A. Autonomy as access to options

To begin, consider the idea that the autonomous thinker has, while the heteronomous lacks, access to various options from which she can choose her own convictions. This conception of autonomy emphasizes the relationship between a student and her views. The heteronomous person is somehow limited to a single or very narrow range of ways of viewing the world. The autonomous person, by contrast, has sufficient options among which she can make the choice that best suits her. The old-fashioned view is objectionable insofar as it denies students this access, and “locks them in” to a single view. This provides one way of understanding what Phillips might find worrisome about education that amounts to mere “transmission of knowledge.” The problem, we might think, is that what is transmitted is a single, fixed view. This could also be the content of the worry that what the old-fashioned view is really engaged in is not education at all, but “indoctrination.”

Here is one sense in which a practitioner of the old-fashioned model might seek to deny students access to options: she might suppress or deceive them about the existence of disagreement over the matters that she is teaching, when such disagreement exists. We might think that autonomy requires access to options in at least the sense of awareness that the view being taught is not the only possible view, nor even the only actual one. Relatedly, suppressing awareness of the fact that there is political space for disagreement, that dissent from the view on offer is not subject to political sanction, could be a way of objectionably denying access to alternatives.

These practices may indeed be objectionable, but they do not seem to capture the sense in which progressives think that the old-fashioned view denies students options. The old-fashioned view does not require suppression of options in either of these senses. An educator who seeks to bring you to love what she judges to be good need not hide from you that not everyone loves what is good; nor need she hide the nature of a pluralistic polity in which a wide range of views enjoy political protection. So, while some of the most extreme, insular approaches to education may engage in these tactics, they are not a feature of the old-fashioned view as such.

Moreover, avoiding these practices would not satisfy detractors from the old-fashioned view. When advocates of the progressive view call for access to options, they usually have a stronger condition in mind. It is common to speak of the further access condition in psychological-sounding terms: a person has access to options only if they are live possibilities for her. ^{Footnote 5} The autonomous person is able to choose among various views or commitments. The heteronomous person, on the other hand, takes her views to be inevitable, natural, or unshakeable. ^{Footnote 6} She can gain no critical distance from them, and so has no choice — no live option — but to affirm them. Let’s grant that, in seeking to bring students to value what is good, the old-fashioned approach does aim to put alternatives out of reach in this way. It may then be on these grounds that critics believe that this approach violates autonomy and should therefore be rejected.

But we cannot understand the purported problem with the old-fashioned view as the inculcation of unshakeable, psychologically inescapable, beliefs. That’s because many of our beliefs and commitments seem to be psychologically inescapable, and unproblematically so. Commitments to well-established scientific and historical facts, and to a wide range of mathematical truths are like this. Most of us cannot give up these convictions at will. We are locked in to them and have no access to alternatives. Moreover, those who taught us likely aimed to get us to regard these commitments in just this way. But neither this aim nor the relationship between students and their beliefs that results should seem objectionable. In aiming to get me to believe, in a way that I cannot seriously question, that George Washington was the first president of the United States, my teachers did not violate my autonomy or place themselves in any objectionable relationship with me.

So if anything like this characterization of the objection to the old-fashioned view is going to work, we need a way of further specifying when seeking to lock students in to a view is objectionable. The best candidate for this further specification may seem obvious. The examples above concern empirical matters, while worries about the old-fashioned view usually arise from its approach to normative or ethical claims and ideals. So, we might try this: inculcating certainty about normative matters is an objectionable violation of students’ autonomy. With respect to questions having to do with what is of value, what deserves our allegiance, and how it is best to act and to live, students must be given options among which they can make up their own minds.

But lacking access to alternatives to a range of normative views, in the psychological sense above, also seems to be unproblematic. We learn early on that harming others requires some very significant justification: we cannot do so just because we feel like it, or because we cannot be bothered to pay attention. As with uncontroversial historical facts, many of us cannot seriously doubt that this is true, or take up some alternative view at will. Again, this is often precisely what those who taught us to regard harming others in this way were aiming to accomplish. And again, this does not seem to amount to an educational failure, or an objectionable violation of autonomy. If this is right, then aiming to inculcate certainty, and in this sense deny students access to options, with respect to normative matters is also not problematic per se.

It thus seems that people who worry that a student is being denied access to alternatives cannot really be objecting to the inculcation of certainty as such. Of course, where historical, scientific, or mathematical questions are genuinely open, it would be objectionable to lead students to firmly embrace a given view. Similarly, many normative questions are much more difficult than: May I harm others just because I feel like it? Insofar as the answers to such difficult questions are treated as easy or obvious, we might object. But the objection here is not to closing off access to options as such, but closing off access to particular possible answers to particular questions. It is not that a student is being taught to regard “her beliefs” as unshakeable or natural or inevitable. It is that she is being taught to regard some particular beliefs or convictions in this way.

This changes the nature of the objection, and what would be required to evaluate it. Determining whether a practitioner of the old-fashioned view is doing something problematic now requires considering the content of the convictions she seeks to instill, and defending a judgment that it is inappropriate to deny students live alternatives to that content. One might make such an argument in one of two ways. First, if you think that there are several live options with respect to a question, such that no one should have a settled view about it, then you will probably think that students should be taught in a way that leaves all of these options open. Second, you might think that the matter in question is one of genuine value pluralism. That is, you may think that a variety of incompatible options would be equally or incommensurably valuable, and that determination among these would rightly be made according to individual preferences, talents, or other idiosyncrasies. For example, within broad limits, choice of career or life’s work is almost certainly like this. Insofar as this is the case, it is clear why it would be objectionable to teach a student to regard just one, or a very narrow range, of career possibilities as live options.

Consider how this applies to the paradigm case of religious education — that is, education that is not merely about religion, but aims to instill a certain religious point of view. Those who worry about the appropriateness of religious education normally cast their objection at a high level of generality, claiming that the problem is that students are taught a single outlook. ^{Footnote 7} I am suggesting that something more particular must be at stake. The question is whether it is objectionable to seek to instill a religious outlook, or perhaps even whether it is objectionable to instill the particular religious outlook in question. One might suppose that it is objectionable on either of the two grounds above. Objectors to religious education often believe that the claims being communicated are false, or at least take the weaker view that we cannot know whether such claims are true, so that the question is genuinely open, and agnosticism is the appropriate attitude. Others think that religious outlooks are a matter of value pluralism like careers: There are several incompatible, but equally valuable commitments that one might have, and the selection is rightly determined by individual idiosyncrasies. My point is that these are not merely formal commitments about access to options. They are instead substantive normative views about how an educated person should regard religious commitments. Conspicuously, most supporters of religious education will reject them. They will not endorse agnosticism about religious matters, or think that it is right to regard religion as a matter of taste. By their lights then, instilling the commitments at issue does not deny students’ access to options they ought to have, and thus would not violate their autonomy on the conception under consideration.

The lesson here is that, if we understand autonomy as access to options, we cannot characterize the difference between old-fashioned and progressive approaches as disagreement over its value or appropriateness as an educational ideal. While the old-fashioned view transparently aims to pass on certain unshakeable normative convictions, only a very extreme and unattractive position would hold that this is never appropriate. To avoid saddling advocates of the progressive model with this extreme view, we need to interpret them as objecting that some particular instances of the old-fashioned approach teach students to have settled views about things that it is inappropriate to be certain about, or to regard their options as limited with respect to matters in which they ought to have multiple possibilities and be free to express their individual tastes or idiosyncrasies.

We cannot then teach for “autonomy” construed as access to options, without relying on some way of settling the question: With respect to which matters should students be given live options? The answer is plausibly: just those matters that are either open questions or areas of value pluralism rightly settled by individual idiosyncrasies. Traditionalists and progressives can agree up to this point; their disagreement is over which matters these are. This is a substantive, not a merely formal, disagreement. Of course it is not appropriate to teach students to view something as a settled matter when it is not appropriate to regard the matter as settled. But which matters are appropriately so regarded calls for substantive normative judgment about what, and not merely how, the educated person ought to think.

B. Autonomy as an ability to provide justification

Consider now a second attempt to specify the difference between the old-fashioned and the progressive views, this time focusing on the relationship between student and teacher. On this view, what characterizes the autonomous person is that he does not believe things just because someone told him to, but is able to articulate the reasons for or justification of his convictions himself. To teach for autonomy is to encourage this kind of independence.

As in the case of the first conception, this sort of independence from others is not a perfectly general ideal. Another’s testimony is often perfectly adequate reason to believe something. We routinely depend on others’ testimony about everyday matters, as when people tell us their names or give us directions. It is widely recognized that getting by without such dependence would be extraordinarily difficult, and it strikes me as pathological rather than commendable to try. Arguably, general refusal to trust what others say, in addition to being imprudent, manifests an attitude toward them that deserves to be called disrespect, and violates a compelling moral ideal. And reliance on testimony occupies a routine and almost certainly ineliminable role in education. Many historical and scientific claims (though perhaps not mathematical ones) can again serve as examples here. Most of what most of us believe about these matters is based on the testimony of our teachers or on the books that they selected. The view that the intellectual ideal at which education aims amounts to freedom from this kind of dependence strikes me as implausible.

Again, some will want to suggest that, while reliance on others for empirical information is fine, matters are quite different once we enter the realm of the normative. It might be thought that we ought not accept normative ideals or convictions on others’ say-so, that this would amount to criticizable heteronomy. ^{Footnote 8} The old-fashioned view seeks to direct students’ loves. So it might be thought that the objection to it could be captured in the concern that it encourages students to form normative commitments based just on the say-so of others.

This way of putting the objection suggests that the old-fashioned approach amounts merely to reporting normative claims, and expecting students to memorize and so internalize them. Talk of “indoctrination” may also conjure up this picture. But indoctrination in this sense would be, at best, a very poor enactment of the old-fashioned view, ill suited to achieve its stated aims. Consider an example. Suppose you wanted to shape a student’s normative attitudes about music. Contrast two approaches to this undertaking: on the first, you give your students a list of great composers and their finest works. You require students to memorize the list and test their mastery of it. This might result in students forming normative beliefs like “Mozart was a great composer.” But surely, taken in isolation, this would fall far short of achieving your goal. You cannot shape someone’s loves merely by reporting to him what is worth loving. If you aim to get people to value the music of Mozart, you need to expose them to the music itself. Likely you will also want to provide them with some further guidance, directing their attention to the features and qualities of the music that make it great. If you succeed, your students will come to appreciate the music directly, experiencing its greatness, rather than merely believing the proposition that it is great. This is a better example of the old-fashioned approach to education, and it is clearly not well described as the students coming to love the music just because you told them to. ^{Footnote 9}

But one might think that this still fails to develop the autonomy that fans of the progressive view seek. In fact, someone might worry that the approach I have just commended to the practitioner of the old-fashioned view — exposing students to things purportedly of value and directing their attention in ways that aim to get them to appreciate this value — might just as well describe an effective plan of indoctrination as one of responsible education. It’s all well and good when the object held up for appreciation is a Mozart sonata, but perhaps the example is misleadingly benign. What if the object were, instead, the glories of the dictator’s reign?

Of course, that kind of case would be objectionable, and a stellar candidate for the pejorative “indoctrination.” But the lesson we should draw is this: as above the question of what is appropriate depends not on the abstract or formal approach to education, but on the content of the things taught. When we worry about students being taught to value the glories of the dictator’s reign it is because we think that the reign ought, rather, to be resisted. On the other hand, when we think that being taught to value Mozart’s music is unobjectionable, it’s because we think that this music is indeed wonderful. These are nothing other than substantive judgments about what deserves to be valued and appreciated, or what makes for a good life.

But perhaps this conclusion is too quick. So far in this section I have focused on only one aspect of the conception of autonomy under consideration, the suggestion that an education that demands that students accept normative commitments just on the say-so of their teachers would violate autonomy. But the conception under consideration also includes the idea that the autonomous student can give the justification for her normative commitments. The appreciation envisioned above may not be sufficient for an ability to state the reasons for valuing the thing in question. So, perhaps this ideal can capture the disagreement between the two approaches.

But, once again, I think that we should be skeptical of the generality of this ideal. It is far from clear that inarticulacy about one’s normative commitments is always a problem. Consider again the person who accepts, indeed is absolutely convinced, that it is not acceptable to harm another person merely for convenience. Faced with a skeptical challenge to this commitment, she might well find herself inarticulate. I think most people would. But continuing to affirm her normative commitment in spite of this does not seem to mark any sort of fault in her.

Under pressure from this kind of example, we might try backing off to a weaker position. Perhaps autonomy does not require the ability to give a complete defense of all of our normative commitments, but something more like acknowledging a responsibility to be responsive to skeptical challenges. Andrea Westlund develops this idea into what she calls a “dialogical” conception of autonomy. ^{Footnote 10} On her view, the autonomous person recognizes a standing responsibility to engage in dialogue about the reasons for her beliefs. The objectionably heteronomous person is the one who, when pressed for reasons, defers immediately to some other authority rather than making any attempt to articulate her own justification. ^{Footnote 11}

We might question whether we have even this scaled back responsibility with respect to all of our normative commitments. Not all skeptical challenges obviously warrant the dialogical response that Westlund envisions. Perhaps some challenges, like that to the basic moral commitment considered above, can be appropriately dismissed. Another possibility is that at least some such challenges are better met in ways other than an attempt to articulate an argument or state reasons for one’s commitments. ^{Footnote 12} If either is true of at least some skeptical questions, then while advocates and detractors of the old-fashioned view may disagree over the range of cases that are well characterized this way, they again need not disagree over the value of autonomy as such.

But suppose that we grant that the autonomous person recognizes a dialogical responsibility to defend any normative view. The content of this responsibility has now been weakened enough that it is unclear that anyone need dissent. Notably, it entails no reason to abandon or doubt your view if you are currently inarticulate about its justification. Conversely, this approach makes heteronomy quite an extreme deficiency. ^{Footnote 13} Westlund characterizes the person who fails to meet the dialogical standard as one whose “deliberation itself takes a distinctly deferential form: . . . it is as if she simply cannot hear questions meant to challenge that deference. She is deeply unresponsive to critical challenges.” ^{Footnote 14} Eschewing education that would have these effects will not rule much out. Most importantly, it will not rule out anything approaching the full range of applications of the old-fashioned view, but at most only its most extreme instantiations. In fact, educating someone on the old-fashioned model —conscientiously bringing him into contact with and helping him to appreciate what he now values — would arguably make him more, rather than less, likely to meet the dialogical standard. In conversation, he can draw on and reference his experience of the appreciation of these values, and there is no reason why he should not.

Like the value of access to options, neither the value of refusal to take our normative commitments on the say-so of others, the ability to articulate reasons for them, nor the recognition of a responsibility to try, characterize points of disagreement between the educational aims of the two models. The question that divides them is apparently not whether students ought, in general, to have reasons for their views besides the testimony of their instructors, or ought to be able to defend their views by articulating these reasons. Disagreements will remain over which normative commitments students should be encouraged to form and which skeptical challenges they should be encouraged to take seriously. But these are, once again, substantive normative judgments about what rather than how the educated person ought to think.

C. Autonomy as reasons responsiveness

I turn now to a final attempt to characterize the autonomy that the progressive approach nurtures and the old-fashioned view thwarts. According to this conception, whether or not she can articulate the reasons for her view, the autonomous person has beliefs and convictions that are properly related to the reasons supporting them. This would allow the defender of autonomy to account for many of the conclusions drawn above. She could grant that teaching for autonomy is compatible with aiming to get students to form some unshakeable commitments (just when there is good reason for doing so) and with asking students to take some things, even many things, on testimony (just those for which testimony provides sufficient reason).

An educational approach would violate autonomy, so conceived, if it aimed to instill commitments regardless of the reasons for them. The old-fashioned view is at least occasionally represented as doing this, and characterizing it as indoctrination may suggest that it does. ^{Footnote 15} Instances of this sort of indoctrination are familiar. Here’s a paradigm case: Hoping to maximize sales of a certain soft drink, I design an advertising campaign that aims to get people to believe that the beverage in question is superior to all alternatives, and preferred by superior people. I aim to get them to purchase and consume my drink, and to share it with others. I aim to get them to value and love the drink, and even to shape their self-conception around this love. It is of no concern to me whether the beverage in question merits these attitudes. What I care about is sales, and the view that I aim to inculcate is nothing other than a means to this end. The presence of this extrinsic goal makes the ad campaign a clear instance of seeking to form normative attitudes without regard to reasons. This aim makes intelligible the motive to inculcate values and convictions that may come wildly apart from one’s own views about what is valuable.

But it should be obvious that the old-fashioned approach to education is singularly unlike this. Its practitioners do not seek to shape students’ allegiances for some merely instrumental reason. Helping students get it right about what is valuable, and come to love what is good, is the characteristic aim and self-conscious goal of the old-fashioned view. Quite unlike those heading up the advertising campaign, practitioners of the old-fashioned view aim to communicate the convictions and loves that they do precisely because they believe there to be good reason for just these commitments.

One’s convictions about which commitments are well supported by reasons might, of course, diverge from the convictions of these practitioners. But nowhere is it clearer that this is a substantive normative disagreement about content, not a formal disagreement about educational approach. It is, once again, a disagreement about what the well-educated person should think and what she should value. ^{Footnote 16}

We have discovered a familiar pattern. The progressive model purports to offer an alternative to the old-fashioned approach, one that eschews substantive commitments about what is worth loving or what makes for a good life. Instead of educating students with the aim of inculcating any such convictions, the progressive approach aims to make them autonomous and so able to settle these questions for themselves. But we have encountered difficulties in filling in a conception of autonomy that plausibly captures the critique of the old-fashioned view and provides an appealing basis for an alternative.

We have surveyed several candidates: (1) the autonomous person has access to alternative options; (2) the autonomous person does not rely solely on testimony, and has an ability to articulate the reasons for her views or at least recognizes a responsibility to do so; (3) the autonomous person is appropriately responsive to reasons. In each case, the advocate of the old-fashioned view seems just as much able to affirm these ideals as the advocate of the progressive view, and the disagreement between them is really over further judgments about how to apply the ideal. We make normative judgments not only about what to affirm and what to reject, but also about which matters to take to be settled, which to regard as matters of taste, what alternatives to consider, what to regard as plausible, when to accept something merely on testimony, when to seek to articulate the reasons for our beliefs, and a host of other matters. These are substantive ethical positions and our views about them are part of our overall normative outlooks. That is to say, once we start to fill in the details of our understanding of autonomy, we quickly run up against the very sorts of normative disagreements that the progressive approach tries to avoid, disagreements about what particular normative commitments an educated person should have.

One could take an extreme view about some or all of these questions. One could claim that a person should regard all of her commitments as doubtful, such that she could abandon them at will. Or one might think that no view can appropriately be taken on testimony, or that we should always be able to give an argument in defense for our views. We can call this a view that prizes autonomy if we like, but it doesn’t offer an alternative to an educational ideal that makes presumptions about what sort of person we should educate students to be. Instead, it advances one particular such ideal: that we should educate students to be thoroughgoing distrustful skeptics. Neither this, nor positions on the other extreme — that we should never doubt, never require independent corroboration of testimony, never feel responsible for offering a defense — will strike many people as appealing. The views that deserve attention and careful thought fall in the wide middle range that holds that each of these attitudes is appropriate with respect to some convictions and inappropriate with respect to others.