Philosophical Studies

, Volume 162, Issue 3, pp 645–664 | Cite as

Can liberal perfectionism justify religious toleration? Wall on promoting and respecting



Toleration is perhaps the core commitment of liberalism, but this seemingly simple feature of liberal societies creates tension for liberal perfectionists, who are committed to justifying religious toleration primarily in terms of the goods and flourishing it promotes. Perfectionists, so it seems, should recommend restricting harmful religious practices when feasible. If such restrictions would promote liberal perfectionist values like autonomy, it is unclear how the perfectionist can object. A contemporary liberal perfectionist, Steven Wall, has advanced defense of religious toleration that grounds perfectionist toleration in an innovative account of reasons of respect. He thus defends perfectionist toleration on two grounds: (i) the appropriate manner of responding to perfectionist goods like autonomy and membership is to respect the religious choices of others; (ii) citizens can acquire reasons to respect the religious choices of others through internalizing a value-promoting moral and political code. I argue that both defenses fail. The cornerstone of both arguments is the connection Wall draws between reasons to promote value and reasons to respect it. I claim that Wall’s conception of the relationship between promoting and respecting value is inadequate. I conclude that the failure of Wall’s defense of perfectionist toleration should motivate liberal perfectionists to develop more sophisticated accounts of normative reasons. The viability of a truly liberal perfectionism depends upon such developments.


Perfectionism Liberalism Religious toleration Liberal perfectionism Steven Wall Toleration 

Liberalism is committed to the toleration of a broad range of religions and religious practices, despite the fact that citizens significantly disagree about which religions are true, false and harmful. Toleration is perhaps the core commitment of liberalism, from its inception in figures like John Locke, through luminaries like John Stuart Mill, to contemporary liberals like John Rawls (Locke 2009, p. 1; Mill 1978, p. 9; Rawls 2005, p. 10).1 This seemingly simple feature of liberal societies creates tension for liberal perfectionists, who are committed to justifying religious toleration primarily in terms of the goods and flourishing it promotes. Perfectionists, so it seems, should recommend restricting harmful religious practices when feasible.2 For instance, why shouldn’t perfectionists advocate, as Gerald Gaus has suggested, “a regulation that would allow only adults into church services?” (Gaus 2003, p. 149). If church attendance in childhood makes it harder for citizens to become autonomous adults, liberal perfectionists should regard church attendance as a great harm. If such a regulation would promote liberal perfectionist values like autonomy, it is unclear how the perfectionist can object.

A contemporary liberal perfectionist, Steven Wall, has advanced an engaging defense of religious toleration that departs from standardized debates between political consequentialists and deontologists by arguing that citizens must tolerate others for the right reasons. Wall’s defense of perfectionist toleration thus promises to include what is attractive about deontological approaches: they attach importance to acting on the right sorts of reasons. The defense is arguably the most powerful among liberal perfectionists. Should it fail, the case for perfectionist toleration is weakened along with the prospect for a truly liberal perfectionism. A successful defense would be a great boon for liberal perfectionism. Suspicions of authoritarianism surround opposition to perfectionist political theories; a compelling perfectionist defense of religious toleration would help defuse these suspicions.

Wall defends perfectionist toleration on two grounds: (i) the appropriate manner of responding to goods like autonomy and membership is to respect the religious choices of others; (ii) citizens can acquire reasons to respect the religious choices of others through internalizing a value-promoting moral and political code. I argue that both defenses fail. The cornerstone of both arguments is the connection Wall draws between reasons to promote value and reasons to respect it. However, Wall’s conception of the relationship between promoting and respecting value is inadequate. I conclude that the failure of Wall’s defense of perfectionist toleration requires liberal perfectionists to develop a more sophisticated account of normative reasons.

I will proceed in five parts. In Sect. 1, I explain Wall’s conception of toleration. In Sect. 2 I outline the structure of Wall’s liberal perfectionism and explore Wall’s understanding of the relationship between the good and respect. Section 3 rejects Wall’s appeal to the perfectionist goods of autonomy and membership to defend religious toleration. Section 4 criticizes Wall’s rule-teleological defense of tolerating disvalue. Section 5 concludes by examining how a liberal perfectionist defense of toleration might adapt to the challenges presented in Sects. 3 and 4.

1 Wall’s conception of toleration

An individual is tolerant of religious differences when she refrains from interfering with religious practices that promote values incompatible with her own. For example, a Christian is tolerant of an atheist when she refrains from interfering with atheistic practices that discourage the value of faith in God; similarly, an atheist is tolerant of a Christian when he refrains from interfering with Christian practices that promote the value of faith in God. The tolerant person must refrain, therefore, from repressing some actions that are harmful in the sense of reducing the achievement of perfectionist goods. In Wall’s view, a tolerant person must also permit others to engage in religious practices out of respect for them, not merely because she doesn’t care what they do. Tolerance is more than “nonchalance” (Wall 1998, p. 63). The tolerant person must also avoid the attitude of the “efficient persecutor” who is inclined to “repress other people because they are engaging in conduct he deems wrong” and who only restrains himself “because he has been persuaded that in the present circumstances repression would be either ineffective or counterproductive” (Wall 1998, pp. 63–64).3 An efficient persecutor only resists repression if it is costly or ineffective. Wall believes these are the wrong reasons to tolerate.

A truly tolerant attitude falls “between nonchalance and the efficient persecutor” (Wall 1998). The tolerant person both believes that some religious practices are harmful and restrains his motivation to repress those practices. He “not only restrains his disposition to repress, but is morally motivated to do so.” This moral motivation includes “the recognition that others are entitled to respect”; the tolerant person believes that if John thinks it is permissible “to engage in some type of conduct, then this provides a reason … for the tolerant person to let [John] do it” (Wall 1998, pp. 64–65).4 If Reba recognizes a reason to tolerate John, she is tolerant when that recognition “motivates [her] restraint” (Wall 1998, p. 65). Following a standard view, Wall claims that a person is tolerant if:
  1. (i)

    He disapproves of certain types of conduct, behavior, social practices, etc.;

  2. (ii)

    He has a disposition to repress that which he disapproves of;

  3. (iii)

    He has at his disposal—and is aware that he has at his disposal—some effective means for repressing what he disapproves of; but,

  4. (iv)

    He restrains his disposition to repress; and,

  5. (v)

    He does so for the right type of reason (Wall 1998, p. 65).5


2 Liberal perfectionism and reasons of respect

Perfectionism holds that political institutions have a core duty to promote well-being or human flourishing. Wall defines perfectionism as the view that “political authorities should take an active role in creating and maintaining social conditions that best enable their subjects to lead valuable and worthwhile lives” (Wall 1998, p. 8).6 He takes perfectionism to consist of four commitments:
  1. (i)

    That some ideals of human flourishing are sound and can be known to be sound;

  2. (ii)

    That the state is presumptively justified in favoring these ideals;

  3. (iii)

    That a sound account of political morality will be informed by sound ideals of human flourishing;

  4. (iv)

    That there is no general moral principle that forbids the state from favoring sound ideals of human flourishing, as well as enforcing conceptions of political morality informed by them, when these ideals are controversial and subject to reasonable disagreement (Wall 1998, p. 8).

Wall argues that a political value or ideal of human flourishing is sound when there is reason to believe that the ideal or value is sound and the ideal or value possesses some property or set of properties which make it sound (Wall 1998, p. 9).7Liberal perfectionists hold that freedom is a “distinct value” (Raz 1986, p. 19). Thus, liberal perfectionists must defend liberal principles by arguing that traditional liberal freedoms are either constitutive of well-being or instruments towards promoting or respecting well-being. Perfectionist defenses of religious toleration are therefore instrumental, intrinsic (non-instrumental) or both. Religious toleration is justified if it promotes or respects some sound value or ideal or because it is a sound value or ideal itself (or both).

Liberal perfectionism is not a merely teleological political theory despite maintaining that the state is justified in “favoring … ideals” and helping “subjects to lead valuable and worthwhile lives” (Wall 1998, p. 8). While perfectionism affirms that one way to appropriately respond to value is to promote it, Wall claims one can also appropriately respond to value by respecting it, since a number of goods “command respect for their own sakes” (Wall 2003, p. 253). Often reasons to respect values like autonomy do not outweigh reasons to promote it (Wall 2003, p. 237). However, sometimes “we have reason to respect the good even if not doing so would lead to a greater good” (Wall 2003, p. 238).8 Some reasons of respect permit or require sacrificing a greater good: “we [sometimes] have reason to respect the good even if not doing so would lead to greater good being promoted” (Wall 2003, p. 238).9 Thus, reasons to promote and respect value can conflict and must sometimes be traded-off.10

It bears noting that Wall distinguishes between reasons to respect (i) persons and (ii) the good.11 Reasons to respect persons require us “to not hinder, suppress or destroy” a person’s good along with doing so out of respect for that person (Wall 2003, p. 238). But reasons to respect the good only seem to require us not to “hinder, suppress, or destroy” value; it is not clear whether an attitude of respect is required to respect goods, though nothing hangs on this for now. I will assume that reasons to respect persons and goods can overlap in lieu of a statement to the contrary. For instance, there might be two distinct reasons of respect to not suppress autonomy—a reason to respect the good of autonomy per se and a reason to respect persons by not suppressing their autonomy. Not insignificantly, Wall only claims that reasons to respect goods can be traded-off with reasons to promote goods. He does not speak to whether reasons to respect persons can be traded-off with reasons to promote the good of persons. I shall assume that Wall affirms that such reasons can be traded-off in the absence of reasons otherwise.

The introduction of reasons of respect into liberal perfectionism is innovative and potentially powerful. The liberal commitment to religious freedom raises the familiar “paradox of toleration” which holds that we must permit activity we think is wrong, or to follow Bernard Williams, “Toleration … is required only for the intolerable” (Williams 1996, p. 18). For Wall, the paradox of toleration creates a problem for “explaining how a person with sound evaluative beliefs could have reason to respect objects that he rightly disesteems” (Wall 2003, p. 232; Christiano 2008). Steven Smith puts it simply: “Given the choice, why should we knowingly put up with error?” (Smith 2008, p. 243). Wall’s response requires the crucial idea of a trade-off of reasons to respect and promote value. To justify perfectionist toleration, Wall must defend the existence of reasons to respect the good and persons that outweigh reasons to promote value in cases of conflict. If some entities command great respect, then there may be reason to tolerate disesteemed religious practices since sometimes there will be sound reasons not to promote value. Accordingly, the paradox of toleration may dissolve as we should knowingly tolerate error out of respect for the good of others. Reasons of respect may thereby enable perfectionists to defend religious toleration.

Wall claims that religious toleration is justified because it respects significant goods like autonomy and membership. For his argument to succeed, reasons to respect value must reliably outweigh reasons to promote value in many cases of conflict. For example, if John’s reason to promote Reba’s good is often stronger than his reason to respect her disvaluable religious practices, his reason to respect cannot ground toleration; liberal religious toleration requires a readiness to sacrifice the promotion of immense value on behalf of respecting others. Further, and also critical, Wall must establish that reasons to respect others necessitates a respectful attitude towards their religious choices. Otherwise, he has failed to justify genuine toleration by his own lights. To be tolerant, John must have an attitude of respect towards Reba beyond the mere recognition of a reason not to repress her belief.

3 Autonomy and membership

Wall’s first argument for perfectionist toleration begins with an examination of the candidate goods autonomy and membership. Wall must contend that the most appropriate response to autonomy and membership is to respect them, that is, to refrain from suppressing them for the right reasons to the degree required by a truly liberal level of religious toleration. Three criticisms show that autonomy and membership are unsuccessful candidates.12 First, for any moral code that requires religious toleration, there exists an intolerant code that promotes the candidate goods more effectively. Call this the criticism from Comparative Inferiority. Second, supplementary pragmatic defenses of religious toleration are inconclusive and unsatisfying; arguments that rely on the imprudence of intolerance fail. Call this the criticism from Pragmatic Inconclusiveness. Finally, the candidate goods cannot justify toleration for the right reasons. Call this the criticism from Heteronomy.

3.1 Autonomy

Liberal perfectionists argue that autonomy is a central part of personal good.13 For Wall, autonomy is an,

… ideal of people charting their own course through life, fashioning their character by self-consciously choosing projects and taking up commitments from a wide range of eligible alternatives, and making something out of their lives according to their own understanding of what is valuable and worth doing (Wall 1998, p. 128).

The autonomous person chooses and develops her own form of life. Wall notes that liberal perfectionists often endorse a number of distinct claims about autonomy, including that autonomy is intrinsically valuable or a central component of a good life.14 Nonetheless, all liberal perfectionists affirm the paramount value of autonomy.

Some liberal perfectionists argue that living autonomously requires religious freedom, but Wall counters that the strength of an autonomy-derived defense of toleration depends on how the value of autonomy is understood. According to Wall, autonomy can be valued either as an ideal or as a set of rights. An ideal of autonomy holds that a life is valuable when it is lived autonomously. Wall rejects the ideal-based defense of toleration since it only justifies tolerating a person’s choice of religious practice when respecting the choice to live in an “objectively worthless way … is more important … than getting him to pursue a more valuable way of life” (Wall 2003, p. 236). This is hard to show, given that living a valuable life appears to outweigh the value of living a “worthless” life. Yet liberal toleration requires tolerating objectively worthless lifestyles.

Wall thinks that a rights-based conception of autonomy stands a better chance of grounding religious toleration. If persons have a right to autonomy, then if “respecting [that] right is always more important than promoting a person’s good, autonomy-derived toleration will extend even to choices by persons to destroy … their lives” (Wall 2003, p. 236). Similarly, autonomy-derived toleration will extend to choices to practice disvaluable religions. If the most appropriate way to value autonomy is to treat it as a right, the value of autonomy (of religious choice) may trump our reason to promote a person’s good. Liberal perfectionism claims that autonomy is a great good; if so, autonomy’s immense value can explain why we have reason to refrain from forcibly promoting others’ good. Consequently, liberal perfectionism will vindicate a right to choose significant disvalue for oneself, including disvalue associated with religious practice.

The rights-based view also faces significant challenges. First, it must assume that autonomy rights reliably trump promoting the good, which is “not easy to defend” in principle (Wall 2003, p. 236). One can easily imagine an autonomy-restricting law that produces, for instance, more good as a result. If the state regulates or suppresses autonomy rights, it could generate more good for individuals who would have otherwise chosen great disvalue. The rights-based account also cannot protect those who engage in non-autonomous practices, despite the fact that truly liberal toleration requires tolerating many such cases. After all, if individuals were barred from choosing non-autonomous forms of life, most world religions would come under the suspicion of the liberal perfectionist state (Wall 2003, p. 237). Without such a right, the liberal perfectionist state would be permitted to persecute these religions if it could overcome pragmatic barriers. Certainly no reasonable member of a liberal society could accept this.

3.2 Comparative Inferiority

Wall acknowledges that these concerns “raise a few doubts” about an autonomy-based defense of toleration (Wall 2003, p. 235). There is reason, however, to have more than a few doubts. A perfectionist defense of toleration must justify strong rights of religious freedom to rebut the case for religious intolerance. Wall employs reasons to respect the good of autonomy to ground these duties; but these reasons are too weak.

Let’s begin by stating the criticism from Comparative Inferiority informally. Comparative Inferiority holds that for any feasible tolerant code there is a feasible intolerant code that is rationally preferable to it. A “code” is some set of laws, regulations, policies or regime types backed by force of law or some set of moral norms backed by moral ostracism, blame, punishment and so on. A code is “feasible” when implementing the code incurs manageable social cost. To succeed, Wall’s defense of toleration must locate a feasible tolerant code that is justified by reasons of respect grounded in the value of some candidate good; thus, the balance of reasons to respect and reasons to promote should justify implementing a code sufficiently tolerant to count as liberal. But, the criticism holds, we can always conceive of an intolerant code that promotes a degree of value that can override reasons to respect autonomy. If such a code is feasible, then the balance of reasons favors the intolerant code over the tolerant one. The criticism will generalize, thereby showing that there is a feasible, intolerant code on balance superior to each member of a substantial set of tolerant codes. In response, Wall must maintain that there is a range of feasible, tolerant codes for which there exists no feasible, intolerant code that promotes more overall value. This claim, we shall see, is hard to defend.

The formal argument deploys a number of variables and axioms that require clear definitions. First, define a degree of value promoted as a P-value; second, define that degree of value created when autonomy is respected as an R-value. Next, since Wall admits that reasons to respect certain goods can sometimes trump reasons to promote those goods, we shall treat him as acknowledging a trade-off ordering between P and R-values. Wall is committed to a trade-off ordering if he combines the possibility of a trade-off of P and R-values with the modest decision theory axioms of completeness, asymmetry and quasi-transitivity. The axioms merely assume (a) that some degrees of value are greater than others (and not vice versa) and (b) that the remaining quantities of value are equivalent for all P and R-values. With these three assumptions all P-values and R-values can be weakly ordered, that is ranked according to greater than, less than and indifference relations. The weak ordering is the trade-off ordering. It is fair to assume that the ranking is cardinal and not merely ordinal, which is to say that we can understand P and R values as degreed. Without this assumption, there is no way to gauge P and R-values save by comparison. Liberal perfectionists, however, think that value can be judged in absolute terms, as objects and states of affairs can be judged as disvaluable or valuable without reference to another good.15

To simplify, I will represent the trade-off ordering in abstract, formal terms, as representing some general value V. V symbolizes sums of R-values and P-values or the net value produced by some degree of respecting and promoting value. For our purposes, V picks out the degree of value produced when the candidate good, autonomy, is respected and when some general amount of value is promoted. V is increased when autonomy is respected and when good is promoted; conversely, V is decreased either when autonomy fails to be respected or when the good fails to be promoted or reduced. It is crucial that the argument not presuppose that the value of autonomy promoted and the value of autonomy respected are the same kind of value. V is merely a formal representation of both forms of value. If we equate the two forms of value, we will represent the value of respecting autonomy as a value to be promoted, which misrepresents Wall. Since it is wholly formal, V can be generated by any substantive theory that specifies a trade-off ordering. The formal argument characterizes both forms of value in terms of whatever substantive normative theory is used to determine trade-offs.

Indifference curve analysis can elucidate the argument.16 Define indifference as follows: X is indifferent to Y when X is not preferred to Y and Y is not preferred to X.17 Given a weak, cardinal ordering, there will be some value achieved by promoting (a P-value) and value achieved by respecting autonomy (an R-value) that are indifferent to one another in terms of V. Together, the R-value and P-value form a pair {R, P}. Figure 1 plots the indifference point. The R-value is represented on the X-axis and the P-value on the Y-axis.
Fig. 1

Respect–promote indifference point

This pair is indifferent to another PR pair when no reason shows one pair to be more choice-worthy than the other (in terms of V). The possibility of a trade-off and the decision theory axioms of completeness, asymmetry and quasi-transitivity entail that that there at least two RP-pairs are indifferent, which is to say that there is no reason to prefer one to the other. Since R-values and P-values are both weakly ordered and cardinal, an infinite number of points indifferent to {R, P} can be generated. If we increase the degree of value promoted, then by the trade-off ordering we can derive a second indifference point {R1, P1}. To illustrate, suppose that R is increased by some amount N. To deny the existence of a second indifference point is to deny that there exists any P-value that can offset the increased R-value required to generate a second indifference point. Given the trade-off ordering between P and R-values, this is impossible. Hence, the resulting range of indifference points compose an indifference curve represented in Fig. 2.18 As in Fig. 1, R-values are plotted on the X-axis and P-values on the Y-axis. Points {R, P} and {R1, P1} demarcate the line.
Fig. 2

Respect–promote indifference curve

Recall that each indifference point can be represented as comprising a discrete value V. Now stipulate that C is a legal and moral code that produces the degree of general value V represented by {R, P} and that C1 is the legal and moral code that produces the general value V represented by {R1, P1}. Also stipulate that C is an appropriately tolerant code due to its high degree of respect and stipulate that C1 is an intolerant code due to its low degree of respect but that nonetheless promotes great value. Since, by stipulation, {R, P} is indifferent to {R1, P1}, C is indifferent to C1. Consequently—and this is central—Wall has no rational reason to favor C over C1 and thus no reason to prefer a tolerant code over an intolerant code.

Suppose that there is another pair, {R2, P2}, that is superior to every point on the RP indifference curve. Since all points on the curve are indifferent, any point superior to {R, P} is thereby superior to {R1, P1} in terms of V. Thus, if legal and moral code C2 represents point {R2, P2} (see Fig. 3) then C2 is superior to C and C1 by stipulation. Let us further stipulate that C2 is unacceptably intolerant, like C1. If there is a feasible C2, then by Wall’s own lights, he must prefer the intolerant code C2 to the tolerant code C. The criticism from Comparative Inferiority is this: There is a feasible C2 such that Wall must rationally prefer it to C. Generalizing, the criticism claims that for any tolerant code of autonomy rights, there is a feasible intolerant code that is rationally preferable to it. To defend perfectionist toleration, Wall must argue that there exists a range of feasible, tolerant codes for which there is no rationally superior intolerant code. That is, he must argue that there are feasible, tolerant codes like C such that no feasible code Cx promotes a high enough degree of V to justify overriding reasons to respect autonomy. Let us represent the foregoing in Fig. 3.
Fig. 3

Feasible, intolerant codes

Point {R2, P2} is now represented. The solid vertical line represents the divide between tolerant and intolerant codes. All intolerant codes are located to the left of the tolerance line and all tolerant codes to the right. Assume that the line curves at extremely high degrees of P-value as promoting such a high degree of value will, at some point, justify a smaller degree of respect. I assume this because Wall does not hold that respect has absolute priority over promotion of value. The shaded area represents the set of feasible RP values represented by the set {Rx, Px} that require intolerant codes.

When the argument is put in formal terms, the weakness of Wall’s defense should become clear. It is implausible to maintain, as Wall must, that the entire set {Rx, Px} (represented by the shaded area) is empty. This is to claim there are feasible tolerant codes (like C) for which there exists no feasible intolerant code Cx (representing the set {Rx, Px} that generates greater value. Indeed, there is positive reason to think the converse. Religious persecutions have often been successful, so restraining a disvaluable religion to promote value seems obviously feasible. This is particularly obvious if we distinguish between groups. Imagine a religion that has pernicious social effects. Suppressing or restraining that religion to the point of intolerance could easily increase the amount of value in that society as a whole even if it reduces value for its members. Moreover, even if Wall can maintain that there are no feasible, intolerant codes that are of greater general value than feasible, tolerant codes, the argument is clearly much more tendentious than it initially appears.

3.3 Pragmatic Inconclusiveness

Wall should counter that religious intolerance has great pragmatic costs thus casting doubt on the claim that there are feasible, intolerant codes. It is not implausible to think that liberal citizens and states may be incompetent to eradicate only value-reducing religions. Terence Cuneo speculates that the state might be “incompetent to discern whether the valuable aspects of participation in certain religious traditions are defeated by the bad aspects of such participation” (Cuneo 2005, p. 126). In these cases, the costs of suppression would exceed the benefits. Value-reducing religions will be protected only because the state is not smart or competent enough to separate the good religions from the bad (or the good aspects of a religion from its bad aspects). If Wall can show that intolerant codes are often infeasible then perhaps the set of feasible, intolerant codes is small and hard to identify.

But again, religious persecution often succeeds. From the eleventh through the fifteenth century, the Roman Catholic Church maintained supreme religious power in Western Europe through the aid of local kings and the Holy Roman Emperor; its size and power often made eliminating heretical movements a small matter. While early Christian heresies like Arianism often produced social conflict, the following heretical groups were successfully suppressed: Albigensians, Henricians, Waldensians, Taborites, Euchites, and Lollardists.19 Readers are probably unfamiliar with these groups because, while not inconsequential in their time, they were effectively wiped from history. Consider then that modern states are more powerful than the Medieval Church and Holy Roman Emperor combined. While primitive forms of communication helped these institutions control religious belief, the examples illustrate that pragmatic considerations will not always bar religious persecution. Accordingly, pragmatic arguments cannot justify an appropriate degree of toleration save on a case-by-case basis.

Wall has also raised his own concerns about pragmatic defenses of toleration. “[C]ost-based justifications” of toleration lead to worries that “[i]njunctions to weigh costs call to mind a scale on which different interests could be measured. But no such scale exists. Our judgments on these matters must be rough and impressionistic” (Wall 1998, p. 69). Weighing costs faces difficulties because commensuration relies on the “types of costs … involved” being comparable. If many values are incommensurable, then their associated conception of social costs will be too. Consequently, the state will lack a rational basis upon which to weigh the costs and benefits of persecution (Wall 1998, p. 70). While perfectionist values can be traded-off in principle, weighing these values requires vague and tendentious judgments. Making such judgments might not be worthwhile.

However, if making such judgments is so difficult, then Wall’s candidate good-based defense of toleration seems to fail immediately. The defense requires judging that reasons to respect goods like autonomy often override reasons to promote value. If cost-based judgments are necessarily “rough and impressionistic” then it is unclear how the necessary judgments could be made. Wall could reply that if such judgments cannot be made, then the perfectionist need not recommend moving from a tolerant code to an intolerant one. However, while pragmatic considerations might require more tolerance, they might also require less. In the same way that pragmatic considerations may give us reason to stick with a tolerant code rather than moving to an intolerant one, so pragmatic considerations could give us reason to stick with an intolerant code rather than moving to a tolerant one. The argument cuts both ways. So suppose a society practices feasible, tolerant code C and is considering whether to practice intolerant code C2 on the grounds that C2 promotes enough value to justify reducing respect for the autonomy of religious citizens. Pragmatic considerations could be cited against making such a move. Conversely, if a society practices C2 and considers moving to some tolerant code C3, pragmatic considerations could also be cited against this move. Wall should be unsatisfied with a defense of toleration that only justifies practicing toleration when a society is already practicing it. Liberal perfectionism must identify reasons why societies with intolerant codes should adopt tolerant ones.

3.4 Heteronomy

The criticism from Heteronomy holds that pragmatic defenses of perfectionist toleration cannot justify toleration for the right reasons. Wall believes that citizens of liberal societies have reason to consciously respect the religious practices of others, including those whose religions generate net disvalue. Again, tolerant citizens are “morally motivated” to permit religious practices they regard as disvaluable (Wall 1998, p. 64). If Alice only has religious freedom because her fellow citizens cannot effectively persecute her due to pragmatic costs, she is not truly tolerated. It is cold comfort to a member of a potentially persecuted religion that she is allowed to practice her faith only because it is beyond her society’s ability to suppress. Surely no member of a liberal society should have to fear religious persecution should the benefits of persecuting her outweigh the costs. The criticism from Heteronomy seems powerful, particularly since Wall builds his account of toleration around respect for the right reasons. Since only pragmatic considerations can bar the state from persecution, the criticism from Heteronomy shows that perfectionist toleration is inadequate. Toleration must come from the right motives, yet toleration for merely pragmatic reasons cannot.

The criticisms from Comparative Inferiority, Pragmatic Inconclusiveness and Heteronomy build upon one another. Pragmatic Inconclusiveness shows that pragmatic arguments against the feasibility of intolerant codes fail and so cannot block the criticism from Comparative Inferiority. Heteronomy shows that even if pragmatic arguments were successful, they would fail to ground the appropriate attitude of respect. Taken together, these criticisms make a strong case that autonomy is an inadequate candidate interest; thus, autonomy-based defenses of perfectionist toleration fail.

3.5 Membership

Wall also defends perfectionist toleration by appealing to the good of membership. A citizen has membership or standing in political society “if and only if given his social identity, he can rationally identify with its governing institutions” (Wall 2005, p. 294). Governing institutions are the “main legal and political institutions of the state.” Rational identification requires justifiably believing that one’s institutions are “serving one’s interests and [are] treating one as no less important than other citizens” (Wall 2005, p. 295). Wall illustrates with the case of an anarchist who refuses to identify with his well-governed state. If the state serves his interests, he may have standing within it because his opposition is irrational.20 On Wall’s view, a person has a rational belief “if his belief is properly responsive to the evidence available to him and if he has made an adequate attempt to gather, or become aware of, evidence relevant to the truth or falsity of the belief” (Wall 2005). This is a partially externalist conception of rationality: to be rational is to be responsive to the best reasons that there are. However, requiring awareness adds an internalist element. Wall stresses that members of false religions can rationally identify with their state; their beliefs in the truth of their religion can subjectively justified despite being incorrect, as rational identification involves the ability to access reasons to identify with one’s institutions. If the externalist and internalist aspects of rational identification obtain, citizens can achieve the good of membership.

Aspects of social identity help individuals to define themselves: “a person’s ethnicity, political identification, and sex, as well as his religion, are prime examples of facts that typically have this kind of salience” (Wall 2005, p. 296). An individual’s social identity is determined by how her community perceives her, or how her identity is socially comprised; this is because “it helps us understand when it is and is not rational for a person to identify with the governing institutions of his society.” Hence, social identity marks out interests that justify rational identification; minor interests do not make rational identification valuable. An individual must also have some conception of her social identity before she can determine whether the state promotes her interests.

Membership makes religious persecution problematic, for “[w]hen the state promotes religion, either coercively or non-coercively, it risks making it impossible for some citizens to rationally identify with the governing institutions of their society” (Wall 2005, p. 297). Membership is a perfectionist good because it is bad for persons to be alienated from their governing institutions. Hence, reasons to respect membership are not unlike reasons to respect autonomy: to be enjoyed, both perfectionist goods require religious liberty. By promoting a particular religion, the state will sacrifice membership by restricting paths to achieving it.

Membership has an advantage with respect to autonomy in that it requires rational identification to be enjoyed. Wall can argue, then, that a state policy of religious toleration must be stable for citizens to engage in rational identification with their institutions. If the state regularly reassesses its policy of religious liberty, then members of many religions should not identify with their governing institutions due to reasonable concerns that their religion could fall out of favor. In this case, the state’s disposition to persecute will prevent citizens from achieving a great good.

Membership also has a disadvantage with respect to autonomy. Liberal perfectionists rank autonomy among the highest goods. While it may not outweigh basic needs such as food or healthcare, autonomy is crucial to living well. Accordingly, the state’s duty to promote other goods will not easily override its duty to respect autonomy. Membership lacks this centrality.21 While autonomy is plausibly central to flourishing, membership seems less important. If a citizen has food, healthcare, shelter, clothing, professional accomplishments and autonomy, whether she has membership seems not to matter much. But if membership is less valuable than autonomy, it cannot require sufficient respect to override the promotion of other goods. Hence if autonomy fails to ground toleration, membership will as well.

Furthermore, autonomy and membership can conflict. Since autonomy has greater centrality, it should reliably trump membership in a conflict. To see this, suppose that autonomy requires restricting religious practices. If the state complies, citizens will often be deprived of membership. If the perfectionist state has a duty to ensure that conservative Protestants live autonomously and decides that their faith practices discourage autonomy, the state may reasonably conclude that it has a duty to restrict those practices. If a conservative Protestant criticizes the stricture because it reduces membership, the argument would fail due to autonomy’s greater weight.

The criticisms from Comparative Inferiority, Pragmatic Inconclusiveness and Heteronomy threaten a membership-based defense of toleration. Comparative Inferiority holds that there always exists a feasible, rationally preferable intolerant code for any feasible, tolerant code. Presumably reasons to promote value can conflict with reasons to respect membership, just as they can conflict with reasons to respect autonomy. If so, membership-based defenses of perfectionist toleration are threatened. Pragmatic Inconclusiveness holds that, while pragmatic considerations may block moves from tolerant to intolerant codes, they also provide strong reason against moving from intolerant codes to tolerant ones. The criticism holds for membership if it holds for autonomy. Finally, Heteronomy also undermines a membership-based defense of perfectionist toleration. Citizens must tolerate the religious practices of others for the right reasons. If the state only tolerates for pragmatic reasons, it does not respect the membership of citizens. Membership therefore appears to be an inadequate candidate good. The three criticisms leveled in Sect. 3 show that autonomy and membership are inadequate candidate goods. By implication, Wall’s first defense of perfectionist toleration fails. We turn now to Wall’s second defense.

4 The optimal societal ethic

The second argument for perfectionist toleration begins with the claim that tolerating disvaluable religious practices may be required when the disvalue and value of those practices are inextricably intertwined (Wall 2003, p. 238). If a religion produces value for its members but the value is unavoidably tied to significant disvalue, then a society can only secure the value by tolerating the disvalue. Thus, societies that secure value for their citizens should sometimes tolerate disvalue. For these cases, given that we have reason to respect the good, we have “reason to tolerate practices that we rightly disesteem” (Wall 2003, p. 241). In some situations, a “condition of respecting the good is a willingness to respect the bad.” By themselves, these claims establish only a limited scope for toleration since they only justify tolerating practices that are “inseparably intertwined with valuable objects”; individuals will thereby lack reason to tolerate disvaluable practices not so intertwined or practices whose disvalue exceeds their value (Wall 2003, p. 240).

Wall thinks that this case for toleration can be extended to disvaluable objects and practices if liberal perfectionists can “[explain] the reasonableness of indirectly respecting the good by complying with a societal ethic that, if generally observed by others, would result in maximal respect for the good in one’s society” (Wall 2003, p. 242). An “optimal societal ethic” or OSE can promote perfectionist values. This optimal societal ethic requires toleration even when toleration fails to promote value; for the disposition to tolerate will promote perfectionist values as a whole (Wall 2003, p. 243). In this way, Wall’s defense draws on familiar defenses of rule-consequentialism.22 Perfectionist values are best advanced when a society inculcates a code of toleration that optimally promotes perfectionist values. If the code is inculcated, citizens of a liberal perfectionist state will develop a disposition to respect disvaluable religions.23

Wall’s defense of toleration may appear to assign merely instrumental value to the disposition to tolerate disvaluable religions. Citizens who value the OSE merely instrumentally “will be inclined to depart from it when doing so looks to be a more effective means for respecting what they take to be correct values. This is a recipe for the disintegration of the OSE” (Wall 2003, p. 245).24 In response, Wall claims that citizens who internalize the OSE have noninstrumentalist reasons to tolerate since “the ethic will be more stable and more likely to function well if it is valued for its own sake” (Wall 2003, p. 246).25 So, citizens acquire strong noninstrumentalist reasons to internalize the OSE because the OSE can only function if it is valued noninstrumentally. They then acquire reasons to respect disvalue when they internalize the OSE since the OSE demands respect of this sort. We have reasons to respect disvaluable religious practices because we have reason to value the OSE and the OSE requires respecting disvaluable religious practices.

The reply is problematic because it effectively wishes noninstrumental reasons into existence. Valuing the OSE noninstrumentally may be required in order for the OSE to function but this fact cannot make it the case that there are noninstrumental reasons to value the OSE. No good acquires noninstrumental value because the good could only perform its function if it had such value. Wall’s explanation of the noninstrumental value of the OSE seems backward. If the OSE does not already generate noninstrumental reasons to value it and fails to perform its function if it is valued merely instrumentally, then perhaps the OSE is not very valuable.

In Wall’s second defense of toleration he does not try to justify the attitude of respect that he endorses elsewhere in his work. Instead, respect is taken to consist merely in the disposition not to repress practices one disesteems. However, as Wall has noted, those who resist such repression may be nonchalant or efficient persecutors. Or they might do so merely because they value their OSE. None of these considerations justify a tolerant attitude, as each one is compatible with an attitude of disrespect. Hence even if Wall’s defense of acting as if one were tolerant succeeds, he will still not have vindicated being tolerant on perfectionist grounds.

Wall may reply as follows. Recall that Wall only requires an attitude of respect to tolerate persons rather than the good. So, he could first argue that the OSE provides a reason to have an attitude of respect towards the people who engage in disvaluable religious practices because such attitudes are required to comply with the OSE. Perhaps if the OSE did not require an attitude of respect it would be less effective, since motivating people out of nonchalance, pragmatic considerations and reverence for the OSE may be insufficient. Second, Wall could maintain that the OSE provides a reason to respect disvaluable religious practices in that it provides reasons not to hinder or suppress them. Third, when confronted with reasons not to respect these religious practices, citizens will judge them overridden by the noninstrumental value of the OSE and the general perfectionist value the OSE promotes. Thus, citizens will (i) have reason to respect disesteemed religious practices and (ii) have a justified attitude of respect. The OSE defense of perfectionist toleration can thereby defend genuine toleration.

Even so, the OSE does not justify toleration as Wall has defined it. For Wall, citizens are only tolerant when they are morally motivated to refrain from suppressing practices they disesteem. This moral motivation should be rooted in the recognition of reasons to respect persons since the motivation includes “the recognition that others are entitled to respect”; if Reba thinks she is permitted “to engage in some type of conduct, then this provides a reason … for the tolerant person to let [Reba] do it” (Wall 1998, pp. 64–65).26 The tolerant person is one who recognizes that it is the fact that X wishes to do Y which provides her with a reason to let X do Y. The attitude of respect is in turn justified by that recognition. But the attitude of respect described in the previous paragraph is not so justified. Instead, the attitude of respect required to tolerate is based on the reasons of respect grounded by the OSE. Yet the OSE only gives us reason not to suppress disvaluable practices because the general practice of restraint promotes perfectionist value. So, it appears the sorts of reasons of respect grounded by the OSE do not justify the attitude of respect Wall wants. The attitude of respect required to be a tolerant person is grounded in the recognition of reasons provided by the choices of others, not the recognition of reasons provided by the need to help the OSE promote perfectionist value. The OSE only gives us reason not to suppress disvaluable practices; it gives us no reason to tolerate disvaluable practices because those engaged in the practice are worthy of respect. On the OSE defense, therefore, citizens do not tolerate for the right sorts of reasons of respect. The OSE therefore cannot justify that right reasons to truly tolerate. Thus the second defense of toleration fails.

5 Conclusion

Wall’s liberal perfectionism cannot justify liberal religious toleration. Since Wall’s case for perfectionist toleration is among the most sophisticated, the fate of perfectionist toleration seems tied to Wall’s defense. What’s worse, if perfectionism cannot justify toleration, the possibility of a liberal perfectionism is threatened. A liberal political theory must embrace religious toleration to be worth its salt; the fight for religious toleration, after all, is the progenitor of the liberal tradition. If liberal perfectionism cannot adequately justify the original liberal liberty then how liberal can it be?

Even so, the above arguments are not intended to decisively refute liberal perfectionism. Instead, they challenge perfectionists like Wall to develop an account of normative reasons that can justify religious toleration. So far, Wall’s liberal perfectionism lacks an adequate account of reasons of respect. As long as Wall accepts a trade-off between reasons of respect and reasons to promote value, it is unclear how his first defense can succeed against the criticisms from Comparative Inferiority, Pragmatic Inconclusiveness and Heteronomy. The second defense has two problems: first, there are no reasons to value the OSE sufficient to make it work and second, the OSE cannot justify the right attitude of respect for disvaluable religious practices.

I will conclude with some proposals for how to emend Wall’s account of reasons of respect to account for these criticisms. The first defense can succeed if Wall explains why completeness, asymmetry or quasi-transitivity fail to apply to the trade-offs between reasons to respect some candidate good and reasons to promote value generally.27 For instance, Wall could deny quasi-transitivity by arguing that the trade-offs between reasons of respect and reasons to promote are contextual. In some contexts, reasons to respect a candidate good like autonomy will outweigh reasons to promote value, but in other contexts the same reasons might be ordered in reverse. For instance, Wall could argue that in cases of state policy-making, reasons of respect are sufficiently strong to outweigh reasons to promote under a wide range of circumstances even if, say, in other historical contexts reasons of respect were insufficiently strong. Or Wall could argue that there is a high degree of metaphysical indeterminacy about how to rank reasons of respect alongside reasons to promote. Since the reasons are of sufficiently distinct types, it may be the case that there is no fact of the matter about how to generate an ordering. Wall could then argue that reasons of respect can still justify a high degree of toleration even if they cannot be weakly ordered vis-à-vis reasons to promote. The challenge for the liberal perfectionist is to motivate the falsity of quasi-transitivity. Any such argument will involve original theoretical developments in liberal perfectionist accounts of reasons of respect.

A second way to repair the first defense is to give a pluralistic account of respect-worthy candidate goods. The main criticism of autonomy and membership-based defenses of toleration is that liberal perfectionists should prefer intolerant codes that promote more value over tolerant codes that promote less. If Wall could show that a large number of goods are sufficiently respect-worthy he could motivate the claim that reasons of respect are sufficiently strong to override reasons to promote value generally. He could then argue that respecting these goods as a whole requires a truly liberal degree of religious toleration. The challenge for the liberal perfectionist in this case is three-fold: she must (i) identify each of these goods, (ii) show that the sum of reasons to respect these goods often outweigh reasons to promote value generally and (iii) discover a method of making the trade-offs needed to judge whether reasons of respect override reasons to promote.

One method of avoiding inevitably messy trade-off judgments is to opt, as Rawls does, for lexical priority of some kinds of value over others (Rawls 1971, pp. 36–40). For instance, liberal perfectionists could claim that reasons to respect persons have lexical priority over reasons to promote perfectionist value generally. Some may maintain that to adopt such a degree of priority abandons liberal perfectionism, but it need not. After all, Wall has already introduced reasons of respect into liberal perfectionism. If reasons of respect need not prevent the promotion of value generally, a theory that contains them could embrace some kinds of lexical priority. The challenges to a lexical priority view are fairly obvious, however. Wall has already embraced a trade-off principle between reasons to respect and reasons to promote; he would have to abandon the principle in at least some cases. Further, lexical priority is such an extreme form of priority that it will be difficult to defend. Wall must also show that a lexical priority view does not bar the state from promoting sound ideals of flourishing.

The second defense can be salvaged if Wall can provide noninstrumental reasons to value the OSE apart from the fact that the OSE needs such reasons to function. While Wall’s claim that the OSE has noninstrumental value is justified on poor grounds, there is no reason to think that Wall cannot find other reasons to show that the OSE has noninstrumental value. It is not clear what such reasons would be, so the perfectionist must identify the required reasons. She must then show that reasons to noninstrumentally value the OSE override reasons to suppress disvaluable religious practices.

Another interesting move may save Wall’s second defense. To justify toleration, Wall needs more than mere reasons of respect, as even these reasons must be of the right sort. Reasons of respect grounded in the OSE are not the sort that can justify having a tolerant attitude. Such respect is required by the OSE and not by the recognition that others merit respect. Wall needs a further distinction between types of reasons of respect that may be delineated by their ground. The OSE defense provides reasons of respect generated by the recognition of perfectionist value but only through the mediating framework provided by the OSE. Other reasons of respect may be provided without mediation but simply by the recognition of the value of persons. Perhaps Wall can locate reasons of respect grounded in the unmediated recognition of the value of persons which could ground the appropriate attitude of respect. Even if the mediation relation cannot correctly classify different types of reasons of respect, some other relation may be up for the job.

Again, the challenges posed here are meant to push liberal perfectionists to develop more sophisticated accounts of reasons of respect. One powerful reason to do so is that Wall’s rather developed account of reasons of respect cannot justify religious toleration. To be truly liberal, however, liberal perfectionism must justify religious liberty. The failure of Wall’s arguments suggests that a defense of religious toleration is a significant challenge to liberal perfectionism.


  1. 1.

    Rawls makes toleration the conceptual core of liberalism: ‘political liberalism applies the principle of toleration to philosophy itself’ (Rawls 2005).

  2. 2.

    I appeal to a broad conception of harm below as reductions in the achievement or enjoyment of perfectionist goods. The harms with which I am concerned can be committed without wrongdoing. They may also include harms to self.

  3. 3.

    Wall uses “the term ‘repress’ very broadly to stand for repress, persecute, harm, offend, insult, etc.” (Wall 1998, pp. 63–64).

  4. 4.

    I will assume that Wall’s “moral motivation” requires an attitude of respect throughout the paper.

  5. 5.

    See King (1998, pp. 21–72) for a classic account of toleration.

  6. 6.

    Wall notes that this account is extremely general because it does not specify a conception of a valuable and worthwhile life.

  7. 7.

    In the same passage, Wall argues that perfectionists can be value pluralists and that there will be no direct connection from sound ideals of the good to the structure of political institutions without substantial empirical work.

  8. 8.

    Wall also maintains that he will not defend this claim about the independence of reasons to promote value and reasons to respect it.

  9. 9.


  10. 10.

    The question of trade-offs still arises even if respect for a candidate good promotes perfectionist value to a great degree. In the first defense, Wall does not argue that respect for certain values tends to promote more value than disrespect on the whole.

  11. 11.

    Wall’s conception of respect for persons is defended in Wall (1998); the defense of respecting the good is located in Wall (2003).

  12. 12.

    Wall does not give an “autonomy-derived” defense of toleration and even suggests difficulties for it. However, Wall thinks it contains some “measure of truth” (Wall 2003, p. 235). Reviewing problems with the good will prove illuminating nonetheless.

  13. 13.

    For Raz’s understanding of autonomy, see Raz (1986, p. 155). For arguments that attempt to establish a connection between autonomy and toleration, see Raz (1988).

  14. 14.

    Wall makes a detailed prima facie case for six distinct claims about the value of autonomy (Wall 1998, pp. 144–161).

  15. 15.

    The points I make here are essentially identical to the ones Rawls makes about the application of decision theory to rational intuitionism (Rawls 1971, pp. 30–36).

  16. 16.

    Indifference curve analysis is emphatically not the same as cost–benefit analysis. Indifference curve analysis can be employed at a completely formal level to clarify any number of problems of rational choice.

  17. 17.

    Here I follow Stanley Benn’s indifference curve analysis of choosing between incommensurable goods (Benn 1988, pp. 43–64).

  18. 18.

    Following standard decision theory, we may also generate multiple indifference curves for different values of R and P. Higher and lower aggregate values of R and P will also admit of indifference points, and thus their own indifference curves.

  19. 19.

    Some qualifications arise. The Waldensians survive in small numbers and the doctrines of some other groups merged into Protestantism.

  20. 20.

    Wall often speaks of interests and of values. I shall understand serving human interests as promoting value for humans.

  21. 21.

    Sometimes autonomy requires membership, but not always.

  22. 22.

    Wall’s strategy is surprisingly similar to Brad Hooker’s strategy which focuses on evaluating the value of inculcating a code with norms of respect that promotes well-being (Hooker 2000, p. 32).

  23. 23.

    One might worry that we should inculcate a code with exceptions for disvaluable religions, but Wall notes that “[t]oo many exceptions will undermine the ethic” (Wall 2003, p. 245).

  24. 24.

    Note that the OSE is valued as a set of dispositions not as a set of propositions.

  25. 25.

    Wall grants that this response will not provide everyone with a non-instrumental reason to tolerate. But his case does not require showing “that all persons have reason to value toleration for its own sake.” Instead, “instrumental considerations as well as those derived from indifference or skepticism may be sufficient to motivate support for the ethic” (p. 247).

  26. 26.

    Emphasis mine.

  27. 27.

    Larry Temkin has denied that the axiom of transitivity applies to practical reasoning (Temkin 1996).



I am grateful to a number of people who provided me with feedback on this article, including Nathan Ballantyne, Thomas Christiano, Chris Freiman, Jerry Gaus, Uriah Kriegel, Mark LeBar, Jonathan Quong, Steven Wall and members of a 2009 University of Arizona writing workshop, along with an anonymous reviewer. I especially thank Gregg Keithley for pressing me to think hard about the relationship between perfectionism and toleration and to Aira Burkhart for graphic illustration.


  1. Benn, S. I. (1988). A theory of freedom. New York, NY: Cambridge University Press.CrossRefGoogle Scholar
  2. Christiano, T. (2008). Does religious toleration make any sense? In L. Thomas (Ed.), Contemporary debates in social philosophy (pp. 171–191). Oxford: Blackwell.Google Scholar
  3. Cuneo, T. (2005). Can a natural law theorist justify religious civil liberties? In T. Cuneo (Ed.), Religion in the liberal polity (pp. 108–130). Notre Dame, IN: University of Notre Dame Press.Google Scholar
  4. Gaus, G. (2003). Liberal neutrality: A radical and compelling principle. In S. Wall & G. Klosko (Eds.), Perfectionism and neutrality: Essays in liberal theory (pp. 137–166). Lanham, MD: Rowman and Littlefield.Google Scholar
  5. Hooker, B. (2000). Ideal code, real world. Oxford: Oxford University Press.Google Scholar
  6. King, P. (1998). Toleration. London: Frank Cass Publishers.Google Scholar
  7. Locke, J. (2009). A letter concerning toleration: Humbly submitted. New York: Classic Books America.Google Scholar
  8. Mill, J. S. (1978). On liberty. Indianapolis: Hackett Publishing Company.Google Scholar
  9. Rawls, J. (1971). A theory of justice. New York, NY: Oxford University Press.Google Scholar
  10. Rawls, J. (2005). Political liberalism. New York, NY: Columbia University Press.Google Scholar
  11. Raz, J. (1986). The morality of freedom. Oxford: Oxford University Press.Google Scholar
  12. Raz, J. (1988). Autonomy, toleration and the harm principle. In S. Mendus (Ed.), Justifying toleration: Conceptual and historical perspectives (pp. 155–177). Cambridge: Cambridge University Press.CrossRefGoogle Scholar
  13. Smith, S. D. (2008). Toleration and liberal commitments. In M. Williams & J. Waldron (Eds.), Nomos xlviii: Toleration and its limits (pp. 243–280). New York: New York University Press.Google Scholar
  14. Temkin, L. (1996). A continuum argument for intransitivity. Philosophy & Public Affairs, 25(3), 175–210.CrossRefGoogle Scholar
  15. Wall, S. (1998). Liberalism, perfectionism and restraint. New York, NY: Cambridge University Press.CrossRefGoogle Scholar
  16. Wall, S. (2003). The structure of perfectionist toleration. In S. Wall & G. Klosko (Eds.), Perfectionism and neutrality: Essays in liberal theory (pp. 231–256). Lanham, MD: Rowman and Littlefield.Google Scholar
  17. Wall, S. (2005). Perfectionism, public reason and religious accommodation. Social Theory and Practice, 31(2), 281–304.CrossRefGoogle Scholar
  18. Williams, B. (1996). Toleration: An impossible virtue? In D. Heyd (Ed.), Toleration: An elusive virtue (pp. 18–27). Princeton, NJ: Princeton University Press.Google Scholar

Copyright information

© Springer Science+Business Media B.V. 2011

Authors and Affiliations

  1. 1.Philosophy DepartmentBowling Green State UniversityProvidenceUSA

Personalised recommendations