Liberal Neutrality: Constructivist, not Foundationalist

In defending the principle of neutrality, liberals have often appealed to a more general moral principle that forbids coercing persons in the name of reasons those persons themselves cannot reasonably be expected to share. Yet liberals have struggled to articulate a non-arbitrary, nondogmatic distinction between the reasons that persons can reasonably be expected to share and those they cannot. The reason for this, I argue, is that what it means to “share a reason” is itself obscure. In this paper I articulate two different conceptions of what it is to share a reason; I call these conceptions “foundationalist” and “constructivist.” On the foundationalist view, two people “share” a reason just in the sense that the same reason applies to each of them independently. On this view, I argue, debates about the reasons we share collapse into debates about the reasons we have, moving us no closer to an adequate defense of neutrality. On the constructivist view, by contrast, “sharing reasons” is understood as a kind of activity, and the reasons we must share are just those reasons that make this activity possible. I argue that the constructivist conception of sharing reasons yields a better defense of the principle of neutrality.


INTRODUCTION
The liberal principle of neutrality holds that the state ought to remain neutral among competing conceptions of the good. For most liberals, I suspect, this principle gains whatever plausibility it has from the intuitive appeal of a more general moral principle-call it the Kantian principle-according to which proper respect for the moral status of persons requires that we do not coerce them in the name of reasons that they themselves cannot reasonably be expected to share. 2 Since all state actions are in some sense coercive, the principle of neutrality follows from the Kantian principle almost immediately, provided only that conceptions of the good are among the reasons that persons cannot reasonably be expected to share. But while many friends of the Kantian principle have the intuition that conceptions of the good are such that others cannot be expected to share them in the right way, whether and why this should be the case is notoriously difficult to explain.
In this paper, I hope to make some headway towards such an explanation by examining what it means to share a reason. By way of beginning, I'd like to make a rough-and-ready distinction between two different ways in which persons can be said to share a reason, and two corresponding approaches to political philosophy. In one sense, two persons can be said to share a reason when the same reason applies to each of them independently. In this sense, they share it simply because they both have it, in the same way that a child might be said to "share" her mother's eyes. Philosophers who understand sharing reasons primarily in this sense tend to adopt an approach to political philosophy that I will call "foundationalist." On the foundationalist approach, the reasons that persons have are taken as basic, or foundational, and the reasons we share are just the reasons we each have independently (or the reasons we each have in the right way, however that is spelled out).
These shared reasons, once identified, are supposed to serve as the foundation for a theory of justice.
But there is another sense in which two people might be said to share reasons. They might be said to share reasons, not because a state of affairs obtains where each happens to have the same reasons as the other, but rather because they are mutually engaged in the activity of offering and accepting reasons back and forth. This sense is closer to what we ordinarily mean when we speak of "sharing," although its implications for neutrality have not (I think) been sufficiently appreciated. This understanding of what it means to share reasons belongs to an approach to political philosophy which I will call "constructivist," following John Rawls and Christine Korsgaard. 3 On the constructivist approach, the reasons we share for political purposes are not taken as given; rather, they are understood as emerging from the activity of offering and accepting reasons from one another.
In this paper, I attempt to show that the foundationalist approach to political philosophy fails to satisfy the demands of the Kantian principle (and by extension the principle of neutrality) because it is tied to a crudely individualistic conception of what it means to share reasons. "Sharing as each having," I will claim, is too weak to provide a shared foundation for political life. I will argue that the constructivist approach fares better because it understands sharing reasons as an activity we engage in together. By examining the practical presuppositions of this activity, we can together construct the reasons we must share in order for the activity to unfold in a way that is free and fair to everyone. My conclusion will take the form of a conditional: if we are moved by the Kantian principle and are thus concerned to interact with others according to shared reasons, then we have reason to abandon the foundationalist approach to political philosophy and become constructivists.

POLITICAL FOUNDATIONALISM
For the foundationalist, the first question of political philosophy is the question of what reasons we have. Once these reasons are identified, erecting a theory of justice is simply a matter of weighing and combining these foundational reasons in the right way. On this characterization, all varieties of political-moral realism (or at least all varieties I'm aware of) are forms of foundationalism; however, for the purposes of this paper I am concerned only with those foundationalists who are also neutralists-those who make some distinction between the reasons we have simpliciter and the reasons we may reasonably expect everyone to share for political purposes.
The best-known example of a foundationalist who also endorses a form of neutralism is probably Thomas Nagel, who distinguishes in his Equality and Partiality between the values a person may be justified in holding and the values she may be justified in imposing on others. Nagel makes this distinction because he believes that the stan- dards of justification appropriate to an individual when deciding how to live her own life are lower than the standards appropriate to society as a whole when deciding what values to coercively impose. On his view, some values are sufficiently justified that an individual could reasonably accept them for her own life, but not sufficiently justified that she may reasonably employ the coercive power of the state to force everyone to do the same (Nagel 161-2). Nagel's commitment to neutrality-or "toleration," to use his preferred term-is motivated by the Kantian principle mentioned above, according to which we ought not coerce others in the name of ends they cannot reasonably be expected to share (Nagel 159). He understands "reasonableness" as an epistemological concept: we treat others as "mere means" when we force them to serve an end that the evidence does not compel them to accept (Nagel 159-61).
Yet Nagel recognizes that disagreement about which ends are supported by the evidence is common, especially when it comes to arguments about conceptions of the good. And finding agreement on some public standard of evidence for assessing ends would be just as difficult as finding agreement on the ends themselves, so the epistemological route is a dead-end. To circumvent this problem, Nagel introduces a quasi-contractualist procedure to separate the reasons we can reasonably be forced to share from those we cannot: when faced with the question of whether to use the power of the state to impose some value over the objections of others, Nagel argues that respect for those others requires that we re-phrase the issue in neutral language-i.e., language acceptable to all concerned, language that doesn't bias the question in favor of one answer or another-and ask ourselves, "how would I like it if someone did that to me?" (Nagel 162-4) So, for example, if we want to know whether we may establish our Catholic faith as the state religion and force everyone to go to mass, we do not ask ourselves how we would like it if someone were to impose "the true religion" or even "the Catholic religion." We do not ask these questions because, as Catholics, we would have no problem with someone else doing either of these things; these ways of phrasing the question are biased towards the answer we favour. Instead, we ask how we would like it if someone were to impose "their religion," a description of the proposed policy that Catholics and non-Catholics alike may be expected to accept. 4 And as Nagel notes, as soon as we ask ourselves how we would like it if someone did that to us, we see quite clearly that we wouldn't like it very much at all (Nagel 162).
The first thing to notice about this procedure for resolving disputes about ends is that it is clearly biased toward the answer Nagel desires. If to re-describe the policy in neutral terms is to abstract from the reasons we have for favouring it, then the procedure will always yield a policy of abstention. Once we abstract from the (putative) value of the end, it's clear that we no longer have reason to support it; the appropriate response toward neutral value is indifference. And when we ask ourselves how we'd like being forced to serve some neutral value, we should respond with indignation.
Strangely, Nagel does not explain why in this role-reversal exercise "maximizing aggregate utility" or "promoting virtue" are most neutrally re-described as "promoting one's own conception of the good," something we surely wouldn't like another to do to us, while "protecting basic rights" or "preventing inequality" (i.e., preferred liberal values) are neutral as stated. Presumably this reflects nothing more than his own confidence that liberal values are conclusively justified from the impersonal point of view, whereas non-liberal values are not.
For present purposes, what is important about Nagel's argument for neutrality is how quickly the reasons Nagel finds us to share collapse into the reasons he thinks we have. This is not accidental. For the foundationalist, the reasons we share are just a sub-set of the reasons we have. And how can we non-arbitrarily pick out the reasons we share from all the reasons we have, except on the basis of the reasons the foundationalist thinks we have? And if the reasons we have are the very matter in dispute, it's not clear that we've made any progress towards satisfying the Kantian principle's requirement of providing those who disagree with reasons they can share.
To be clear, my objection to Nagel's argument is not that he is wrong about what reasons we share; my objection is that he has not given the person who disagrees with him any reason to think he's right. This is because the only reasons Nagel can appeal to in justifying his contractualist procedure are the reasons he thinks we havereasons which, ex hypothesi, the person who disagrees with him rejects.
I have presented Nagel's view too crudely, no doubt, but the particular contours of his view need not concern us too much here. I think Nagel's view illustrates a general problem with foundationalist attempts to carve out a space of neutral reasons. On the foundationalist's view, the question of what reasons we share boils down to the question of what reasons we all have; thus the foundationalist's own views about what reasons we have will necessarily determine the reasons she finds us to share. Her own private assessment of the reasons provides not only the raw data to which the standard of shareability is applied, but also (and more troublingly) the content of the standard itself. The problem, again, is not that she may be wrong in her assessment of the reasons; the problem is that she has nothing more to say by way of justification to someone who disagrees with her assessment-nothing, that is, beyond "that's how I see it." And this seems to be inconsistent with the Kantian ideal of interacting according to reasons we can share. When the foundationalist says that a reason is one that we cannot reasonably expect others to share, what she means at bottom is that it's not a good enough reason. This becomes quite clear in Nagel's argument for toleration when he says: I think the problem is that there is no higher-order value of democratic control or pursuit of the good abstractly conceived which is capable of commanding the acceptance by reasonable persons of constraints on the pursuit of their most central aims of self-realizationexcept for the need to respect this same limit in others. (Nagel 164) Here Nagel flatly denies the truth of all non-egoistic teleological moral theories; there simply isn't a reason to pursue the good whatever it is, he says, or at least there isn't a reason strong enough to override our personal interest in self-realization. (To this, I think, the teleologist is perfectly justified in responding: "Yes, there is"). This insistence is especially clear in Nagel's view, but I think it's a necessary feature of all foundationalist defenses of the principle of neutrality. If the reasons we share are just the reasons we (really) have, then the reasons we can't share are just the reasons we don't (really) have.
The objection I am raising here against the foundationalist picture can be summarized as follows: on the foundationalist view, the first political question is the question of what is "really" valuable, what reasons we "really" have. But this question can only be answered from the perspective of some particular person-you or I, for example. If you and I disagree about the answer to that question, we then have a further question to answer: the question of whose perspective on reasons should be the authoritative one for us. The obvious answer to the question of which perspective should be authoritative is, of course, "the correct one." But that answer only pushes the question back: whose perspective on the impersonal point of view is correct? Any answer to this question must always presuppose authoritative access to the impersonal point of view, it seems to me, and thus must always fail to persuade those who see things differently.
The foundationalist may be able to persuade herself that she is acting reasonably and that her use of the coercive power of the state is justified, and she may indeed be right. But if the aim of political justification is, as the Kantian principle suggests, not merely to satisfy ourselves when we exercise power but also to give reasons to those over whom power is exercised, then foundationalism fails on that score (Rawls develops this point in PL 100-101).
The underlying reason for this failure is that the foundationalist conceives of justification in terms of each person's relationship to the impersonal reasons; to see things objectively is, on this view, to see things from the point of view of the impersonal reasons. What is lacking on the foundationalist view is an objective yet shared perspective from which we can consider the reasons together. The foundationalist picture cannot provide such a perspective, at least not one sufficient to satisfy the Kantian principle, because the shared perspective is (for the foundationalist) the perspective of the impersonal reasons. Until we agree about the reasons we have, we don't yet have a shared perspective, and yet the problem posed by the Kantian principle is precisely that we can't agree about what impersonal reasons we have.

POLITICAL CONSTRUCTIVISM
Constructivism is often thought to begin in skepticism about the existence of a realm of independent moral truths, or in skepticism about the possibility of reliable access to that realm. Although I must confess that I have those kinds of doubts, political constructivism need not depend on them. Political constructivism, as I understand it, neither affirms nor denies either the existence of or the possibility of our access to construction-independent moral truths. Instead, what political constructivism denies is the possibility of a non-constructed shared point of view from which these values can be publicly assessed and balanced.
In the previous section, I tried to locate political foundationalism's failure to provide such a public standpoint in its conception of what it is to share a reason. On the foundationalist view, we share reasons just in the sense that each of us independently has the same reasons.

154
Thus until we know what reasons we have, we cannot know what reasons we share, which is to say we cannot take up the shared point of view. Thus the shared point of view cannot be used to settle disagreements about the reasons we have-or at least not in a way that will satisfy anyone. Political constructivism avoids this particular problem by understanding "sharing reasons" not as a state of affairs but as an activity. On the constructivist view, we share reasons in the sense that we are constantly giving them back and forth. Political constructivism begins by examining our practices of sharing reasons: of offering reasons to one another, of coming to agreement (or not), of holding each other responsible for the reasons we give and the agreements we make, and so on.
Rather than understanding a reason as something independent of these practices, something to which our practices must conform, political constructivism employs a conception of reason in some sense internal to these practices: on the constructivist view, a reason is a claim that one person makes against another (or against herself). 5 Unlike on the foundationalist view, where our access to reasons is something ineluctably private, constructivism operates with an essentially public conception of a reason.
What is wanting on the constructivist view is some criterion of valid claims, some objective standard of what claims we all must recognize if this practice of sharing reasons is to take place in a way that's free and fair to everyone. This standard cannot consist in some external standard of good reason, for then we would encounter the same problem as on the foundationalist view. The standard must be, in some sense, internal to our practices.
Rather than looking outside the practice, then, we must begin by examining the practice itself and what conditions are necessary in order for it to function well. We see that in order for persons to be engaged in the practice of offering and accepting reasons, two conditions must be met. First, persons must be understood as having some conception of the good, some rational plan of life, that they are concerned to advance; in the absence of such a conception they would have no basis for making claims against others. Second, persons must have some capacity to be moved by the claims of others when these claims are legitimate; they must have some sense of justice. These two conditions, a capacity to form a conception of the good and to pursue it on terms fair to others, are just the two moral powers Rawls attributes to citizens in his model conception of a well-ordered society, and they emerge naturally from a conception of social cooperation based on the give-and-take of reasons.
The route from this conception of the person to a political conception of justice is familiar. Rawls's device of the original position is meant to represent the way that the idea of persons as equally givers and receivers of reasons leads to certain concrete political principles. The original position is not meant to identify the reasons that bind us through their construction-independent force; it is a "device of representation" meant to draw out the presuppositions of our shared practices. The problem facing the denizens of the original position is to find the conditions under which the claims of one person can harmonize with the claims of every other in a systematic way-to find the conditions of possibility of what Rawls calls a well-ordered society.
Rawls's solution, roughly, is that such a systematic harmony is possible only against a political background which provides each person with an adequate scheme of basic rights and liberties and the largest possible share of the social product. In other words, it is possible only where each person has adequate freedom from the interference of others in the setting of her own plan of life (consistent with the like freedom of others) and has the greatest possible means to execute that plan (consistent with availability of like means for others). It is not the purpose of this paper to defend Rawls's two principles as the only solution or even the best solution to this problem; I only want to emphasize the way in which Rawls's view is a natural development of a certain understanding of what it is to share reasons, one which need not appeal at any point to an external, material standard of value.
The constructivist aims to construct a shared point of view from which the claims persons make against one another can be adjudicated-to construct, in other words, the "we" of the political community. This is why Rawls frequently speaks of his two principles of justice as providing an "objective point of view" (e.g. at PL 115). There is a certain sense in which the very idea of reasons as claims already presupposes the existence of some "we;" it presupposes, at least, that persons who interact by giving each other reasons share a common moral space, the space of reasons. Political constructivism inquires into the background conditions that must exist for persons to live in this space together, to take responsibility for the claims that they put to themselves and others. On the constructivist view, at a minimum, treating people as givers and takers of reasons requires a background which protects them from coercion, manipulation, and at least the more severe forms of economic hardship (although of course people may disagree about the particulars). The point is that when we take up the public point of view, the point of view of the "we," we are deliberating about what is fair to persons conceived in a certain way, what we think is needed for people to interact with each other as reasonable and rational creatures.
In the closing section I want to say a few words about how constructivism is superior to foundationalism in satisfying the Kantian principle. Before I do that, however, I want to anticipate an objection. 6 It may seem that the constructivist view requires at least one foundational reason in order to get off the ground, namely, the reason we have (whatever it is) to engage in the practice of sharing reasons in the first place. If true, this would immediately give rise to a dilemma: either there is some such foundational reason, in which case constructivism falls before the very same objection I raised against foundationalism in the second section of this paper, or there is no such reason, in which case our practices of sharing reason are arbitrary. But I don't think constructivism need posit any such foundational reason. Rawls leaves the question of what reason we have to engage in the practice of sharing reasons to each person's private comprehensive doctrine. Any reasonable person will recognize that she has some reason to make claims against others and to respect the claims others make against her, whatever that reason may be (this is actually true by def inition on Rawls's view-see PL 49f). Furthermore, for those who are not reasonable in this sense-for those who deny that they have reason to share reason-there doesn't seem to be any reason, foundational or otherwise, that could satisfy them; we have nothing to say to them, and we must coerce them in the name of reasons they unreasonably reject.
It might seem that the foundationalist could give the very same answer in response to my objection to her view: "It's true," she might say, "that I have nothing to say to the person who denies that he has a reason to respect the rights of others or to contribute to the common good, but such a person is manifestly unreasonable. He obviously does have such reasons, and the mere fact that he is too foolish or too stubborn to recognize it should not prevent us from forcing him to respect those reasons in spite of himself. The construc-tivist view helps itself to the very same conclusion." But I think the constructivist view is different in an important respect. When a foundationalist declares a person "unreasonable," it is because that person rejects this or that particular reason-our reason to respect the rights of others, for example, or our reason to contribute to the common good. On the other hand, when a constructivist declares a person unreasonable, what she means is that the person in question is not willing to participate as an equal member in our practice of sharing reasons-that person refuses to listen to reasons generally, not merely to this or that reason. These are quite different things.
I do not mean to suggest that, within the constructivist framework, there will not be disagreement as to which background reasons are essential to our reason-giving practices and which are not, and thus as to which reasons persons may be forced to share (or forced to act as if they share). Nor do I mean to suggest that, when we step back to the objective point of view of the reasonable and rational citizen in order to settle these disagreements, there will not be further disputes about how to specify that point of view. But I do claim that the constructivist view has a few important advantages over the foundationalist view when it comes to managing disagreements about reasons. First, it sets out a distinct point of view from which such disagreements may be settled, namely, the point of view of the reasonable and rational citizen; this constrains the scope of the disagreement in important ways and may make a resolution possible. The foundationalist picture, by contrast, offers no such standpoint: on the foundationalist picture, disagreements about reasons are always immediately about the reasons themselves (or so I have argued). Second, even if such disagreements cannot be settled, they are nonetheless likely to be narrower in scope because of the constraints imposed by the constructivist procedure, hopefully to the point where any available solution is likely to be reasonably (if not fully) just.

CONCLUSION
By way of closing, I would like to situate the foundationalist and constructivist pictures with respect to their interpretation of the Kantian principle with which I began, the principle according to which respect for persons requires that we not coerce them in the name of ends they cannot by expected to share. The foundationalist interprets this as a moral constraint. For the foundationalist who is not also a neutralist, it is at best a trivial reminder that we shouldn't act on bad reasons; if we shouldn't act on bad reasons then a fortiori we shouldn't coerce others in their name. But for the neutralist foundationalist, for whom there may be some difference between the reasons that are good enough for me and the reasons that are good enough for everyone, the principle requires that we constrain our political justifications in certain ways. The objection I've raised against this view is that, no matter how we constrain our justifications, in the end the foundationalist must always rest her case on the reasons as she sees them. If the Kantian principle implies that this is not enough in the first instance, I don't see how adding more steps to the argument will ever suffice. The constructivist, by contrast, understands the Kantian principle as more than a constraint; for the constructivist the Kantian principle describes an ideal of social cooperation according to the giveand-take of reasons rather than the threat of force. If we are to realize a world where people interact according to shared reasons, then there are certain reasons that we must share. The reasons we must share are given, not by the independent realm of value, but by this ideal of social cooperation itself. The reasons we must share are the reasons that protect and foster our capacities to give and take reasons.
If what I have said so far is right, then the foundationalist approach to neutrality has it exactly backwards: we do not begin doing political philosophy by weeding out the reasons or values that seem objectionable and then build our political order on the values that remain. We begin instead by constructing the political point of view itself; it is this point of view which will tell us the reasons we share for political purposes.
As I said at the outset, my conclusion in this paper is conditional: if we are moved by the Kantian principle and thus concerned to interact with others according to reasons we can share, then (I suggest) we have reason to abandon the political foundationalist project and become political constructivists. It is a commonplace in philosophy that one person's modus ponens is another's modus tollens, and some may see in this paper reason to abandon the Kantian principle because it costs too much (philosophically speaking) to satisfy. In response, I would stress that one need not be a constructivist "all the way down" to appreciate the distinctly political problem posed by the privacy of foundationalist reasons and thus to be a political constructivist. And insofar as the Kantian principle expresses a powerful moral ideal, I think it is one we would be mistaken to abandon.