The Evanescence of Neutrality

In this short commentary, I draw on Alan Patten’s chapter about liberal neutrality in Equal Recognition, as well as the companion paper “Three Theories of Religious Liberty,” to argue that neutrality has a well-defined, yet limited, place in liberal theory, particularly in relation to controversies about the rights of religious citizens. ¹ I suggest that neutrality becomes evanescent the closer we get to hard cases of religious recognition and accommodation.

Alan Patten defends neutrality as a constraint on the pursuit of perfectionist policies. A liberal state has a strong (if defeasible) reason not to favour one conception of the good over another, and this reason is grounded in respect for people’s fair opportunity for self-determination (henceforth FOSD). The state should not favour one rival good over another out of respect for people’s interest in pursuing the conception of the good they hold. The conception of neutrality Patten favours is what he calls neutrality of treatment. Neutrality of treatment improves both on neutrality of impact and justification. For example, it can explain why some intuitively non-neutral policy, such as religious establishment, is wrong even though it can be justified neutrally (ER, 113). Patten argues that neutrality demands that the state give equal treatment to different conceptions of the good—Christianity and Islam, cricket and softball, to use his favourite examples. The state maintains neutrality between rival conceptions when, relative to an appropriate baseline, its policies are equally accommodating of those conceptions—without however equalizing outputs or impact (ER, 115).

One advantage of the theory is that Patten deflates some implausible claims sometimes made on behalf of neutrality as a central value of liberalism. Neutrality has a relatively modest place in the liberal architecture of justice. First, neutrality is a downstream value (ER, 108–9). It is not a foundational or Archimedean value but merely a pro tanto reason: it can be defeated by substantive liberal values such as FOSD. This is because neutrality is itself derived from FOSD. So conceptions of the good that deny or contradict FOSD have no good claim to be treated equally. Neutrality is not grounded in moral scepticism, but in robust liberal commitments such as FOSD. Second, neutrality is an inconclusive value. It does by itself not tell us which policy to adopt. Conceptions of the good must be treated equally—but sometimes this requires privatisation, at other times even-handed support (ER, 119). Separation between church and state is an example of privatisation (downward equality); positive provision of cultural goods is an example of even-handed support (upward equality). Different policies are required depending on the nature of the good in question.

Neutrality’s place in Patten’s liberal theory is modest, but it is robust. Neutrality provides guidance in real-world cases of fair treatment by the state of different conceptions of the good. Here are two illustrations:

Cricket v softball. Patten asks us to imagine a local authority that allows any team sport to sign up to use a newly released public field. Cricket and softball players are treated equally. It might well be that they will not be as equally successful in playing their favoured game: perhaps there are not enough cricket players to put together a proper game of cricket. But this would be worrying only if we were committed to neutrality of impact. Neutrality of treatment only requires that all sets of players have roughly equivalent opportunities to play their game (ER, 117).

Muslims v Christians. In “Three Theories,” Patten asks us to imagine a city with a strict zoning regime that reserves large areas for residential dwellings, yet has made exemptions for Christian churches. Neutrality of treatment requires that similar exemptions be offered to Muslim mosques too (TT, 22-3).

In such cases, the state has a fairness-based reason not to favour one conception of the good over another. Without some ideal of neutrality, we would not be able to explain what is wrong with reserving zoning exemptions to Christians. After all—we would say—Muslims enjoy full rights of freedom of religion, and there might be good public reasons for maintaining some historical tradition of Christian privilege. What Patten’s theory suggests is that unequal treatment of Muslims involves a distinctive breach of neutrality, and it is incompatible with FOSD properly conceived.

So far, so good. But let me now introduce a problem for liberal neutralists. Many liberal theorists accept that religious beliefs and practices should sometimes be singled out for special protection. Is this compatible with neutrality? The state should not favour cricket over softball, or Christianity over Islam, but can it permissibly favour religion over non-religion? The problem arises in acute form in justifications for religious exemptions. If the state exempts religious citizens from neutral laws of general application, is it not acting unfairly towards non-religious citizens? Is it not forcing some to subsidise the expensive preferences of others, in breach of FOSD? How can we reconcile neutrality with the notion that religion is special, that it is worthy of special respect? Many liberal theorists have run into what I call the Ethical Salience problem.

Consider, for example, Ronald Dworkin. Dworkin rejects exemptions on the grounds that neutrality requires that no discrimination be made between different kinds of views, in particular, between religious and non-religious views. Yet—I argue elsewhere—Dworkin does not follow through this premise. In his discussion of the peyote case in Employment Division v Smith (1990)—about whether Native Americans who use a sacramental drug, peyote, should be exempted from general drug legislation —he concedes that “equal concern requires legislature to notice whether the activity it proposes to prohibit or burden is regarded by any group as a sacred duty.” Sacred duties, presumably, are ethically salient in the way that purely recreational uses of drugs are not—and therefore should not be burdened unfairly. Dworkin also accepts ethical salience in his theory of state neutrality, which restricts the domain of neutrality to important matters of personal ethics, from religion to sexuality, and does not apply it to less salient matters such as mere preferences. The special ethical salience of certain commitments, then, poses a challenge to the liberal claim to neutrality. ²

Patten has a thoughtful response to the Ethical Salience challenge. First, he explicitly accepts, more readily than Dworkin, that some interests are weightier than others. Second, he suggests that these interests are not only religious but include a range of personal, ethical, sexual, and cultural commitments (ER, 168). This is a plausible (and, to me, compelling) answer. Religion is not uniquely special but belongs to a class of especially weighty interests. Call this the Weightiness argument. The Weightiness argument allows Alan Patten to justify both cultural accommodations (in Equal Recognition) and religious exemptions (in “Three Theories”). FOSD for citizens with weighty cultural or religious interests sometimes requires that they be specially accommodated. Patten also (rightly, in my view) suggests that culture and religion are not distinct categories of social or ethical experience, but rather lay on a continuum of weighty interests. Therefore, theories of cultural accommodation and religious exemptions can be articulated by reference to similar principles of liberal fairness.

So how does the Weightiness argument relate to neutrality? Patten argues that “neutrality applies more robustly [to weighty interests]” (ER, 136). It is easy to see why. People’s interests in self-determination in areas of religion, family commitments, sexuality, cultural attachments, and so forth, are particularly weighty. So it is crucially important that the state not favour one type of sexuality, religion, culture, or family model over another. Breaches of neutrality would be particularly egregious there. FOSD is infringed upon more seriously if one religion is favoured over another, less seriously if cricket is favoured over softball. This is because leisure preferences are not as weighty—that is, ethically salient—as religious beliefs.

So far, so good. But there is an intriguing ambiguity in the expression “neutrality applies more robustly” to weighty interests. It has an Orwellian ring to it: it seems to imply that some interests are more equal than others. If we are to balance a weighty interest against a non-weighty interest, it is difficult to see what guidance neutrality is able to provide. In what follows, I suggest that neutrality of treatment becomes evanescent the closer we get to hard cases concerning the place of religion in the liberal state.

Consider the following hypothetical case. A local authority must decide whether to allocate a newly released piece of land, either to a cricket club in search of a suitable field, or to a Muslim association in order to build a mosque. Depending on the facts of the case, there is a range of permissibly just solutions to the Mosque v Cricket case. However—or so I argue—neutrality of treatment gives little guidance in any of these scenarios, because the interests at stake cannot be measured up against a simple baseline of equal treatment. Here are some possibilities:

Cricket but not Mosque: this is OK if the authorities are otherwise committed to a principle of no support for religion as religion, perhaps because of a non-establishment norm (the reason might be that freedom of religion is best guaranteed by a “wall of separation” between state and religion). Neutrality applied to religion and neutrality applied to sport might support different policies, because neutrality is inconclusive as to which policies are most appropriate to the good in question. There is no interesting sense in which neutrality of treatment is applied between cricketers and Muslims.

Mosque but not Cricket: this is also OK if two conditions hold: support for religion is accepted as compatible with FOSD (here the reason might be that citizens cannot enjoy the worth of freedom of religion unless they have actual opportunities to practice it); and the building of other religious facilities (say, Christian churches) has already been accommodated by public authorities. If we add the premise that religious practice is more ethically salient than leisure activities, it seems fair to grant Muslims the newly released land. Is neutrality of treatment relevant in this case? Only in the limited sense that it underpins equality between different religions, qua weighty interests. But, again, there is no notion of equality between religion and sporting preferences.

Mosque and Cricket: this would also be OK, but it would not be directly derived from a principle of equal treatment between religion and sport: we would first need to know how other religions, and other leisure activities, are treated in the society in question. So, in all these cases, we need to “zoom out” to identify the appropriate baseline for equal treatment, but this baseline is good-specific. And because different goods require different policies of neutral treatment, it is impossible to know what neutral treatment across different goods entails: the thought that neutrality applies “more robustly” to weighty interests gives little guidance.

At this point, defenders of neutrality might reply that neutrality is not totally irrelevant to the Mosque v Cricket case. This is correct. Imagine the local authority acts out of hostility towards Muslims. Or imagine it holds cricket to be a superior, more elevated sport than softball. Or imagine it thinks that people should be pious believers rather than sporting hedonists. In all these cases, the subsequent decisions are illegitimate. But note that the conception of neutrality that does the work here is neutrality of intentions, not Patten’s preferred neutrality of treatment. This might suggest that there is more of a role for neutrality of intentions than Patten admits. At any rate, it is clear that, in the Cricket v Mosque case, neutrality of treatment does no work at all.

At this point, Patten might reply that I have misconstrued the domain of neutrality—the range of goods to which it applies. He could insist that neutrality of treatment properly and exclusively applies to rival goods. Neutrality of treatment is important between cricket and softball, and between Christianity and Islam, because people can be either Christian or Muslims, and cricket players or softball players. If the state exclusively accommodates Christians, or cricket players, it does not respect FOSD for Muslims or softball players. But—the argument would continue—neutrality need not apply in the same way to the Mosque v Cricket case, because these goods are not rival in the same sense. One can be a cricketer and a Muslim—a cricket field and a mosque are not rival goods and, therefore, neutrality of treatment is not required between them. ³

What is notable about this answer is that it makes neutrality irrelevant to a wide range of controversies about the just place of religion in the liberal state. This is because the set of issues raised by the Mosque v Cricket case can be generalised. One of the greatest puzzles of contemporary liberal political theory, it is sometimes thought, is how to treat “religion” and “non-religion” equally. Consider the following controversies. Should the state subsidize religious groups providing social services such as health care on the same basis as non-religious groups? Should it fund faith schools as well as secular schools? Should the teaching of Intelligent Design be treated on a par with the teaching of Darwinian Evolution? Should the state grant exemptions from general laws to non-religious as well as religious citizens? Can it grant more privileges to religious than to non-religious associations (say, by exempting them from anti-discrimination norms)? Can the state promote a national religion, as well as a national culture? In all these cases, some people think that “religion” and “non-religion” are rival goods, and they raise a question about equality of treatment. But this is mistaken: equality of treatment will play some role in adjudicating these disputes, but it will be a limited, indeed evanescent, role.

This is because “religion” and “non-religion” are no more rival goods than the mosque and the cricket field in my example. Fair resolution of these controversies requires “zooming out” and locating them within a broader account of the justice of welfare entitlements and institutions, the purposes of a liberal education, the fairness of burdens on conscience, the scope of freedom of association, the bases of national identity and citizenship, and so forth. To be sure, the ideal of neutrality will rule out animus towards religion or appeal to its truth (neutrality of intention), and it will rule out accommodating one religion over another (neutrality of treatment). But it will tell us very little—if anything at all—about what it means to treat religious and non-religious citizens equally. In all these controversies, it is not very helpful to see religion and non-religion as rival goods.

I conclude that neutrality has an even more modest place in the liberal architecture of justice than Patten suggests. And, in fact, when Patten discusses such controversies as exemptions from drug use regulation on religious grounds (Smith), corporate exemptions from the contraception mandate (Hobby Lobby), or conscientious objection (Sherbert), he makes little appeal to the idea of neutrality of treatment. Substantive liberal ideals such as rights to health care, gender equality, fairness in the distribution of costs, and freedom of religion do all the work. This is exactly as it should be. If Patten’s claim is that religion and non-religion are not, in the relevant sense, rival (which I think is correct), then it means that neutrality has become evanescent just when its guidance is often seen to be most required—in controversies about the just place of religion in the state.