Les ateliers de l'éthique The Ethics Forum Perfectionism, Economic (Dis)Incentives, and Political Coercion

May a government attempt to improve the lives of its citizens by promoting the activities it deems valuable and discouraging those it disvalues? May it engage in such a practice even when doing so is not a requirement of justice in some strict sense, and even when the judgments of value and disvalue in question are likely to be subject to controversy among its citizens? These questions have long stood at the center of debates between political perfectionists and political neutralists. In what follows I address a prominent cluster of arguments against political perfectionism—namely, arguments that focus on the coercive dimensions of state action


ANTI-PERFECTIONIST ARGUMENTS FROM COERCION
Perfectionist moral theories are theories that ground moral reasoning at least partly in some objective account of what is good or valuable for human beings.Such accounts may be based on a conception of human nature, although that need not be the case: they may simply rely on the identification of certain human activities, capacities, attributes or pursuits as valuable, without commitment to a more robust theory of human nature or human essence. 1 Perfectionist political theories extend perfectionist moral reasoning to the sphere of politics, maintaining that the entire expanse of moral considerationsthose rooted in objectivist accounts of value included-may, indeed should, guide all action, whether political or nonpolitical.
What anti-perfectionists, or neutralists (I will use the terms interchangeably), typically claim, by contrast, is that certain kinds of moral consideration ought not to guide, or ought to be excluded from, the justification of certain types of action.Anti-perfectionism does not, however, require that all moral considerations be excluded from the justification of the problematic types of action; nor does it require that the problematic moral considerations be excluded from the justification of all action.Any anti-perfectionist argument must therefore begin by answering two basic sets of questions: (1) To which class of considerations does the exclusionary restriction apply, and why to that class alone?(2) To whose actions, or to which sphere(s) of action, does the restriction apply, and why only to those?
The typical neutralist answer to the first question is, roughly, that the restriction applies to moral considerations rooted in conceptions of the good-in the thick or comprehensive accounts of value provided by perfectionist moral theories-but not to the rest of morality.The typical answer to the second question is that the restriction applies to state or government action (or to political action), but not to individuals acting either privately or nonpolitically.In other words, doctrines of neutrality challenge the move from moral to political perfectionism, demanding that states or governments should not appeal to perfectionist conceptions of value when justifying their policies. 2ome of the difficulties facing neutralist doctrines are connected to the first of our two sets of questions.In particular, there is the familiar worry that there might be no defensible way of dividing the realm of moral considerations into two parts, such that one part may serve as the justificatory basis of political or state action while the other may not.It has been argued, for example, that even the likeliest candidates for membership in the part of morality neutralists deem appropriate for political reasoning-e.g., requirements of justicecannot but rely for their force and content on the other part of morality, on particular conceptions of the good, and that doctrines of neutrality are therefore doomed to incoherence. 3hese, however, are not the worries I wish to discuss here.(Let me just note that I do not find them as obviously damaging to the neutralist case as some have taken them to be.) 4 Instead, I want to address our second set of questions: To whose actions, or to which sphere(s) of action, should anti-perfectionist restrictions apply, and why?Why is it political action, or state action, that merits special concern?If we have compelling reasons to encourage certain conceptions of the good and discourage others, why may we not do so politically, through the state, as well as nonpolitically?Unless anti-perfectionists can argue for some morally relevant distinction between political and nonpolitical action-unless they can drive a wedge between moral perfectionism and political perfectionism-it remains unclear why conceptions of the good should be excluded from political reasoning.
Not all of the by now familiar arguments for state neutrality meet this very basic requirement: arguments from moral skepticism, for example, do not. 5Much better at identifying the requisite distinction, however, are arguments that point to the coercive character of state action.
Anti-perfectionist arguments from coercion come in a variety of forms.They vary in their conceptions of what it is that we should worry about when we worry about coercion.(Ought we worry about freedom in some thin sense of arbitrary choice, of not being subjected to the will of others, and so forth?Or should we protect some thicker ideal of autonomy as reasoned self-direction?)They can also vary in their conceptions of how freedom-and coercion-related considerations function within morality.(Can some loss of freedom be offset by other sorts of gain?Or do considerations of freedom enjoy some more privileged status?) Concerns about coercion are often connected to the idea that governments may act only on reasons that the governed-at least the reasonable among them-share, or could share, or cannot reasonably reject.Such concerns are sometimes associated with the Kantian charge that coercing people for reasons they do not or could not share treats them merely as means rather than also as ends; with the Lockean worry that attempts to 'coerce people into the good' are self-defeating; or the Millian idea that uncoerced experimentation with various conceptions of the good is instrumental to finding out what is in fact good. 6f course, hardly anyone thinks that all coercive acts are unjustified: some actions ought to be prohibited, and the prohibitions coercively enforced.The apprehension, however, is that to ground such prohibitions in perfectionist reasoning is to expand the use of coercion to matters that ought to remain free of coercive regulation.
But although anti-perfectionists so regularly base their arguments on the coercive aspects of state action, what they typically argue for excluding are not just coercive legal prohibitions grounded in perfectionist reasoning, but any kind of political action so grounded.In doing so, they often ignore the fact that perfectionist action need not be the "coercive imposition of a style of life": Conferring honours on creative and performing artists, giving grants or loans to people who start community centres, taxing one kind of leisure activity (e.g., hunting) more heavily than others, are all cases in which political action in pursuit of conceptions of the good falls far short of the threatening popular image of imprisoning people who follow their religion, express their views in public, grow long hair, or consume harmless drugs. 7 would seem, then, that arguments from coercion do not suffice for a thoroughgoing, morally principled rejection of political perfectionism. 8This is underscored by the fact that economic (dis)incentives-subsidies, taxation of specific activities-as well as other noncoercive policies are not marginal components of perfectionist practice, but rather commonplace instruments of perfectionist state policy.Now, to point to the fact that perfectionist policies need not be coercive is not yet to discredit all coercion-based arguments for state neutrality: it is only to show that coercion-based arguments fail to support strong versions of neutralism, according to which no government policies may be justified on perfectionist grounds.It does not show that coercion-based arguments could not support some weaker principle of neutrality which applied only to coercive state action. 9(To reject the weaker principle, one would have to argue that perfectionist coercion can sometimes be justified.)But perhaps even 'strong' neutralists could respond to the above perfectionist challenge.To argue that coercion-based arguments support even 'strong' neutralism, neutralists would have to resort to one or more of the following strategies: (1) They could question whether the very distinction between coercive and noncoercive state action is plausible (the argument from ultimate coercion).(2) They could argue that even if such a distinction can be made, allegedly noncoercive policies such as economic (dis)incentives do in fact coerce people into accepting (or rejecting) conceptions of the good, or into acting (or refraining from acting) in accordance with such conceptions (the argument from direct coercion).(3) They could argue that even if perfectionist policies are not coercive, they are nevertheless manipulative, and that the same reasons that weigh against coercive state actions also weigh against manipulative ones (the argument from manipulation).Or perhaps (4) they could claim that however we classify policies such as subsidization or activity-specific taxation, what is objectionable about them is that they induce people to act for the 'wrong reasons', and that such influence, whether properly speaking 'manipulative' or not, similarly undermines people's autonomy (the argument from inappropriate motivation).My main aim in what follows is to respond to these challenges. 10I will conclude, however, with a few remarks on the implications of my arguments for the special case of government intervention in religious practice.

SHIFTING THE BURDEN OF ARGUMENT?
Before I do so, however, let me address one possible objection to the general argumentative strategy I employ here, which, to recap, is this: I have said that if we have reason to promote valuable goals, and if doing so by nonpolitical means is acceptable, anti-perfectionists need to explain why doing so by political means is not.To do so, they need to point to some relevant distinction between political and nonpolitical (or state and non-state) action.And if the distinguishing feature to which they point is the state's coercive character, they need to defend their view against the objection that many state actions of the sort perfectionists recommend are in fact noncoercive.
But an anti-perfectionist objection different from the ones just noted might be available.Instead of challenging the move from moral to political perfectionism, neutralists might try to block an earlier move within moral perfectionism.That is, neutralists might claim that what- ever the differences between political and nonpolitical action, in neither case do we have a reason to intervene in order to encourage others to adopt worthy ideals or discourage them from adopting unworthy ones.As Jeremy Waldron puts it: …the connection between [perfectionist] evaluation and perfectionist intervention is not nearly so straightforward.That an ideal is unworthy provides a person with a reason not to choose it as her ideal, but it is not at all clear that it provides others … with a reason to discourage her from choosing it. 11 It is unclear, however, what might justify such a view-unless, that is, it rested on a moral theory that was thoroughly egoistic, at least with respect to perfectionist goals. 12Without here arguing against egoistic versions of moral perfectionism, let me just note that apart from what I take to be their utter unattractiveness, they are also at odds with the very point of theories of political neutrality, which is to defend a certain relation between political morality and morality in general (one in which some moral considerations-appealing as they may be-are excluded from politics), and not to lay out the content of basic morality itself.
Let me also add that my claim concerning a reason to intervene should not be confused with the much stronger claim that the unworthiness of a person's ideals imposes on others a duty to discourage her from pursuing them.The stronger claim may be vulnerable to Waldron's above objection (with 'duty' substituted for 'reason').But my own initial claim sidesteps this difficulty by invoking only a reason, not a duty, to intervene.And so, the request that anti-perfectionists provide overriding considerations still stands.To the abovementioned attempts to provide such considerations I now turn.

THE ARGUMENT FROM ULTIMATE COERCION
The first anti-perfectionist response to the claim that perfectionist state action need not resort to coercive measures (and thus to the claim that the state's general coercive character fails to provide the requested overriding considerations), is to question whether it is at all meaningful to divide the means available to the state into coercive and noncoercive ones.Because all influence wielded by the state ultimately rests on the state's coercive power, this line of reasoning goes, all state actions are ultimately coercive, even when no overt coercion is involved: "The state may not show its guns when it takes [seemingly noncoercive] action.But ultimately it is what it is on account of its guns." 13ow seriously should we take such claims?Not very, I think.Many state actions and policies (taxation, military conscription, mandatory schooling) are coercive in nature; many others (designation of national holidays, granting of public honors) are not.(That actions of the latter type are sometimes funded by taxation, which is genuinely coercive, does not make them coercive: museum subsidies force no one to visit museums or to adopt the conceptions of value that such visits embody, even when the subsidies are funded by coercively collected taxes.)The idea that the state's considerable coercive power contaminates all state action is no more plausible than the Austinian view that all legal rules are commands backed by coercive threats.Like this long-rejected legal-philosophical position, the argument from ultimate coercion blurs distinctions precisely where distinctions ought to be drawn.
Whatever initial appeal the argument from ultimate coercion might have derives, I think, from the fact that there are cases in which superficially noncoercive recommendations and requests in fact function coercively.Requests we would otherwise treat as noncoercive can function coercively when issued by a superior-a supervisor, a teacher, an officer-to a subordinate.Such requests need not be communicated explicitly (or perhaps even insinuated) as coercive threats in order to function as ones.Arguably, to have a coercive effect, they need not even be intended as coercive threats, nor interpreted by their recipients as so intended.Instead, it might be enough that the recipients have reason to fear that failure to comply might result in some (perhaps yet uncontemplated) act of reprisal.(The idea that sexual harassment laws should indiscriminately prohibit all sexual communication between superiors and subordinates relies on precisely this sort of reasoning.) The question, then, is whether perfectionist state policies that are not coercive in any apparent or straightforward manner may nevertheless be coercive for the reasons just described.The type of case analogous to the ones just discussed would be one in which a government issued a recommendation without communicating in any way an intention to penalize non-compliers, and yet citizens had reason to fear that failure to comply might carry penalty after all.But should such clear instances of foul play form the basis of a principled position on the kinds of moral reasoning admissible in politics?Should we not rather consider cases in which the government does not violate its own publicly announced rules, i.e., cases in which the state refrains from using any guns it has not previously shown?Are there any principled, coercion-related reasons for rejecting political perfectionism even then?Not in any way that the argument from ultimate coercion can establish.

THE ARGUMENT FROM DIRECT COERCION
To challenge the claim that perfectionist state action need not be coercive, then, one must point to more than just the state's general power to coerce: one must point to some genuinely, directly coercive aspect of seemingly noncoercive policies.Since this argumentative strategy can only proceed piecemeal, one policy at a time, even those who find it promising should not expect it to yield a complete case against political perfectionism.But even partial success would be of consequence if the policies exposed as coercive were central enough instruments of perfectionist legislation.
One such instrument-the one most clearly susceptible to the present type of argument-is the taxation of specific activities that a government wants to discourage.Although activity-specific taxation falls far short of criminalizing the discouraged activities, its impact on individuals' choices, the argument goes, is nevertheless that of coercive threats: What does it mean to discourage an activity by imposing a tax?It means … that the state raises the costs that a person must incur if she wants to pursue the activity.… Now, certainly, altering the costs and payoffs of an activity looks like coercion.After all, what a threat does … is precisely to add an artificial cost to an activity.The threat, 'Your money or your life,' … adds the cost of death to the option of keeping one's money. 14 call this an argument from direct coercion, since the suggestion here is that activity-specific taxation directly coerces people to refrain from the taxed activities, and is not merely contaminated by coercion in a more roundabout way.) Is the challenge reasonable?The argument, as presented in the quote, equates activity-specific taxation and other (dis)incentives with the adding of 'artificial' costs; 15 the adding of 'artificial' costs with threats; and, implicitly, threats with coercion.But although each of the last two moves (in one version or another) has had its defenders in the literature on coercion and threats, by no means are they unquestionable.(Let us grant for now that the first move-the claim that government (dis)incentives add 'artificial' costs to the activities they target-is warranted.I will return to this point in the next section.) The first thing to note is that not every proposal to add costs to an option can plausibly be said to constitute a threat to those who want to pursue the option.Let us suppose, however, that this difficulty is addressed by the reference to 'artificial' costs, which implies the existence of some 'natural' or otherwise appropriate baseline of costs and benefits, deviation from which is necessary if an act is to count as a threat.Of course, we would then need to address the very difficult problem of where and how to fix the baseline.
Note, however, that we are able to circumvent this entire set of problems, since what concerns us here is coercion, not threats.The question whether or not the (dis)incentives under discussion are best conceptualized as threats-and thus the question of how to distinguish threats from cost-adding proposals that are not threats-would require our attention only if we accepted the idea that all threats were coercive.But this is false: on any promising account of coercion, whether or not a proposal is coercive depends on whether the (dis)incentives attached to it are of such magnitude as to apply to its recipient a degree of pressure sufficient to eliminate as a genuine option the activity that the proposal aims to discourage.A proposal fails to be coercive, then, if the attached (dis)incentives are not significant enough to apply the requisite degree of pressure.It is a question, of course, beyond what threshold of refusability an option should count as 'genuine'-as something that the person in question really could choose to accept or decline.(It is also a question whether we should think about the refusability of options in purely descriptive or at least partly in normative terms.)But whatever view we take, it does not affect my point.If what I've said so far is correct, then the question we should be asking is not whether perfectionist (dis)incentives count as threats, but simply whether they need be coercive-to which the plain answer is no.Any plausible view of the degree of pressure that a proposal need apply in order to count as coercive would not designate our normal subsidization and taxation practices-say, a $1 tax on a $10 activity-as applying coercive pressure on their addressees.

MANIPULATION AND INAPPROPRIATE MOTIVATION 5.1 Manipulation
There remains the possibility that trying to influence people's choices by altering the incentive structure of the options available to them might be objectionable for reasons other than coercion yet normatively very similar to it.One way of formulating the revised complaint is this: the decision to pursue an activity or to avoid it ought to be based on a certain type of consideration-namely, on considerations stemming from the activity's intrinsic value, from its intrinsic merits and demerits.'Artificially' to add other considerations to the deliberative balance is manipulative, the objection goes, even if not coercive.And manipulation, while not identical to coercion, is objectionable for the same reason coercion is: it diminishes or violates the autonomy of its victims.
How exactly might economic (dis)incentives be construed as manipulative?Here, for example, are a few ways in which Waldron spells this out.Activity-specific taxation, he writes, "is necessarily manipulative, for it influences a person's decision by distorting the individual's understanding of the merits of the choice."The same is said to apply to subsidies, which "give those who benefit from [them] a misleading and distorted picture of the real costs and benefits of engaging in the subsidized activity."These policies thus "interfere with the way people form their beliefs about value"; they undermine one's ability to "respond to value by choosing for reasons the person apprehends."Finally, by "messing with the options that one faces," manipulation undermines one's ability to choose among options "for the right reasons." 16anipulative intervention motivates people for the wrong reasons, then, and the above remarks focus on one technique in which this can be done, namely deception: manipulative acts can exert influence by concealing the appropriate reasons (reasons stemming from an option's real value), or by presenting inappropriate reasoning as appropriate.
But need perfectionist policies, if they are to succeed, distort people's perceptions of the genuine merits or of the "real costs and benefits" of an activity? 17The answer, I think, is straightforward: perfectionist policies need not have this epistemic effect.Although they could (if so designed) distort the information people have about the desirable and undesirable features of various options, they need not do so in order successfully to encourage or discourage the options.(Certainly, the epistemic concern can serve as no objection to measures-e.g., state-sponsored media campaigns-the entire point of which is to inform people of the value or disvalue of an option.)Not all manipulation relies on deception, however: one could be manipulated into choosing for the wrong reasons while not for a moment losing sight of the right ones.Manipulation, as we normally understand the term, can also consist in the exploitation of a person's weaknesses, with no epistemic component involved.To respond fully, then, to the charge that economic (dis)incentives are manipulative, we will have to consider whether (dis)incentives might manipulate in this second way. 18To make the idea clearer, we will need to spell out in a bit more detail in what exploiting someone's weakness consists (and how influencing someone in this manner differs from manipulation by deception, as well as from coercion and from rational persuasion).I see two ways of doing so: one focused on some peculiar facts about the manipulated party, the other on the degree of pressure that the manipulative action exerts.That is, (1) we can sometimes say that A is exploiting some weakness of B when A offers B an (dis)incentive which, though not normally irresistible (i.e., not irresistible to people in general), is irresistible to B due to some distinctive or idiosyncratic fact about, e.g., B's personality or psychological makeup.Or (2) perhaps we can sometimes say that A is trying to play on some weakness of B when A offers B an (dis)incentive which B is expected to find very hard to resist, although the degree of irresistibility is not quite of such magnitude as to completely undermine B's ability to make a genuine choice. 19his account is extremely sketchy, of course, and would need to be supplemented by further conditions in order to pick out manipulative actions only (and exclude, e.g., acts of coercion).And of course, we should expect the boundaries between coercion and weaknessexploiting manipulation to be somewhat vague.But it seems to me that with some fine-tuning, (1) idiosyncrasy and (2) degree of irresistibility are the two features we should focus on in order to carve out conceptual space for a class of manipulative actions that are neither deceptive (as other manipulative actions are) nor simply coercive.
It is hard to see, however, why either of these two features must be present in such perfectionist policies as subsidization or activityspecific taxation.The policies need not be designed to exploit any peculiar weaknesses of the people whose behavior they hope to influence.And the (dis)incentives involved need not be especially hard to resist.Although the point of such policies is to change the balance of costs and benefits attached to various options, such changes need not be dramatic: sufficiently mild (dis)incentives can avoid applying great (and perhaps morally excessive) pressure on those who strongly resist the encouragement or discouragement they offer, yet have the desired influence on many other people who are more vulnerable.

Inappropriate Motivation
The charge that perfectionist (dis)incentives are necessarily manipulative may be unfounded, but it rests on an even more basic worry which, in its most general form, has not been addressed yet, namely, the worry that perfectionist (dis)incentives might motivate people to act for the wrong reasons.That is, even if a (dis)incentive is not manipulative in any ordinary sense of the term-even if it is neither deceptive nor exploitative in the above-discussed sense-it might still motivate inappropriately, and that, goes the objection, is still sufficiently troubling.
The worry about inappropriate motivation-as formulated, for example, in some of the above excerpts from Waldron-depends on the persuasiveness of the distinction between an option's intrinsic merits (which provide the 'right' reasons for choosing it), on the one hand, and factors introduced to it 'artificially' or 'from the outside', on the other.Now, let us assume for a moment that the distinction is unproblematic.(I will return to this assumption in a moment.)Even so, it is unclear why considerations shaped by state action should fall on one side of the line, and considerations shaped by other social forces on the other.Why should incentives designed and administered by the state be considered any more extraneous to the value of an activity than incentives created and sustained by other, non-state agents?Consider a composer who needs to decide whether to devote her time to composing concert music or to recording advertisement jingles.One class of considerations in favor of the former option is that it is more aesthetically valuable, more personally fulfilling, more conducive to the full realization of one's talents, etc.; a different kind of consideration might be that the state offers generous grants to composers on the basis of artistic merit.Meanwhile, a consideration in favor of the latter option might be that advertising companies pay extremely well.Surely, if there is a line to be drawn here between two different groups of considerations, it should separate the first consideration from the latter two, not the second from the third.
Recall the condition that anti-perfectionist arguments must satisfy: they must explain why it is the state that ought to banish perfectionist considerations from the justification of its actions.But this is precisely what the argument from inappropriate motivation has failed to do so far: any restrictions it might justify would apply to state and to non-state action alike. 20 think, however, that we can do better than this sort of response (which concedes that (dis)incentives 'motivate for the wrong reasons' and merely adds that it is no more objectionable when the state uses them than when non-state agents do).To see how, let us return to the distinction between the 'intrinsic' and the 'extraneous' advantages and disadvantages of an option.
One possible response to the 'inappropriate motivation' objection is to deny that the distinction can (or can in most cases) be madeto insist that the notion of "intrinsic merit … independent of social conditions in most cases is hard to sustain" 21 -and thus to undermine the basis for a demarcation between 'appropriate' and 'inappropriate' motivation.
This seems too strong, however.Returning to our composer example, I do not see a difficulty in distinguishing, as I have, between two classes of considerations on which the composer's choice could be based: the intrinsic merits or demerits of each of the two options, and (in this example) the financial considerations attached-but extraneous-to each option. 22ut we need not reject the distinction in order to respond to the inappropriate motivation objection.The assumption that we do rests, I think, on a conflation of two distinctions that the objection in fact involves: (1) a distinction between the intrinsic (de)merits of an option and the associated-yet-extraneous costs and benefits; and, (2) within the extraneous costs and benefits, a distinction between 'artificial' and 'authentic' cost/benefit balances.We should not conflate, however (as Raz's response seems to do), the claim that the notion of an option's intrinsic merits is unsustainable with the very different claim that there might be no way of demarcating an option's 'authentic' or 'natural' cost/benefit balance.And if that is correct, then it remains possible-in fact, I think it is the case-that (in a great many instances at least) we can distinguish between an option's intrinsic (de)merits and the extraneous costs and benefits attached to it, even though within the latter category there is no way of distinguishing between 'artificial' cost/benefit balances and ones that are 'natural', 'authentic', or uninterfered-with.
It is very rarely, if ever, of course, the case that we can make choices without taking extraneous (at the very least, financial) considerations into account.Our choices are normally motivated by a mixture of such considerations, on the one hand, and the perceived merits of the options before us, on the other.The more, however, we are guided by an appreciation of the intrinsic (de)merits of options rather than by the extraneous (dis)incentives attached to them, the better.And while I do not think we can distinguish between 'authentic' and 'artificial' cost/benefit balances (surely we should not associate the former with market-and the latter with government-regulated ones!), we can distinguish between cost/benefit balances that make it easier for us to act for the right reasons-to base our decisions on the intrinsic merits of the options before us-and ones that diminish our ability to do so.
Where does all this leave us with respect to the perfectionist policies we've been discussing?In some cases, the use of economic incentives clearly lacks the effect of adding extraneous considerations to the overall deliberative balance (or of making such considerations weightier).On the contrary, they can enable the recipients to be more rather than less responsive to the intrinsic merits of the options before them.This is the case when subsidizing an activity makes financial considerations loom less large in people's deliberations and thus frees them to focus on and respond to the activity's value.(This effect is especially clear when those affected by a subsidy are eager to engage in the subsidized activity anyway, but without subsidization are likely to be deterred by the cost.)In other cases, those who might be affected by the subsidy (and this applies equally to taxation) might initially be indifferent to the encouraged option, and, insofar as they choose it, might indeed choose it initially for the 'wrong' reasonsbecause it is inexpensive enough to be chosen on a whim, and so forth.In such cases we may hope, however, that exposure to the encouraged option will convince the initially indifferent of its merits, and that later engagement will thus take place for the right reasons. 23ut what about the problematic cases, those in which people want to pursue a taxed activity (or, perhaps less problematically, do not want to pursue a subsidized one)?Well, to the extent that we do not want people to be motivated by the wrong considerations, the answer is that government (dis)incentives should be set at such a level as to affect the eager and the indifferent, but not the resistant. 24That is the concession that supporters of perfectionist (dis)incentives would need to make if the worry about inappropriate motivation proved to be sufficiently troubling. 25

THE CASE OF RELIGION
I have argued that instruments of perfectionist policy, such as economic (dis)incentives, can leave individuals perfectly free to decide which conceptions of the good to endorse and in which activities to engage, and thus, in at least this crucial respect, need not diminish their autonomy.I hope to have shown, then, that appeals to the state's coercive character cannot by themselves ground a thoroughgoing case against political perfectionism. 26he conclusions for which I've argued so far may seem alarming, however, when applied to religious practice.Have I not committed myself, one might worry, to the idea that it is legitimate, for example, for a government to subsidize a religion it favors but not others?Or, even worse, to the idea that a government may tax the practitioners of a religion it disfavors (as long, that is, as the tax is not so onerous as to coerce the followers of the religion to stop practicing it)? 27 Or, at the very least, to the claim that a government may publicly avow its approval of one religion and aversion to another?If I object to such policies, on what grounds may I object, given my previous arguments?
My response is that I do object to such policies, but not because they need involve any objectionable coercion (which, in my view, they need not).Such policies do not coerce anyone to practice the encouraged religions or to stop practicing the discouraged ones-unless, that is, there is reason to fear that the policies will cause, or will be accompanied by, further acts of reprisal against followers of the discouraged religions. 28The regrettable history and reality of inter-religious relations is such, of course, that government endorsement of one religion over others is often likely to lead to acts of coercion and intimidation against followers of the discouraged religions (either by official representatives of the state or by other parties).But this regrettable state of affairs is unique to religion (and perhaps to a few other comparable areas of human life); it is by no means shared by all of the activities and pursuits that political neutralists habitually classify as 'conceptions of the good' and wish to exclude from the sphere of government action.There are strong moral considerations, then, against certain government policies concerning religion; but these considerations do not apply to the broader realm of 'conceptions of the good'.
There may, of course, be other general features of the state besides its coercive character, in virtue of which special restrictions might apply to its actions.One such feature is the state's special symbolic significance: evaluative judgments bearing the state's imprimatur often carry a symbolic significance for its citizens which similar judgments made by private individuals or organizations do not.This is so even when the judgments in question are not accompanied by coercive measures of any sort.The possibility of insult in such cases is considerable, and in many cases should weigh conclusively against state intervention.The gravity of the insult varies, however, from case to case: it is usually especially grave when religious matters are at stake, less severe when other matters (e.g., certain cultural pursuits) are at issue, and perhaps nonexistent in yet other cases.Concerns about the state's symbolic significance fall quite short, then, of justifying the thoroughgoing principled restrictions on state action that neutralists typically support.
The above remarks-about religion and state, about the state's symbolic import-are of course all too brief.Much more needs to be said in order to make them more precise and compelling.But enough has been said, I hope, to show that my arguments in the rest of the paper do not commit me to an endorsement of government intervention in religious matters.We may oppose such intervention and yet insist that a broad range of perfectionist state action is justifiable.(3) Whose actions are to be subject to neutralist restrictions: The actions of public officials?Of all citizens to the extent that they are acting politically (e.g., as voters)?(For simplification, I will use the term 'political action' in the restrictive sense, as referring only to the actions of governments, or states, and their official representatives.I will therefore use the terms 'political action', 'state action', and 'government action' interchangeably.) 3 These worries have been the main motivation for attempts to provide a 'neutral' justification of principles of neutrality, that is, an anti-perfectionist argument independent of perfectionist value claims.See Ackerman, 1980, pp.11-12; Larmore, 1987, Ch. 3.  4 The core anti-perfectionist claim-that there is a class of moral considerations which are appropriate for moral reasoning outside politics but inappropriate for political reasoning-need not be self-contradictory, even if it is conceded that any piece of moral reasoning must involve or rely on perfectionist considerations.Even if it is true that morality cannot be divided into conceptions of the good on the one hand and moral considerations that do not depend on any conception of the good on the other, this does not imply that the realm of moral considerations cannot at all be defensibly divided into two parts, such that it is at least coherent to think that one part may legitimately be applied to political action while another may not.As Jeremy Waldron (1993, p. 159ff.)puts it, it is not incoherent to be neutral among some conceptions of the good without being neutral among all; and if this is true, then it is not self-contradictory to offer a 'non-neutral', or perfectionist-dependent, justification for principles that retain the basic exclusionary, two-class structure of anti-perfectionist doctrines-so long, of course, as this is done with the understanding that the doctrine thereby becomes more modest in its aspirations.5 Whatever else may be said in objection to skeptical arguments for neutrality (e.g., that moral skepticism is false; or, that even if true, it cannot coherently be invoked to support any moral doctrine; etc.), taken by themselves they fail to identify any relevant difference between political and nonpolitical action.If such arguments apply to the former, they should likewise apply to the latter: they should place a constraint on all moral reasoning, or they should place no constraint at all.(See Waldron, 1993, pp.158-59; Raz, 1994, p. 99.)   6 Neutralist arguments that give pride of place to considerations of coercion are found (to mention a few prominent examples) in Rawls (1971, pp.328-29;  1993, p. 216ff.);Nagel (1987); Kymlicka (1990, pp.222-23); and possibly Larmore (1987), given a suitable interpretation of Larmore's 'norm of rational dialogue'.7 Raz, 1986, p. 161.See also Hurka, 1993, pp.158-60; Sher, 1997, Ch. 3;  Chan, 2000.8 'Morally principled', for there may well be reasons of a more pragmatic nature for limiting the power of governments to promote what they take to be good.

222
(On the limitedness of pragmatic arguments, however, see Sher, 1997, Ch. 5.) 9 Cf.Sher, 1997, p. 37. Arguing against strong neutralism is nevertheless worthwhile, I think, not only because it, rather than weaker variants, is the type of doctrine that neutralists most often defend, but also because it is the stronger variant that is truly distinctive in terms of its practical implications (i.e., in terms of the government policies it excludes), whereas the practical implications of weaker neutralist doctrines are all but indistinguishable from the practical implications of non-neutralist principles of toleration.10 Versions of these claims are presented most thoroughly by Waldron (1989), to whom I will devote the most attention, and more recently by Gerald Gaus  (2003).They are also implied by Rawls's claim that "all political power is coercive power" (Rawls, 1993, p. 216).Though I will not address a fifth challenge here, I will say a few words about it: Anti-perfectionists could also claim that even when perfectionist state policies neither coerce nor manipulate individuals into accepting or rejecting conceptions of the good (or the activities associated with them), the state needs to resort to coercive means-primarily taxation-in order to fund these policies; and that such coercive means, though defensible when the state excludes conceptions of the good from the justification of its actions, are objectionable when it does not.One easy responsetoo easy-would be to say that perfectionist government programs can be funded on a voluntary basis, for example by lottery proceeds.Any attempt to defend a wider range of real-world funding methods, however, would have to acknowledge the presence of coercion, yet explain why such coercion can sometimes be justified.One way to do so is to show that disputes about which conceptions of the good the state should support can be waged, or are typically waged, against the background of higher-level unanimity about the importance of the state supporting some conceptions of the good, and thus to show that the political resolution of such disputes ultimately rests on unanimous consent.An alternate strategy is to deny that unanimous consent on any level is a necessary condition of legitimate government action.(For discussion of the first strategy, see Chan, 2000; for a powerful defense of the second, see Raz, 1994, pp.355-69.My own view is the second, though defending it goes beyond the scope of this paper.)11 Waldron ,1989, pp.1136-37.
14 Waldron, 1989, p. 1142.15 Waldron's remarks, I take it, are intended to apply to activity-specific taxation and to subsidies alike; that is, they seem to count as coercive not only as threats but (at least some) as offers as well.16 Waldron, 1989, pp.1145, 1148-49.17 Note that there is a distinction to be made here (which Waldron does not acknowledge) between the merits and demerits of an activity, on the one hand, and the costs and benefits associated with it, on the other.I will return to this distinction shortly.18 This understanding of manipulative acts as either deceptive or exploitative of weakness is roughly in line with Joel Rudinow's (in my view largely successful) account of manipulation (Rudinow, 1978, pp.338-47).I do not attempt here anything approaching an originally developed or comprehensive account of manipulation.Nor do I assess in any considerable detail Rudinow's account (which, while capturing in my view the main features that distinguish manipulation from rational persuasion, on the one hand, and from coercion, on the other, is incomplete in certain ways, particularly with regard to the notion of 'playing on [exploiting] someone's weakness').19 For a discussion of the first, 'personality'-based condition, see Rudinow, 1978,  p. 342.20 Cf.Raz, 1989, p. 1234; Wall, 1998, p. 200.21 Raz, 1989, p. 1234.22 Perhaps the distinction cannot be made in all cases; and perhaps it is more difficult to make when what we have on the supposedly 'extraneous' side are not financial considerations but, say, considerations of social esteem.But that is less significant for our purposes, for what concerns us here are precisely financial (dis)incentives.23 Cf.Raz, 1989, p. 1233; Kymlicka, 1989; Sher, 1997, p. 64; Wall, 1998, p. 200.24 Alternatively, we can hope that those who at first resist an option would later, as a result of some incentive, come to see its merits.Clearly, though, such hope is not realistic in all cases.25 Ronald Dworkin (1985) offers the basis for a different kind of objection to government incentives, according to which government intervention of this sort results in an unfair allocation of resources.Dworkin's well-known argument goes roughly like this: The core principle of "liberalism" is a principle of equality, understood as equality of concern or respect for persons.Treating people with equal concern requires that we allow them equal access to resources.And this, in turn, requires that the allocation of resources be determined by voluntary market transactions, through which people can acquire the resources they need in order to pursue their preferred goals (their conceptions of the good life).(Dworkin's market mechanism is constrained, however, in various important ways: to be fair, it must begin from a starting point of equal allocated shares; each individual's choice of resources should be guided by his or her "internal" rather than "external" preferences; and so on.)Government interference with such voluntary market transactions results, then, in an unfair (because, in the relevant respect, unequal) allocation of resources.There are several ways to respond to Dworkin's argument (see for example Sher, 1997, Ch. 4; Macleod,   1998, Ch. 7), the most compelling of which, in my view, is to question one of its most fundamental moves.Dworkin starts by proposing a highly abstract principle of equality-equality as equal concern for persons-which (as he admits) is open to several more concrete interpretations (perfectionist, liberalneutralist, utilitarian, etc.).He then favors one concrete interpretation of the abstract principle, according to which to treat people equally is to allow them equal access to society's resources, where each person's choice of resources is guided by his or her conception of the good; and this, in turn, supports the ban on perfectionist government intervention.Why, however, should we favor Dworkin's particular interpretation of the abstract egalitarian principle?Why should we not say instead that to treat people with equal concern is simply to treat each in accordance with the entire range of moral considerations that apply to them?The reason Dworkin himself prefers the neutralist interpretation of the abstract principle is that he thinks it best captures, interprets and explains (and, in a certain sense, also justifies) a particular, historically salient, relatively well-delineated cluster of views about politics and public life, namely "liberalism."Two responses to this preference suggest themselves: (a) We might agree that Dworkin's is the best interpretation of "liberalism," yet believe that the proper goals of political philosophy are somewhat more revisionary than is suggested by Dworkin's interpretive method; or (b) we might accept Dworkin's non-revisionary philosophical method, yet deny that his is the most (or the only) satisfactory interpretation of "liberalism."(I thank Colin Macleod for pressing me to state my view of Dworkin's position.)26 Concerning the justifiability of some policies which are coercive-primarily the taxation needed to fund perfectionist subsidies-see n. 10 above.27 Some might object to such actions even when the subsidy or the tax is applied to all religions equally.I am not so sure that such policies are objectionable when the treatment is equal.For the sake of brevity, however, I'll address only the more clearly objectionable practice of subsidizing or taxing some religions but not others.(I am grateful to an anonymous commenter for urging me to clarify my view on this.) 28 See the final two paragraphs of Section 3 above.