Republicanism as critique of liberalism

The revival of republicanism was meant to challenge the hegemony of liberalism in contemporary political theory on the grounds that liberals show insufficient concern with institutional protection against political misrule. This article challenges this view by showing how neorepublican-ism


INTRODUCTION
Quentin Skinner (1998, x) presents the republican, or neo-Roman, theory he started developing in the early 1980s as a challenge to the "liberal hegemony" in contemporary political theory and a critique of its "ideological triumph." He particularly emphasizes liberalism's failure to appreciate the republican concern with institutions protecting citizens' liberties. Alan Patten (1996), however, finds "no interesting disagreement" between Skinner's republicanism and contemporary liberalism. The institutional protection Skinner promotes is no less important in the theories of liberals such as John Rawls and Ronald Dworkin,Patten argues. 1 Since Patten's dismissal of the republican critique of liberalism, Philip Pettit (1997Pettit ( , 2012aPettit ( , 2014 has emerged as the leading proponent of republicanism. Like Skinner, Pettit criticizes liberalism for failing to grasp the importance of institutional protection of citizens' liberties. Liberal theories, he has stressed repeatedly, have a weaker concern than republicanism for institutional protection of people's ability to effectively exercise the basic liberties. In this article, I explore Pettit's theory and consider how its institutional requirements compare with liberal institutions. Pettit aims his criticism at several liberal thinkers, but I shall focus on Rawls and his political liberalism because of its centrality in contemporary political theory, and because, as we shall see, Pettit himself contrasts his republicanism with Rawlsian liberalism in particular. I find that Pettit's theory is no more demanding than Rawls's theory in terms of institutional protection. Pettit has therefore not provided a republican critique of liberalism. I show that the reason for this compatibility between Pettit's republicanism and Rawls's political liberalism is that their institutional requirements are both constrained by a liberal conception of the common good. The view that legitimate institutions promote the common good lies at the core of the republican tradition. Pettit (2001a, 156-60) understands the common good to be expressed in people's shared, publicly avowable interests. More substantially, these are interests in a protected ability to effectively exercise the basic liberties (Pettit 2012a(Pettit , 2014. This account of the common good is both procedurally and substantively compatible with Rawlsian political liberalism. And since Rawls also sees the common good as the basis for legitimate institutions, the institutional requirements for his political liberalism are compatible with those of Pettit's neorepublicanism. 2 1 Patten (1996) also discusses Charles Taylor's republicanism, but I do not engage with Taylor's theory here, as it is not usually associated with the Roman tradition I focus on. 2 More precisely, Rawls requires legitimate institutions to promote principles based on an overlapping consensus, which is the set of the values reasonable persons share despite their different comprehensive doctrines-that is, the sets of convictions they hold about how to lead a good life, how one ought to treat others, and "much else that is to inform our conduct, and in the limit to our life as a whole" (2005,13). The terminological differences between Rawls and Pettit's theories, however, do not distinguish their institutional requirements.
Of course, this observation does not imply that there can be no republican critique of liberalism. The next step in the article is to show how Pettit's theory can be modified so as to actually formulate such a critique. By detecting fundamental similarities between Pettit's republicanism and Rawls's political liberalism, I identify neutrality between conceptions of the good as the element in neorepublicanism that must be removed to turn it into a critique of liberalism. 3 By weakening neutrality, republicans can define a conception of the common good that is compatible with a high level of political participation by citizens devoted to actively protecting their society against abuse of political power. Rawls (2001, 144) emphasizes the importance of citizens' role in maintaining just institutions, which Patten (1996) also notes when he points out, against Skinner, that Rawls appreciates the significance of citizens' political participation. But by rejecting neutrality in their own theory, republicans can argue that neutrality makes Rawls's demand for civic virtue too weak. Republicanism can then demand greater protection against dominating forces in society than liberalism can.
This move implies a restriction on the common interests a legitimate government must promote and therefore on what motivations people may act on. After all, greater institutional protection means the individuals operating the institutions must be more committed to making the institutions function in this way. On this more restrictive account of republicanism, the aim is to stimulate citizens' political engagement to ensure a high level of protection against political misrule. Republicans will thus reject neutrality and instead promote a particular comprehensive doctrine so that republicanism applies comprehensively to how citizens lead their lives. 4 I shall therefore refer to this nonliberal republicanism as "comprehensive republicanism," which I contrast with Pettit's "liberal republicanism." 5 I also contrast comprehensive republicanism with an account of republicanism that gives some priority to certain conceptions of the good while remaining tolerant toward all conceptions (Lovett and Whitfield 2016). This account also fails to break out of the liberal framework.
3 A person's conception of the good, Rawls (2005, 104) explains, consists of the ends and purposes the person considers worthy of her or his pursuit over a complete life. 4 Liberalism can also promote a comprehensive doctrine. For liberal perfectionists, individual autonomy is a crucial part of a good life and should be promoted in political decisionmaking. Nonliberal republicanism, however, applies comprehensively to individuals' lives to ensure that society be sufficiently protected against political misrule. 5 Pettit (2012a, 11n8) himself admits that "there might be some merit" in referring to his theory as "republican liberalism or liberal republicanism." A problem with comprehensive republicanism is that its implementation might be infeasible due to the plurality of conceptions of the good that characterizes modern society. That is, the required institutions might be unrealizable or unsustainable. This is Rawls's main reason for endorsing liberal neutrality. Pettit (1997, 173) also defends neutrality on these grounds by believing that attempts to extensively shape people's preferences will involve more arbitrary interference-that is, interference conflicting with citizens' interests-than they prevent. His republicanism therefore seeks "a relatively neutral brief for the state-a brief that is not tied to any particular conception of the good" (Pettit 1997, 120). But even if Pettit is right, a possible role for comprehensive republicanism remains: it can provide a basis for evaluating citizens' preferences and for criticizing liberalism for its failure to recognize the significance of the content of these preferences. So, even if Pettit is right that comprehensive republicanism cannot serve as a feasible target for actual institutions, it can nonetheless offer a critique of liberalism.

LIBERAL REPUBLICANISM
Pettit's republican theory revolves around the ideal of freedom as nondomination. On this account, you are free insofar as you live in a society where institutions deny others the opportunity to accumulate power to interfere with you as it pleases them. Interference in accordance with your interests, however, does not compromise your freedom. As a political ideal, republican freedom requires institutions ensuring that interference occurs only in accordance with citizens' common interests. For Pettit (2012a, 20), institutions are just and legitimate to the extent that they reliably promote and protect these shared interests.
What are these interests? In no society do the inhabitants share the exact same interests, especially not in a large, modern society characterized by a diversity of conceptions of the good. The fundamental question for both Rawls and Pettit, as Alan Thomas (2017) points out, is how to legitimize political institutions under such conditions. Pettit proposes a model of common interest that allows for sensitivity to this pluralism. On this account, an interest is a common interest to the extent that it is compatible with treating others as free and equal members of one's society and can therefore be avowed in public without embarrassment (Pettit 2001a, 156-60). Thus understood, a common interest is anyone's interest regardless of her or his conception of the good. A government promoting these interests is therefore neutral between the different conceptions of the good that exist in a society (Pettit 1997, 11). Pettit also declares republican freedom "a neutral political ideal" (97).
This neutrality is central in political liberalism, which takes institutions to be legitimate insofar as they promote a political conception of justice consisting of principles supported by public reasons. Public reasons are reasons every person can accept regardless of her or his conception of the good, insofar as she or he is reasonable-that is, willing to treat everyone as a free and equal member of society and to cooperate on fair terms. A legitimate government, then, serves the interests that citizens actually share, just as in Pettit's republican theory. Christine Korsgaard (1993, 50n47) unintentionally identifies this connection between Rawls and Pettit's theories when she writes that public reason justifications can be offered "to anyone without embarrassment." Pettit also admits to being influenced by Rawls. His own way of identifying common interests, he says, is "broadly contractualist in spirit; it owes much in particular to the interpretation of Rawlsian contractualism" (Pettit 2001a, 157n1). We see, then, clear procedural similarities between Pettit and Rawls's theories. Institutions are to be founded on interests each citizen can recognize as her or his own, otherwise they will feel alien to the citizens. For Pettit, such alienation is a source of domination, while for Rawls (2005, 98, 374), it conflicts with the ideal of political autonomy, which requires that political power be based on principles the citizens themselves endorse rather than principles externally imposed on them.
But despite procedural similarities, perhaps the two theories' institutional requirements are nonetheless substantively different. For Pettit (2012aPettit ( , 2014, the central requirement that emerges from the procedure is that of making sure everyone enjoys the protection and resources necessary for passing "the eyeball test." People pass this test when they are "able to walk tall, live without shame or indignity, and look one another in the eye without any reason for fear or deference" (Pettit 2012a, 3). Having this ability is to be free from domination. A dominated person cannot look his superior in the eye and talk straight to her as his equal. He must instead choose his words and actions carefully to maintain the more powerful person's goodwill toward him (59-61). Precisely how much protection and resources are needed for passing the eyeball test will vary from one society to another, Pettit explains (84).
The level of protection and resourcing must be sufficient to ensure everyone's ability to effectively exercise the basic liberties (Pettit 2012a). Once we see that this is Pettit's substantive institutional requirement, we recognize that Pettit's theory is also substantively similar to Rawls's political liberalism. Rawls does not explicitly mention any eyeball test, but the main institutional requirement of his political liberalism is to provide the means people need to effectively exercise the basic liberties. 6 Which principles will constitute a political conception of justice, Rawls (2005, 450-51) says, depends on the particular interests of the reasonable citizens of a particular society. But in any society, a political conception must contain a list of basic rights and liberties, assign special priority to these rights and liberties, and specify measures for ensuring that all citizens have the means necessary for effectively exercising the basic liberties.
The connection between Pettit's neorepublicanism and Rawls's political liberalism is further strengthened by their compatible accounts of the basic liberties. For Pettit (2008bPettit ( , 2012a the basic liberties are the liberties any person needs to determine for herself how to act without restricting anyone else's capacity to choose for himself how to act. Liberties to make oneself superior to another are not basic. The basic liberties enable people to determine for themselves how to behave without restricting anyone else's capacity to autonomously pursue her or his ends in life (Pettit 2012a, 94-95). For Rawls (2005, xii, 19, 52, 72, 81, 104, 302), similarly, the ability to exercise the basic liberties is necessary for developing the two moral powers-that is, to formulate and pursue a personal conception of the good and to develop a "sense of justice," which is the ability and willingness to respect others as equally entitled to pursue their conceptions of the good. 7 Samuel Freeman (2007) implicitly also identifies a connection between Rawlsian "high liberalism" and Pettit's republicanism when he says Rawls's concern with ensuring that everyone can effectively exercise the basic liberties is founded on the concern that no one should have to ingratiate oneself to others. Rawls, on Freeman's (2007, 187) reading, demands institutions that make everyone "socially and economically independent, so that no one need be subservient to the will of another." 6 But one liberal thinker, Joel Feinberg (1970, 252), refers quite explicitly to the eyeball test when he notes that institutional protection of liberal rights "enables us to 'stand up like men,' to look others in the eye, and to feel in some fundamental way the equal of anyone." 7 Pettit and Rawls also provide similar lists of basic liberties. For Pettit (2008b, 220;2012a, 103), the basic liberties include at least freedoms of thought, expression, religious practice, association, assembly, personal property, employment, movement, as well as to take part in public life as a voter, candidate, or critic. For Rawls (1999, 53), "important" basic rights and liberties include "political liberty (the right to vote and to hold public office) and freedom of speech and assembly; liberty of conscience and freedom of thought; freedom of the person, which includes freedom from psychological oppression and physical assault and dismemberment (integrity of the person); the right to hold personal property and freedom from arbitrary arrest and seizure as defined by the concept of the rule of law." This independence, Freeman says, is the primary end of justice in the high liberal tradition (45).
But while Pettit (2008b) admits inspiration from Rawls in his account of the basic liberties, he insists that Rawls fails to recognize the importance of the political liberties. Rawls "downplays the political liberties," Pettit (2012a, 144n12) argues, by "casting them as 'subordinate to the other freedoms' and does not suggest that they ought to ensure control in that sense." 8 For republicans, by contrast, the political liberties are of crucial importance for ensuring citizens' ability to effectively influence and contest political decision-making, thus avoiding subordination to arbitrary power.
On closer inspection, however, we see that Pettit quotes Rawls out of context. Rawls (1999, 205) does say that "the political liberties are indeed subordinate to the other freedoms that, so to say, define the intrinsic good of the [citizens]." But he then goes on to say that "of course, the grounds for self-government are not solely instrumental. Equal political liberty when assured its fair value is bound to have a profound effect on the moral quality of civic life." The political liberties would have been subordinate if they had only instrumental importance for ensuring good governance of society. However, Rawls thinks they also have intrinsic importance for the good of the citizens. As he says, these liberties "strengthen men's sense of their own worth, enlarge their intellectual and moral sensibilities, and the basis for a sense of duty and obligation upon which the stability of just institutions depends" (Rawls 1999, 206). People need these liberties to feel like full members of their society and to develop the reasonableness that Rawls and Pettit both consider important. Rawls therefore does not downplay the significance of the political liberties. In his later work, he also, like Pettit (2012a, 234), warns against "the curse of money" and argues for restrictions on the use of private wealth in politics to prevent unfair inequality of political influence (Rawls 2005, 456). Pettit (2012a, 254n5) also tries to distinguish his theory from Rawls's by arguing that while he thinks shared interests should motivate all political decision-making, Rawls (2005, 235) thinks they should only inform decisions concerning "constitutional essentials and matters of basic justice." Pettit makes this point only briefly in a footnote and leaves it unclear why he thinks the scope of his theory is broader than that of Rawls's theory. We have seen that Pettit's main focus is on providing the resources and protection citizens need to effectively exercise the basic liberties, and this is an issue Rawls (2005, 227) includes among the constitutional essentials. It is therefore unclear how Pettit's theory has a broader scope than Rawls's.
It is also possible that, once clarified, Pettit's point about the scope of Rawls's theory turns out not to be a critique of political liberalism. Jonathan Quong (2004Quong ( , 2011 argues that "the broad view" of public reason as applying to all political decision-making is more consistent with Rawls's theory than "the narrow view" he defends. Since Rawls sees the idea of public reason as an essential part of justifying political power, Quong argues, he should require that all political decisions be grounded in public reason whether they concern fundamental political issues or not. This makes the theory more firmly detached from any particular comprehensive doctrine. Pettit's argument for a broader scope could therefore be more in line with political liberalism than Rawls's own narrower scope. We see, then, that Rawls and Pettit's theories are both procedurally and substantively very similar. We can therefore understand why Thomas (2017) finds that these theories not only ask the same fundamental question, they also give very similar answers.

ROBUSTNESS
However, Pettit's main objection to liberalism is that it, unlike republicanism, fails to recognize the importance of institutions that robustly protect citizens against interference conflicting with their common interests. Liberals, in other words, are less concerned with protecting citizens' ability to effectively exercise the basic liberties than republicans are. Pettit (2012a, 11) grants that most liberals, like republicans, have endorsed institutions like the rule of law and the separation of powers, and they have, to a certain extent, recognized the importance of a contestatory citizenry.
But Pettit (2012a, 70) argues that while liberals value the absence of interference with the basic liberties, they typically overlook the importance of securing people against counterfactual interference. This point is tied to Pettit's view that liberals tend to understand freedom as the mere absence of interference, while the republican ideal of freedom as nondomination also demands institutional protection against interference. The differences between these concepts have been much discussed, and the significance of these differences becomes less clear when we consider the measurement of the two freedom concepts (Carter 2008;Kramer 2008). Institutions might therefore look very similar whether they promote the one or the other freedom ideal (Moen 2020). My focus on comparing the institutional demands of Pettit and Rawls does not require that I go deeper into this conceptual debate and intricate measurements of freedom. But Pettit (1997, 107) takes his desired institutional arrangement to constitute freedom. So, if Rawls defends the same arrangement, he also promotes Pettit's freedom ideal, albeit in different terms (Moen 2022).
But Pettit (2012a, 107-10) denies that Rawls follows republicanism in demanding that people have the resources and protection they need to exercise the basic liberties across a wide range of different circumstances. The result is that people might have to behave deferentially toward more powerful members of their society to gain permission to exercise the basic liberties. We saw in the previous section, however, that a defining feature of a political conception of justice is that it specifies institutional measures for ensuring that all citizens have the means they need to make effective use of the basic liberties (Rawls 2005, 450). And the requirement that they have this ability robustly seems to follow from the central Rawlsian concern already noted-that individuals be socially and economically independent. Such independence from the goodwill of others is exactly what Pettit wants to realize when he stresses the importance of robust institutional protection. Pettit (2012a, 69-73) therefore appears to have no good reason for thinking that political liberalism cannot also demand the resourcing and protection necessary for ensuring citizens' robust ability to effectively exercise the basic liberties. Thomas (2017, ch. 4) gives some support to Pettit on this matter by saying political liberalism does not adequately protect against the concentration of political power in the hands of a few wealthy individuals. An elite group of citizens may end up in a position where they can control others' ability to effectively exercise the basic liberties. This dependence on the goodwill of powerful individuals is exactly what republicans are focused on preventing. But unlike Pettit, Thomas recognizes that Rawls became increasingly concerned about this issue in his later work. This is evident, for example, in his concern about the "curse of money," as mentioned above. Thomas further thinks Rawlsians should adopt the republican measures for protection as a way of better meeting the principles of justice as fairness. Liberal republicanism therefore complements rather than challenges Rawlsian liberalism.
But perhaps Pettit thinks Rawls and other liberals fail to see the importance of citizens' civic virtue in establishing and maintaining the institutions that provide necessary resources and protection. After all, the effectiveness of institutions depends on individuals' commitment to making them work as intended. Pettit emphasizes the importance of citizens being vigilant of how political power is exercised and ready to contest any use of power they perceive to conflict with the common good. "People must be on the watch for proposals or measures that are not suitably supported . . . and they must be ready to organize in opposition to such policies," he says (Pettit 2012a, 226). Skinner (1990, 304) argues that Rawls and other liberals fail to see the importance of civic virtue in protecting institutions that preserve citizens' liberties. Patten (1996, 31) challenges this claim, but Pettit (2012a) continues to make this point against liberalism.
Is this charge against liberalism warranted? Rawls is certainly aware of the significance of virtuous citizens and emphasizes the importance of political participation by informed citizens for protecting society against misrule. Without widespread political engagement, he says, "even the best-designed political institutions will eventually fall into the hands of those who hunger for power and military glory" (Rawls 2001, 144). Citizens wanting to remain "free and equal . . . cannot afford a general retreat into private life," he says (144). They must do what it takes to establish and maintain just institutions (117-18; 2005, 194-95). A well-ordered society, Rawls (2001, 119) says, "encourages a political character that . . . sustains the political virtues of social cooperation." However, neutrality between conceptions of the good limits how much virtue political liberalism can demand from ordinary citizens. The ideal of civic virtue in political liberalism, Rawls says, "presupposes no particular comprehensive doctrine" (157). Citizens devoting a large proportion of their time to monitoring and, possibly, challenging the use of political power will contribute to the institutional protection against political misrule. But requiring a high level of such civic virtue from citizens would be unreasonable, as that would mean imposing a particular comprehensive doctrine on them.
Pettit's political participation requirement is similarly constrained, since he, as we have seen, also endorses neutrality. And this neutrality is compatible with his requirement that the constraint citizens impose on their government is only one of "virtual control," not "active control" (Pettit 2008a, 111-13). Pettit acknowledges that well-designed formal institutions are by themselves insufficient for consolidating popular control. In addition, he says, citizens must "always insist on the authorities going through the required hoops in order to prove themselves virtuous" (Pettit 1997, 264). They must be ready to contest decisions of government officials-elected or unelected-via channels such as the courts, the press, demonstrations in the streets, or by contacting their representative in parliament or an ombudsman (193; 2012a, 237). But they need not actively monitor the powerholders. They can therefore lead their lives as they wish as long as they remain ready to blow the whistle should they become aware of power abuse.
Thus understood, the republican requirement for political participation is compatible with the neutrality Pettit endorses. And this means he cannot demand more political participation than Rawls can. We therefore again see that there is no conflict between liberal republicanism and political liberalism. Pettit's attempt to criticize liberalism for failing to capture the importance of robust popular control of political institutions fails.

COMPREHENSIVE REPUBLICANISM
Pettit's commitment to neutrality makes his republicanism incompatible with a more demanding institutional protection than political liberalism can accommodate. Rejecting neutrality is therefore necessary for delivering on the promise of a republican challenge to liberalism. And neutrality does not seem necessary in republicanism. After all, the classical republicans of past centuries were insensitive to pluralism and saw it as the role of the polity to inculcate civic virtue in the population. They argued for the imposition of a comprehensive doctrine revolving around public service to a particular conception of the common good, which to modern liberal observers would be objectionable (Sunstein 1990, 181). Machiavelli (2008), for example, emphasized the importance of citizens devoting their lives to protecting their city from corruptive private interests as well as from foreign intrusion. In his view, one's own liberty and the greatness of one's city both depend on full devotion to civic virtue.
The differences between this classical republicanism and Pettit's liberal republicanism might not spring from different conceptions of the core ideal of freedom as nondomination. The relevant difference might instead lie in the nature of the society to which the ideal applies. While republican freedom has always required the protection and promotion of common interests, the substantive understanding of common interests has changed with the nature of society. The small medieval and renaissance city-states where classical republicanism was developed were far more conformist than today's large pluralistic countries. It could therefore be the pluralism of modern society, rather than a reconceptualizing of republican freedom, that has led Pettit to a less restrictive account of the common good. But this explanation of the neutrality of Pettit's neorepublicanism makes the theory no less liberal. It instead suggests that as society has evolved into a larger and more diverse set of individuals, republicanism has taken the form of political liberalism.
But to require more constraints on political power, republicans must turn away from a modern, liberal interpretation and back toward the classical form with a stronger emphasis on political engagement. To do so, republicanism cannot be neutral between reasonable conceptions of the good. It must privilege some conceptions of the good over others, particularly those treating political engagement as a significant part of the good life. This view is characteristic of what Rawls (2001, 142-45;2005, 205-6) calls "civic humanism." 9 Civic humanism promotes a particular conception of the good, and Rawls therefore considers it unsuitable as a political ideal. Given his commitment to neutrality, Pettit must also reject this doctrine, which he also does (Pettit 1997, ch. 1;2012a, 11-18;; see also Skinner 1990). 10 A concern for protection more demanding than liberal neutrality can allow for does not imply civic humanism. We need not see political participation as good in itself; we can consider it merely instrumentally valuable for its contribution to protection against political misrule. We can then say that a person can free-ride on others' political engagementthat is, benefit unfairly from the sacrifices of others. In civic humanism, however, such absence from political life is considered bad for individuals themselves.
I call the demanding, nonneutral form of republicanism "comprehensive republicanism," as it applies comprehensively to individuals' lives. It conflicts with liberal neutrality between different comprehensive doctrines, as well as with a perfectionist liberal concern with individual autonomy. 11 As on Pettit's view, it requires that citizens demonstrate against abuse of political power and report it to the courts, their representatives, and the press. But they must be more actively on the lookout for political misbehaviour. For Pettit, republican freedom depends only on citizens' virtual 9 Civic humanism has in recent decades been associated with writers such as Hannah Arendt and Michael Sandel. 10 Pettit gives various other names to this doctrine, including "neo-Athenian," "Franco-German," and "Continental" republicanism. He calls the republicanism he favors "neo-Roman" or "Italian-Atlantic," as well as accepting, as we have seen, "liberal." 11 For a "liberal republicanism" in line with liberal perfectionism, see Dagger (1997).
control-they must be ready to speak out if "the red lights go on," as he says (Pettit 2012a, 136n5). On the comprehensive account, on the other hand, citizens cannot just wait for the red lights to go on. They must be ready to switch on the lights by carefully watching political processes and how decisions are made and carried out. Citizens must also watch each other and put pressure on those perceived to not do their bit in keeping powerholders virtuous. Comprehensive republicanism rejects a division of labor that leaves politics largely to a few professionals while others perform other profitable activities. Citizens must have detailed knowledge of how institutions should operate to serve their common interests, as well as how they actually operate and be ready to speak out if they do not function satisfactorily. They cannot simply put their faith in government officials or hope that others will be vigilant, as that would leave them dependent on the goodwill of others. This account of republicanism is therefore more in line with the old republican motto "the price of liberty is eternal vigilance." Achieving this increase in institutional protection involves shaping individuals' preferences to motivate a higher level of political engagement. Pettit and Rawls both see the need for citizens' political participation to maintain institutions that protect their liberties. And the institutions should induce such behavior to ensure their own stability over time. People's preferences are not "fixed or given," Rawls (2005, 269) says. Just institutions should induce "political virtues," including political participation, to ensure that people remain committed to maintaining these institutions over time, thus ensuring the protection of their own liberties (205). Pettit also defends such measures. He, for example, sees the need for a "program of civics education" (Lovett and Pettit 2009, 23).
The difference between liberal and comprehensive republicanism is a difference in the trade-off between the requirement for political participation and the scope of opportunities people have for pursuing their personal ends. Republicans value protection against arbitrary power, and comprehensive republicanism takes this concern to its maximal extent. It only promotes interests compatible with a strong commitment to political participation and will therefore compromise citizens' ability to pursue reasonable conceptions of the good incompatible with a high level of active involvement in politics. It restricts the set of common interests by excluding interests conflicting with active control, thus ensuring more effective protection against arbitrary rule than with virtual control.
On the liberal account, on the other hand, the protection concern is traded off against neutrality between conceptions of the good. The vigilance requirement cannot be so extensive that it restricts people's ability to pursue any reasonable conception of the good. On Pettit's view, citizens must be ready to contest decisions or practices they believe to conflict with the common good, but they need not actively search for breaches of the citizens' trust. For Rawls (1999, 293-94), we have a natural duty to comply with just institutions and to contribute to their sustainability as long as "this can be done with little cost to ourselves." These liberal views do not require people to protect their institutions to an extent that will compromise their conception of the good.
On either account, individuals may have rights, but comprehensive republicanism will specify how they ought to exercise their rights. If an agent has a choice between doing x or y, we can say she ought to do x without preventing her from doing y. Liberal neutrality is compatible with education that emphasizes the importance of political participation, but it will place emphasis on providing information about the basic liberties and on honoring the various ways people exercise these liberties (Rawls 2001, 156). Comprehensive republicanism will be more specific about how citizens are to exercise their liberties, and it defends civic education designed for creating citizens devoting much of their lives to political activity.
Other commentators have argued not just that republicanism must step away from neutrality to challenge liberalism, but that Pettit's neutrality cannot fit within the republican framework. The republican focus on protecting society against arbitrary power, Frank Lovett and Gregory Whitfield (2016, 127) argue, should involve measures such as compulsory voting, subsidies for political activities like running for office and political organizing, as well as education and other policies designed for inculcating "a patriotic love of republican institutions." Republicanism, therefore, cannot be neutral, they argue. It is bound to give favorable treatment to some comprehensive doctrines. 12 Paul Weithman (2004), similarly, argues that republican measures intended to stimulate civic virtue must be based on a perfectionist argument for civic-mindedness as intrinsically good for individuals. As I note above, however, republicans can defend civic virtue solely for its instrumental value.
But on Lovett and Whitfield's (2016, 126) account, republicanism will satisfy "the principle of toleration," according to which "public policies, institutions, and so forth should impose no special disadvantages on any worthwhile conception of the good." They use the modifier "worthwhile" to exclude conceptions hostile to the value of equal respect, and therefore hostile to toleration itself. 13 Particular worthwhile conceptions of the good can be encouraged. Conceptions emphasizing the value of political engagement can for example be treated favorably by subsidizing political activities. Discouraging other conceptions, however, would violate the principle of toleration. While this view is a step away from neutrality, the concern for toleration is liberal. No conception of the good can be given unfavorable treatment as long as it shows equal respect for other worthwhile conceptions. Lovett and Whitfield's republicanism thus incorporates a liberal constraint on the measures permissible to incentivize political engagement that can enhance society's protection against political misrule.
Comprehensive republicanism will incorporate the measures Lovett and Whitfield see as crucial in any republican theory, as well as other measures that prove to effectively motivate society-wide political engagement. Education is certainly an important part of preparing citizens for a life of expressing their preferences and continuously watching government officials, both elected and unelected, to make sure they serve the common good. But comprehensive republicanism goes further in motivating political involvement. Comprehensive republicanism can impose "special disadvantages" on certain "worthwhile conceptions of the good" if it will serve this end. Or in alternative terms, it will be more restrictive on what it considers "worthwhile." As in Machiavelli's (2008) classical republicanism, we would be mistaken to think not devoting our lives to public service would be worthwhile or beneficial to ourselves (see also Cicero 2008, 7-8). There may, of course, be good reasons for imposing a liberal constraint (a point I return to in section 6), but to the extent that republicans do so, they weaken their concern for institutional protection and their critique of liberalism.

REPUBLICAN ETHOS
People are unlikely to behave in the way comprehensive republicanism requires without a social ethos that noncoercively informs their behavior. A social ethos is constituted by social norms, which are regularities in behavior that most members of society conform to and expect others to conform to. Under a republican ethos, people will win each other's approval for political vigilance and for challenging decisions that conflict with common interests. And they will be met with disapproval for not engaging actively in politics. People can thus motivate each other to become as politically active as comprehensive republicanism requires. This is an important part of solving potential problems of people free riding on each other's political vigilance.
Requirements for a restricted set of permissible preferences and a republican ethos make for a critique of liberal neutrality along the lines of G. A. Cohen's (2008) egalitarian critique of Rawls. Cohen focuses especially on the difference principle of Rawls's (1999) theory of justice as fairness, which requires that the worst off in society be made as well off as possible. 14 Cohen points out that Rawls himself says his theory depends on people's motivation by a sense of justice. If people's behavior is informed by a sense of justice and the difference principle is a principle of justice, then Cohen thinks we should expect them to do all they can to maximize the position of the worst off. Rawls (1999, 68), however, considers certain inequality-producing incentives compatible with justice insofar as they are necessary for motivating the most productive members of society to produce benefits for the worst off. This view conflicts with the requirement that people be motivated by a sense of justice, Cohen argues. It would be more coherent to say that justice, on Rawls's own terms, would be better served if the talented demanded no incentives for improving the situation of the worst off. If they were motivated by an "egalitarian ethos" rather than by inequality-producing incentives, there would be more resources available for assisting the worst off and therefore for making society more just.
The same line of reasoning leads to a defense of comprehensive republicanism. Here the issue is not that of making the worst off as well off as possible, but rather protecting society against arbitrary power. If republicanism is about such protection, then it is better served the more motivated people are to pursuing this end. Informed by a republican ethos, people will be motivated to do what they can in terms of vigilance and contestation 14 The difference principle has a less prominent position in Rawls's political liberalism than in his earlier work on justice as fairness. In political liberalism, as we have seen, the focus is mainly on the basic liberties. Following Cohen, however, we might still think that justice requires individuals' commitment to ensuring everyone's ability to effectively exercise the basic liberties. Their behavior might be informed by a particular social ethos so that they will not interfere with anyone's basic liberties, even when there is no risk of being caught and punished by formal institutions. to make sure powerholders remain virtuous and society, consequently, is protected against misuse of political power.
As noted above, liberal republicanism weakens the concern for protection at the benefit of neutrality. We can make sense of Pettit's weak protection requirement, as well as Rawls's lax interpretation of the difference principle, by noticing that they both build a personal prerogative into their theories. A personal prerogative gives agents permission to choose what is morally right for them beyond the point at which they have met their moral requirements. 15 Rawls includes a personal prerogative permitting the most productive individuals some scope for choosing for themselves the extent to which they will work for the benefit of the worst off (Estlund 1998). 16 Pettit, similarly, grants a personal prerogative when he allows people scope for deciding for themselves to what extent they contribute to protecting society against arbitrary rule. This is a concession to liberalism. And from the perspective of comprehensive republicanism, the personal prerogative weakens the republican concern about arbitrary power. Patten (1996) also notes that to the extent republicanism conflicts with liberalism, it gives individuals weaker rights and is more restrictive with respect to individuals' behavior. It demands more "nonliberal patriotism," he says-that is, greater commitment to political participation and civic virtue to secure liberty-preserving institutions. But this makes for no considerable critique of liberalism, he adds, since "liberals would be wise to stand their ground rather than attempt to take on board the republican point of view" (Patten 1996, 40). This is also Pettit's reason for making no such demand for nonliberal patriotism and for rejecting comprehensive republicanism. Comprehensive republicanism is undesirable, in his view, because of the interference involved in stimulating active popular control in a population of individuals, many of whom would reasonably object to such treatment (Pettit 1997, 173). Such interference would feel alien to the citizens, and therefore conflict with their status as free persons under a government operating on their terms. For this reason, Pettit (2017) also thinks the nonneutral institutions of comprehensive republicanism would fail to motivate stable compliance over time and therefore be infeasible. Ideals should be feasible, Pettit (2001b, 859) says, by which he means they can be expected to attract "full or at least adequate compliance." We should be able to expect citizens to act as the ideal requires, so as to sustain the ideal arrangements once they are in place (Brennan and Pettit 2005).

GOING EVALUATIVE
For Rawls (1999, 398), similarly, an ideal is "seriously defective" if it is so demanding that real people cannot be expected to bring themselves to voluntarily comply with them. We should be able to expect compliance over time solely by the use of democratic means (Rawls 2005, 142-43). And that calls for neutrality between conceptions of the good, given the fact of pluralism-that is, a diversity of reasonable comprehensive doctrines "is a permanent feature of the public culture of democracy" (36). Unless institutions are based on principles everyone can accept, they will not attract citizens' stable compliance. Lovett and Whitfield (2016) share Rawls's and Pettit's views that defining a political ideal must be constrained by considerations of what is realistically achievable in a modern society. But this implies tolerance, they argue, not neutrality. We have seen that republicanism, in their view, can give advantages to some worthwhile conceptions of the good, for example, by subsidizing various forms of political engagement. But while it can encourage some worthwhile conceptions of the good, it cannot give disadvantages to others. Republicanism is not neutral, on their account, but it must be tolerant of any worthwhile conception of the good (Lovett and Whitfield 2016, 126-27). Any contemporary public philosophy or political doctrine, they say, must satisfy a toleration principle.
These constraints on what a political doctrine can require are based on considerations of people's motivations. People, as they are in a modern society, will not commit to the behavioral pattern comprehensive republicanism demands, and trying to make them do so would involve impermissible measures. These observations may quite possibly be accurate, and since we have seen that comprehensive republicanism is not subject to such constraints, it might have bleak prospects as a guide for actual institutional design.
But we have seen that it takes the republican concern with protection against arbitrary power to the full extent. And, as I shall now argue, it can therefore provide a distinctly republican basis for evaluating states of affairs and for criticizing people's motivations and liberal institutions and practices. The sensitivity to facts about people's actual preferences and to the fact of pluralism is clearly important for making prescriptive claims about how society actually ought to be organized. Given what people in a modern society are like, comprehensive republicanism is perhaps both infeasible and undesirable as an ideal to actually work toward. But political philosophers can also make evaluative claims that cannot be falsified by such considerations (Gilabert 2011, 58). So, if we take comprehensive republicanism to make only evaluative, and not prescriptive, claims, then it cannot be dismissed on the factual grounds on which Pettit defends his liberal republicanism. Comprehensive republicanism can serve as a basis for criticizing people in modern society for their lack of republican preferences, and for criticizing liberalism for not showing sufficient concern with the negative impact this has on protecting society against political misrule.
While this evaluative approach has its defenders (Cohen 2008;Estlund 2020;Gheaus 2013), it has probably attracted more criticism than support. Rawls takes reasonable pluralism as an indisputable fact about modern society and would possibly hold that an evaluative ideal in conflict with this fact is no ideal at all. And with Geoff Brennan, Pettit sees no reason for caring about ideals that it "might be counterproductive to try to establish" (Brennan and Pettit 2005, 261). William Galston (2010, 406) thinks a principle that cannot guide real political practice is either false or useless. Gerald Gaus (2016, 16) finds no point in evaluating social institutions without considering how these institutions actually ought to operate. And David Miller (2013, ch. 10) thinks an ideal we do not know how to realize gives us nothing but pointless regret.
However, the evaluative approach can serve at least two useful purposes. First, comprehensive republicans can criticize citizens' lack of virtue when they give their political representatives opportunities for getting away with decisions conflicting with common interests. The fact of pluralism may be an unsurmountable obstacle to realizing the comprehensive-republican ideal, but comprehensive republicans can still say it contributes to making people insufficiently concerned about their vulnerability to arbitrary power. If protection against such power is what we want, and pluralism restricts the extent to which such protection can be achieved, then we have one reason for thinking pluralism is bad, however many reasons there may be for appreciating it. This view does not depend on a theory of how to actually make citizens more virtuous.
Comprehensive republicans might hope their criticism will stimulate more political engagement, but they need no good reason for expecting it. They can consider the prevalence of nonrepublican preferences a disease for society that no known treatment will cure. 17 Just because there is no known cure, republicans need not moderately conclude that the disease is no disease after all. When people lack motivation for making sure their interests are promoted by political powerholders, we can see this as a problem for the protection against arbitrary rule, whether we know how to remedy it or not. Republicans need not take the ultimate ideal to be the minimal level of domination that happens to be achievable given people's actual preferences. They can instead take it to be the level of domination they would enjoy if citizens had ideally republican preferences. Only then is society entirely healthy.
The second function of an evaluative approach is that it can help us make better decisions by making us more aware of what is at stake. Principles from which we can derive only evaluative claims can identify ways in which we would improve the current state of affairs whether we actually choose to pursue them or not. If we only consider principles satisfiable within the constraints of facts and institutional legitimacy, we shall fail to fully appreciate the cost involved in pursuing one course of action rather than another. We fail to see what is at stake and what we must, perhaps inevitably, forgo when we make collective decisions. So, while Miller may be right about an evaluative approach giving us reasons for regret, such regret is not pointless. It is rather a consequence of appreciating the costs involved in making difficult decisions. An ideal can therefore be useful even when we consider it unwise to pursue it (Cohen 2008, 353-61). 18 We can agree with Pettit that comprehensive republicanism is too demanding for a diverse population, but nonetheless find it useful for making us aware of the costs of giving up on the ideal of more politically engaged citizens.
Awareness of such costs can be illuminated particularly by trying to fit together values that are defined independently of each other and without sensitivity to facts about society. In this process, we see how values conflict and we can identify the conditions under which they are cosatisfied. It can thus help us see whether, or to what extent, values' cosatisfaction would be morally permissible and politically possible. Without necessarily 17 I borrow the disease metaphor from Anca Gheaus (2013, 456, 463). 18 Others have made related points. David Estlund (2014, 118) says "the fact that people will not live up to [a theory's standards] even though they could is, evidently, a defect of people, not of the theory." And David Copp (1996, 205-6) writes that "if justice by the lights of a theory is unachievable or unsustainable because of the motivations of those who benefit from something the theory would view as unjust, this may be no objection to the theory. . . . A theory's failure to satisfy the constraint is not sufficient to show it to be unjustified." providing solutions to real-world problems, this process can help us consider to what extent they should inform real political decision-making by specifying what it takes to cosatisfy different principles. 19 Cohen (2008, ch. 5) takes this approach by showing how equality, Pareto efficiency, and freedom of occupational choice are compatible when people hold egalitarian preferences. He then uses this possibility result to criticize actual people for their nonegalitarian preferences: if people had internalized egalitarian values, they would freely choose occupations so as to achieve efficiency without demanding inequality-producing incentives (Cohen 2008, 189-95). Cohen thus identifies the benefits of citizens with egalitarian preferences. But by taking relevant facts into account, the three values will not be cosatisfiable, since people typically do demand inequality-producing incentives for choosing occupations that benefit others.
Similarly, if people have preferences for a strong commitment to protecting society against arbitrary rule, we can expect comprehensive republicanism to be compatible with efficiency and a high level of freedom from interference. But the republican value of freedom as nondomination, on the comprehensive account, clearly conflicts with neutrality, as we have seen. And by taking facts of pluralism into account, it will also conflict with efficiency. People will likely prefer a society where they enjoy a wide range of opportunities to one where they have a narrow but robustly protected range. As Geoff Brennan and Loren Lomasky (2006) note, requiring citizens to devote their lives to political participation necessarily reduces their capacity to produce goods people want. 20 They ask, "Do we really want pilots to spend less time on take-off techniques and surgeons to stint on practicing suture tying so that they can devote the odd hour or two to the consideration of foreign policy?" (Brennan and Lomasky 2006, 233). But while comprehensive republicanism may be implausible as an ultimate ideal for a modern society, it illuminates the loss of protection against misrule involved in endorsing liberalism.
Actual political decision-making must obviously be sensitive to facts, such as the fact that people are not all devoted egalitarians or republicans. Legitimate institutions may therefore compromise comprehensive republicanism or Cohen's egalitarianism. But if we make this concession to liberalism, comprehensive republicanism enables us to see what we thereby forgo: opportunities for stimulating more political engagement in the population to enhance the protection against arbitrary rule.

CONCLUSION
The revival of republicanism was meant to give us an attractive alternative to the liberalism that has been, and remains, a dominant force in political theory. Neorepublicanism is presented as more concerned with protecting people against vulnerability to arbitrary power. However, neorepublicanism, especially on Pettit's influential account, offers no such protection beyond what we get from political liberalism. The revival of republicanism has therefore not posed a challenge to liberalism; it has instead consolidated its dominant position.
In this article, I have shown how neorepublicanism, at least on Pettit's formulation, has collapsed into political liberalism because of its neutrality between conceptions of the good. Because people in a modern society pursue a wide range of such conceptions, the set of common interests that Pettit believes a legitimate government must promote only includes citizens' shared interests in effectively exercising the basic liberties. This is also the main concern of political liberalism. To challenge liberalism, republicans must therefore take a more restrictive view of the interests that ought to be promoted. It must abandon its liberal neutrality, or the liberal toleration in Lovett and Whitfield's account, and restrict the domain of permissible preferences. Comprehensive republicanism considers a high level of political engagement compatible with any citizen's interests and restricts its account of the common good accordingly, thus ensuring firm protection against arbitrary power. By adopting this comprehensive account, republicans can therefore deliver on their promise of a theory that requires more protection against arbitrary power than liberalism does.
But all things considered, comprehensive republicanism may appear neither feasible nor desirable as a guide for political decision-making in a modern society. Pettit (1997, 173) may be right when he says that imposing a comprehensive doctrine on a diverse population would be counterproductive, as it involves the use of more arbitrary power than it prevents. The preference-shaping involved in promoting comprehensive republicanism might conflict with anything we would plausibly consider desirable and feasible in a large and pluralistic society. Comprehensive republicanism therefore does not appear to provide a plausible understanding of the common interests that legitimate institutions promote. But if this is right, then the prospect of a modern day republican critique of liberalism is a chimera. This is ultimately Patten's (1996) conclusion in his discussion of Skinner's republicanism: to the extent that it is a critique of liberalism, it is not one liberals should take any note of. But Patten does not consider how republicans can nonetheless form a critique of liberal society and the preferences of the individuals populating it. When decisions conflicting with stimulating a more virtuous citizenry are made, comprehensive republicanism can bring awareness to this cost. A theory can thus provide a basis for evaluation without making prescriptions about how things ought to be, all facts and values considered. Republicanism therefore need not collapse into liberalism, as it has done in recent decades under the lead of Skinner and Pettit. A republican critique of liberalism remains viable. 21