even more implausible that if they have such a duty, it is the source of teachers' blame for daydreaming or talking during a lesson.

The duty of self-sufficiency looks even less promising. True, there may indeed be a duty not to throw oneself on the charity of fellow-citizens. Education helps a person avoid doing so. But surely the generosity of fellow citizens cannot play a central role in an account of the duty to obtain an education. If their generosity ceased, would the moral importance of education decrease? Would students be blamed less for misreading instructions or not finishing homework?

A duty of social contribution looks more promising. Society is often viewed as cooperative enterprise. Perhaps it is a duty to do one's part in this scheme. If so, there may be a duty to develop cooperative skills, to which education contributes.

Two kinds of doubts can be raised about this proposal. First, if cooperation is conceived in terms of mutual advantage, then any social duty would have to be a duty of prudence.³ Thus it would be vulnerable to objections already mentioned.

Suppose, though, we construe the social obligation as a duty of reciprocity: for the benefits we receive, we ought to give something back. But now the social-responsibility account has an unfortunate consequence. It implies that the strength of a student's duty to obtain an education is proportional to the benefits she has received or expects to receive. Two very unappetizing effects follow from this proportionality. First, students from poor families would have less responsibility, and thus teachers' expectations of them would legitimately be lower. Second, any sustained failure to fulfill one's duty to learn would be self-justifying, because it would reduce one's future expectations and thus diminish the duty of reciprocity.

Neither self-regarding nor social duties seem to do the work we require. What remain are duties to particular people other than oneself. We may wonder how a duty to learn could be owed to some other person, when the person receives no direct benefit. We can see how the child might have a duty to a parent to mow the lawn or to a teacher to behave respectfully, but a duty to learn long division seems very different.

To avoid this difficulty, we introduce the notion of derived responsibility based on a duty transferred from one person to another through an agreement or relationship. Derived responsibility is familiar in everyday life: we can take on some or all of another person's duties by marrying, making a promise, becoming a friend, or accepting employment. Typically, the other person's duty becomes our own if she cannot fulfill her obligation unless we do our part.

Now we need to show that a student's duty to learn is derived in this way. From whom? The teacher is one obvious candidate, but then the child's duty could be avoided by withdrawal from the teacher's class. We need to look further back - to the child's parents. It seems plausible that parents have a duty to ensure, as far as they can, that their child has a decent start in life. In a society like ours, a decent start includes a good education.

Suppose, then, that there is a "duty to see to the education of one's children," as Joseph Raz has suggested.⁴ Is a part of this duty passed on to children? Parents cannot see to it that the child learns if the child is unwilling to cooperate. Thus, if there were no derived duty, parents could not insist on cooperation, and some would be left with a duty they could not fulfill. It seems odd to think of conscientious parents, sitting at dinner brooding over their unfulfilled duty, while the child, having daydreamed several hours in school, blithely eats the meal they have prepared without acknowledging a share of their burden. Would parents accept this?

Some might resist this account because the child would incur responsibility involuntarily. To avoid this, they might add a motivational component to the parents' duty. The duty to see to a child's education would then include a duty to make the child want the education. But then the same difficulty would arise: a parent could not fulfill that duty unless the child responded appropriately to reasons or inducements, and then either the child would incur some of the responsibility, or we would be confronted again with the prospect of unfulfillable duties.

There seems to be no way out unless the child shares the duty. Many of our common intuitions support this. The parent has a duty to see that the child goes to school. But suppose he discharges this duty solely by means of threats or bribes. We recognize that something has gone wrong here: the parent has failed to recognize that the child shares the responsibility.

What if someone still resists, preferring the prospect of an unfulfillable duty to derived responsibilities not taken up willingly? A child's duty to learn can be derived from parental responsibility in a different way, without transferring it from one person to another.

We start with parental duty. As equal opportunity theorists are wont to observe, discharging that duty is difficult for a parent who has not herself had a good education. It will be harder for her to assess her child's progress. She will be less able to help with homework. She may have difficulty communicating with teachers. Now we need only one further premise: children are potential parents. If they have children of their own, they take on the duty to give them a good start in life. Getting a good education themselves will greatly increase the likelihood of their being able to do this.

The argument from parental responsibility avoids the difficulties of the other two suggestions about the source of a child's duty to learn. It does not depend on controversial premises; it does not diverge widely from teachers' practice and customary defenses; it does not demand less of recalcitrant students or students from poor families. That is not to say, of course, that some teachers do not appeal to social and self-regarding duties in urging or demanding their students' compliance. The claim in this section, which is more limited, is that the argument from parental responsibility makes the best fit with those practices that most urgently require justification, and thus renders the responsibility account more plausible than the two alternative sources of duty that we have considered.

Suppose we accept the responsibility account up to this point. The critic may still urge that this conclusion be rejected because the result is too harsh. It places too much of the onus of educational failure on students, and too little on teachers, the school bureaucracy, or the political system. This is a common objection to recent proposals for high-stakes testing: why should children bear the responsibility of institutional failure?

The harshness objection, as I shall refer to this point, may seem to undercut the account presented here. As a matter of fact it poses no such difficulty, and may even lend additional support to the account.

To see why, consider the legal doctrine of excuses, which tells us someone cannot be held responsible for an offense if she could not help what she did. If she cannot be held responsible, she cannot fairly be punished or blamed. It is just these kinds of circumstances to which the harshness objection appeals. Suppose the child misbehaves because of pressure at home, or because he is angry or afraid or hungry. Suppose he refuses to cooperate because teachers in the past have treated him as if he were incompetent, and he assumes he will be treated the same way now. Should he be held responsible in the same way as everyone else? Of course not. The responsibility account will acknowledge that he has an excuse. It therefore avoids the harshness objection, and may even be strengthened by it. As one legal scholar has commented, excuses are "the royal road to a theory of responsibility generally."⁵ Getting clear about what counts as an excusing condition inevitably involves getting clear about the nature and limits of the responsibility from which the offender is to be excused. The harshness objection, in effect, delineates the scope of students' responsibility more sharply.

Hospitability to excuses, then, is one of the strengths of the responsibility account. But recent work in legal philosophy suggests that this feature may also introduce a new difficulty. The list of excuses traditionally accepted in law includes duress, insanity, necessity, and mistake of law. Recent work in legal theory has attempted to expand this list, and this effort is thought by some to undermine responsibility.⁶

Consider the problem of adding alcoholism to the list. Being an alcoholic is not something a person can help; as an addict, he may not be able to avoid drinking; when he is drunk, his self-control is impaired. From the doctrine of excuses, it seems to follow that there is no crime for which the alcoholic can be held responsible, if he has a few drinks before he commits it. Indeed, any element at all that contributes causally to a criminal act can generate an excuse. If so, then how could anyone be held responsible or punished for anything?

The problem for schools would be that many students, especially disadvantaged students, could appeal to an excusing condition. If the appeal is accepted, they cannot be held responsible for the duty to learn. Not being held responsible, they could not be compelled to participate, and thus would be less likely to learn. Can this pressure be resisted? Can a credible line be drawn between circumstances that do and others that do not generate excuses?

This is a vexed question in legal philosophy.⁷ For education, though, the difficulty can be circumvented if we recognize that excuses here are of limited duration and scope; expanding the list is therefore less likely to undermine responsibility than in the case of criminal law.

One reason for this limited character is the way in which responsibility is divided among parents, students, and teachers. If the student is shielded from her share of responsibility, parents and teachers cannot fulfill theirs. A teacher could not say, "Because of her specific learning disability, Jones is excused from learning long division this year; my method of teaching relies on skills of visual discrimination which she lacks." The teacher would be expected to change her method in whatever way was needed to eliminate the excuse. Any other excuse would generate the same expectation. Removal of the excuse would be a matter of the highest moral urgency for everyone concerned.⁸

Excuses in education, then, are self-limiting in a way that they are not in criminal law. No sooner has an excuse been granted than it generates a moral requirement for its removal. This is an important result: it not only enables the responsibility account to avoid the objections discussed in this section, but actually strengthens it. Not only is it fair to hold students accountable, but doing so can serve a corrective function: unfair blame generates excuses, excuses shift responsibility from students to adults, adults are put under pressure to eliminate the excuses and thereby correct the unfairness.

I have argued that blame, scolding, and punishment in public schools - what I have called "the ordeal" - can be successfully defended. Students have a duty to learn, and can be held responsible for violating whatever rules, policies, or instructions are enforced to ensure that they do so.

To justify the ordeal, however, is not to justify every scolding and punishment. The responsibility account mentions several important constraints. Legal scholars have recognized other conditions: that the overall system of rules is just, that the punishment is proportional to the offense, and that rules are general in application, adequately publicized, and enforced consistently and impartially.⁹

There is more at stake in the success of the responsibility account than just defending teachers. The account also bears on the institutional legitimacy of public schools and on two forms of compulsion on which they depend - truancy laws and tax levies. It is difficult to see how children can legitimately be forced to attend school if they do not have a duty to cooperate in the routine a competent teacher prescribes. And it is difficult to see how citizens generally could have a duty to pay for a child's schooling, unless the child also has a duty to help make this arrangement succeed.

The responsibility account fulfills this broader justificatory requirement. If, as I have claimed, it is the account to which teachers appeal in their ordinarily elliptical defense of rules, routines, and punishment, then they are justifying more than their own conduct; they are responding to moral imperatives deeply embedded in the institution of public schools.

For a response to this essay, see Markie.

9. For the fairness of the overall system of laws, see H.L.A. Hart, "Legal Responsibility and Excuses," in Justification and Excuse in the Criminal Law, 38. For the other conditions, see Lon Fuller, The Morality of Law (New Haven, Conn.: Yale University Press, 1964), 39.