The Political Theory of Parties and Partisanship: Catching Up

Despite their centrality to modern democracy, until recently political parties were relegated to the margins of normative democratic theory, taking a back seat to social movements, civil society associations, deliberative experiments, spaces for local participatory government, and direct popular participation. Yet, in the past 15 years, a burgeoning literature has emerged in democratic theory focused directly on parties and partisanship; that is our focus in this review. We locate three main normative defenses of parties: one centered in the special role parties can play in political justification as agents of public reason, a second that looks to the way parties contribute to deliberation, and a third that focuses on the partisan commitment to regulated political rivalry and peaceful rotation in office. In this last connection, we survey work on the constitutional status of parties and reasons for banning parties. We then consider the relation of partisanship to citizenship,and in a fourth section we turn to the ethics of partisanship.Parties and partisanship are interwoven but separable: If partisans are necessary to realize the value of parties,the reverse


INTRODUCTION
In political science, parties are a defining element of representative democracy; as Schattschneider (1942, p. 1) wrote in Party Government, "political parties created democracy and. . .modern democracy is unthinkable save in terms of parties." Yet, until recently, a chasm existed between political science and normative democratic theory, which was silent on parties and partisanship.
A vast literature in empirical political science covers party systems, the function and dysfunction of political parties, campaigns and elections, partisan identity, and much more. In recent work on democracies in crisis, parties again occupy a central place (Cain 2015, Rosenbluth & Shapiro 2018. Strong parties and party systems are "democracy's gatekeepers" (Levitsky & Ziblatt 2018, p. 20), vital for democratic legitimacy and stability. Many political scientists worry that political parties, "once the primary instruments for combining democracy's positive virtues," are at risk (Runciman 2018, p. 214).
Despite their centrality to modern democracy, parties have been "the orphans of political philosophy," as Schattschneider (1942, p. 16) put it more than 75 years ago. They remained at the margins of normative democratic theory, if they appeared at all. They have been largely absent from agonistic, liberal, deliberative, and participatory democratic theories. There is a strong tendency in normative theory to view parties as corrupt and corrupting; as factions ambitious to exercise power; as unreasonable, unrepresentative, and undemocratic. Democratic theorists have entertained dispensing with both parties and elections altogether (Guerrero 2014). They have focused instead on the democratic value of social movements, civil society associations, deliberative experiments, spaces for local participatory government, and direct popular participation-referenda and initiatives (Fung 2012).
Our subject here is the political theory of party and partisanship as it has arisen-one might say finally-in the last 15 years. 1 A normative literature on parties and partisanship has developed, as theorists have begun to assign parties (and, to a significantly lesser extent, partisanship) a place in democratic theory. They invest parties with normative content. And they turn the tables, pointing out the consequences of a paucity of normative standards in the empirical literature (White & Ypi 2016, pp. 10-20;Herman 2017). Today, the distance separating empirical accounts of parties and their normative status is less stark, and increasing numbers of political theorists insist that democratic theorists and party scholars should talk to each other (van Biezen & Saward 2008, Bonotti & Bader 2015. Our subject is the burgeoning literature in democratic theory that focuses directly on parties and partisanship. 2 Our first section below assesses the literature on parties and representation. In the second section, we turn to a central theme: the normative defense of parties. One thread of defense argues that parties play a special role in political justification as agents of public reason. A related thread sees parties as contributors to deliberation. A third thread prizes their commitment to regulated political rivalry and peaceful rotation in office. In this connection, we survey work on the constitutional status of parties and reasons for banning parties. In the third section, our focus shifts from parties to partisanship and its relation to citizenship, and in a fourth section we turn to the ethics of partisanship. This organization indicates that parties and partisanship are interwoven but separable topics in political theory. A defense of parties often comes with a depreciation of partisanship, but partisanship too has defenders. If partisans are necessary to realize the value of parties, the reverse holds as well, and parties are necessary to realize the value of partisanship. Running through this review is our identification of three normative foundations of political parties. One is the value of dynamic political pluralism-diversity of opinion, experience, interest, and identity. The second is the value of institutionalizing pluralism in organized parties for the purposes of politics and governing. The third is the place of parties in acknowledging a legitimate opposition and commitment to peaceful rotation in office. We note where democratic theorists take account of these fundamentals or disregard them, and at what cost. The cost of disregard can be high, for historically (and still today) assaults on parties and partisanship are a principal way of assaulting democracy. Invariably, these assaults take the form of antipathy to pluralism, rejection of parties as a vital framework of pluralism in democratic politics, and insistence that the opposition is not a loyal opposition but a conspiracy against the public good. This is the malignant trio behind antipartyism today. In conclusion we ask, do theorists attend to these mounting forces? Does their work enable us to recognize and address the present danger head on?

PARTIES AND REPRESENTATION (THE PARTISAN CONNECTION)
Canonical work in the history of political thought recognized the role of parties (or proto-parties) in political representation. An abbreviated list of such thinkers includes Machiavelli, who saw in the Roman contest of plebs and elites a dynamic opposition that could function to stabilize a regime (McCormick 2011). David Hume appreciated parties in a similar fashion, and against the republican insistence on sameness-enforced through sumptuary laws and a static economy-saw parties as an inevitable and acceptable part of freedom and commercial society (Sabl 2012, Landis 2017). In the late eighteenth century, Edmund Burke's defense of partisanship was more explicit than his predecessors', which is why some see him as the founder not only of modern conservatism but of modern partisanship (Mansfield 1965). What stands out especially is his famous contrast of delegate and trustee models of representation. Some interpreters see a defense of party even in the writings of Madison, who, though he warned about faction, helped found the first American opposition party as the representative of republicanism (Connolly 2011). Recent work on eighteenth-and nineteenth-century parliamentarism shows how a variety of thinkers, including Burke and Germaine de Staël, believed that the bond of partisan attachment among representatives "moderated the corrupting effect of patronage" (Selinger 2019, p. 72). Scholars of the history of political thought have interpreted and mined these writers and others, paying attention to the theme of representation. Yet, until recently, contemporary political theory did not follow suit.
Four decades ago, Stokes (1968, p. 152) wrote that although "parties and party systems have played an immensely important role in developing the public's control of leaders and conferring legitimacy on the regime," theorists rarely connect parties to "the concept of representation." In Pitkin's (1967, pp. 147-48) classic account, The Concept of Representation, parties appear briefly and disappear quickly. Although, as Disch (2012, p. 605) argues, Pitkin's understanding of representation is more "radical" than many readers identified, Pitkin does not assign a central role to parties and partisans in the dynamic process by which representation is created. Hers is not the only account in which parties remain in the shadows-for in many innovative and compelling recent theories of representation, parties and partisanship remain absent (Guerrero 2010, Montanaro 2012.
Still, more theorists of representation are heeding the call for normative democratic theorists and party scholars to "talk to each other" (Van Biezen & Saward 2008) and recognize-sometimes with careful analysis, sometimes in passing-the central role of parties and partisanship in representation (Disch 2002, Urbinati & Warren 2008, Mansbridge 2011, Wampler 2012, Saward 2014. That said, in many instances, political representation is conceived of as a relationship between an individual citizen and an official representative, with party and partisanship playing a merely facilitative background role. This image of a representative as bound to a constituency in some respects but independent in others is rooted in the nineteenth-century era of parliamentarism-which is the historical location where some ethical dilemmas of representation, such as Pitkin's "mandate-independence controversy," are most appropriately situated (Manin 1997;Selinger 2019, pp. 173-74). Parliamentarianism was characterized by a relationship of personal trust between citizens and their chosen representatives, who were distinguished by their "local connections, their social prominence, or by the deference they provoked" (Manin 1997, p. 202). The legislature was a discursive body in which individual representatives could deliberate together and make decisions independent of their constituents' wishes.
In Manin's (1997) much-cited theoretical account, the connection between parties and representation came with the advent of "party democracy" in the twentieth century. People came to vote for the party rather than the person, and party voting became an expression of class identity (Manin 1997, pp. 208-9). Yet a variety of social changes upset the alliance of class-based cleavages and parties, and by the late twentieth century, it became increasingly difficult to sustain the partisan connection linking the society and the state. This created a need to reconceptualize parties and their role in representation. In empirical political science, there is debate about whether parties are adequately understood as umbrellas or alliances of activists and policy demanders or whether they have a more active and sustained role in mediating between ordinary citizens, activists and policy demanders, and the state (McCarty & Schickler 2018). In both political science and political theory, the decline of party democracy is seen as a crisis of representation.
In Europe, the crisis takes the form of parties that have disengaged from society. As Mair (2013, p. 97) writes, "Parties have reduced their presence in the wider society and have become part of the state. They have become agencies that govern. . .rather than represent" (see also Ignazi 2018). And as parties have disengaged from society, they have become "neither liked nor trusted" (Mair 2013, p. 73). In the United States, ideologically indistinct parties of the mid-twentieth century did not quite fit Manin's (1997) model of party democracy, and the crisis of representation took a different form. Rather than becoming agents of the nonpartisan state, as in Europe, parties in the United States became more ideological and more polarized. This, combined with the paucity of choices offered in a two-party system, created a chasm between partisan elites and ordinary citizens (Disch 2002, Fiorina & Abrams 2011. Parties in the United States have been captured by "the most energized segments of the population [that] attempt to pull government policy to their own preferences" (Cohen et al. 2008, quoted by McCarty & Schickler 2018. This disconnect has the same effect as what Mair describes for European parties; it severs the partisan connection between legislature and society (Fiorina & Abrams 2011, Gilens 2012, Achen & Bartels 2016. Whether the disconnect between parties and society in fact constitutes a "crisis" of representation is open to interpretation, but it does constitute a failure of the elemental linkage function long attributed to parties. This failure sets the stage for current scholarship in political theory on representation, which sees it as not necessarily situated in formal elective offices, and more contestatory and dynamic. Manin (1997, p. 226) puts it this way: As social and class cleavages become more complicated, representatives have to "take the initiative in proposing a line of division"; they have to call to awareness "this or that social division, drawing attention to a split in society that was not previously apparent." Saward (2010) goes further: To represent is to make a "representative claim"-a claim that is open to argument and contestation-and he challenges the assumptions that government is the primary container of representation and that elected officials are the quintessential representatives. Representation is not a fact-it is an event (Saward 2010, p. 39), and the representative claim is made across society, inside and outside the formal realm of official politics, from nongovernmental organizations to advocacy groups to international organizations (Saward 2010, pp. 26-27;Urbinati 2006, pp. 40-44). In this view, political theorists should focus more on whom these groups and associations claim to represent, and how, and on who contests these claims.
It would be a mistake to categorize elected partisan representatives as one set among many who make a representative claim, however. Legislatures are distinctive because of the power vested in them, whether a supreme power, as the Westminster Parliament possesses in principle, or the circumscribed power of legislatures in federal systems or separation-of-powers constitutions. Either way, it is not the legislature as a whole but the largest group within it that rules. The goal of creating a group or a coalition large enough to rule is what causes parties to form, and this irrepressible goal is what makes parties inevitable wherever elected legislatures exist (Muirhead 2014).
Parties and partisans make a distinctive claim: that they seek to govern. This is why the crisis of political representation cannot be elided by multiplying the sites of representation or by attending to the variety of competing representative claims that different actors advance. If parties cannot link the groups vying for power in the legislature with groups in the larger society, then legislatures, which are the heart of representative government, lose their connection to popular interests, wants, and passions, and representative government loses something of its legitimacy. Parties are the institution that successfully democratized representative government, and neither political science nor political theory gives us reason to believe that after their demise a different, more democratic institution will replace them.

DEFENDING PARTIES
Discussion of parties in normative theory only rarely focuses on representation and the institutional context of parties. Instead, a major thrust is a defense of parties for democracy and democratic practices broadly, and within this rubric are several strands of argument. We begin with the defense of parties rooted in political history and institutions. It focuses on the value of parties as nonviolent means of altering government and specifically on the commitment to regulated political rivalry and legitimate opposition. Two other strands of defense are rooted in political philosophy and identify the value of parties in effecting higher-level democratic ideals that are external to parties themselves. These theorists see parties as contributing to "justification," understood as legitimizing coercive laws within the contours of public reason, or to democratic deliberation. This literature stands as a correction to Disch's (2002, p. 108) observation that "today's proponents of deliberative and participatory democracy, schooled in the movement politics of the 1960's, hold party politics to be opportunistic, bureaucratic, and antithetical to citizenship."

Nonviolence, Regulated Rivalry, and Legitimate Opposition
The intellectual history of parties is largely a story of antipartyism and antipartisanship. For every effort to articulate a unified and stable order, parties and partisanship introduce political pluralism and with it division, dissension, and fragmentation. This is why, in its aim to describe a just and stable political order, political thought has been congenitally antipartisan. Rosenblum (2008) identifies two historically recurrent forms of antipartyism, both still prominent today. In one, thinkers averse to social and political pluralism see parties as unwholesome parts; they disrupt some presumptive natural or aspirational unity, or holism. In the other, thinkers accept the fact of pluralism in politics (taking the form of corporatism or the mixed constitution, for example) but see political parties as dangerously divisive. The exception to historical antipartyism is the "last party," a party that is formed in response to an emergency and that disbands once the threat is allayed. But this-a party "to end all parties"-is the exception that proves the antiparty rule (Rosenblum 2008).
The guiding idea behind this defense of party is that if we accept pluralism, we cannot reject parties, and that the acceptance-indeed valuation-of pluralism entails political parties. There can never be one sole set of interests, or one authoritative interpretation of the common good. For that reason, rival positions should not be taken as purely strategic (Gutmann & Thompson 1996, p. 82). Transforming pluralism into ongoing, managed, institutionalized conflict among parties is a hard-won and fragile historical development (Hofstadter 1969). The legitimacy of the other side to compete for power is acknowledged. Rotation in office entails a method for determining winners and acceptance of the results by losers, including acceptance of policies they oppose. Crucially, opposition remains alive and continues; loss is not irreversible; "elections are not followed by waves of suicide" (Schattschneider 1942, p. 91). Regulated party rivalry for office distinguishes party opposition from sedition, treason, conspiracy, rebellion, or civil war.
The pluralist defense of parties is liberal insofar as it rejects political "holism" and any invocation of oneness-be it populism, fascism, theocracy, or a general will (Mouffe 2000, p. 21). It rejects any version of representation that rests on some ideal of totalism or unity, and it rejects every break with party politics motivated by abhorrence of division (Urbinati 2014, pp. 137, 144). It assigns to regulated party rivalry and loyal opposition the crucial business of making pluralism political by structuring conflict and legislative debate, shaping citizen choices, and holding the opposition accountable (Urbinati 2014, p. 229;Waldron 2016, p. 101).
Liberal too is the argument that regulated party rivalry normalizes and institutionalizes the "radical chastening" of political authority (Kateb 1992, p. 37). It undergirds the "provisional nature of political authority" (Rosenblum 2008, p. 363).
Political organization takes many forms, of course, but parties alone bring opposition into the frame of government (Fox & Nolte 1995, Waldron 2016). Waldron emphasizes that acceptance of party division as legitimate and acceptance of the loyal opposition are not adequately understood in terms of tolerance of disagreement, free speech and association, or even organized political opposition to government. A legitimate opposition is characterized by the principle of "positive empowerment" (Waldron 2016, p. 106): The opposition party has an officially recognized role within government-representation on legislative committees, for example (Waldron 2016, pp. 105-6;Rosenbluth & Shapiro 2018, pp. 36-38).
This political justification of parties raises the question of the bounds of loyal or legitimate opposition. How do we assess a charge that the opposition is not legitimate but rather is a conspiracy to undermine the nation or the constitution? If regulated party rivalry requires "a common grammar" (Urbinati 2019a, p. 102), how deep does that go?
The simplest criterion for banning or excluding party opposition is violence. This criterion would include party organizations that are fronts for paramilitary activity and insurrection. One difficulty is that evidence of complicity between armed militants and party leaders or members is always contested. And although violence excludes a party from the status of a loyal opposition, other extraelectoral actions do not-civil disobedience is not disqualifying, nor is using the electoral arena as a forum for mass political mobilization that then extends beyond campaigns and elections (Issacharoff 2015, p. 37).
A major strand of this literature on legitimate opposition addresses the question in terms of militant democracy: anticipatory measures taken by states to prohibit the formation or participation in elections of parties that exploit electoral politics to undermine democracy. The so-called paradox of democracy-restricting rights of political association if they threaten democracy-is not a paradox at all if we assume that "a Constitution is not a prescription for suicide" (Barak 2002, p. 44). The concept of militant democracy is the legacy of Weimar (Loewenstein 1937), but in practice, the standards and defensive measures written into law today closely track each nation's particular historical experiences (Rosenblum 2008, Kirschner 2014, Issacharoff 2015. Studies in this vein probe not only the bounds of legitimate opposition but also who polices these boundsconstitutional courts or legislatures or other entities (Issacharoff 2015, pp. 42 n.43, 60).
Within the rubric of militant democracy, then, standards vary. One measure of legitimate opposition on offer is adherence to established constitutional arrangements. Rawls (1993, pp. 227-30) makes the focus of loyalty "constitutional essentials," which include the idea of loyal opposition itself, arguing that frequent controversy over the structure of government undermines constitutional government. One difficulty here is defining "constitutional essentials." Another difficulty with adherence to constitutional arrangements as a standard of legitimate opposition is the propensity of parties to insert their positions into state or federal constitutions-prohibition of alcohol, for example, or bans on same-sex marriage-with the result that the constitution as currently interpreted cannot serve as the baseline of loyalty (Waldron 2016, p. 117).
Besides constitutional loyalty, a prominent justification for excluding or constraining parties is incitement of hatred toward a racial, ethnic, or religious subpopulation. Here, the concern is that party activity erodes democratic attitudes over the long term; for example, the aim of criminalizing neo-Nazi groups is not only to protect the state from Nazis but also to protect young people from Nazism. Another concern is that without constraint of parties, political hate speech intimidates voters and depresses participation.
Waldron shows the difficulty of isolating criteria for legitimate opposition, including adherence to electoral rules of the game or the amendment process. In place of a doctrinal or behavioral test, he proposes that loyal opposition refers to the way the opposition party must be regarded in a constitutional system. It works "as a sort of admonition to the ruling party" (Waldron 2016, p. 122). That is, the opposition is to be regarded as having constitutional and political standing to run the government (Waldron 2016, p. 120).
Whereas Rosenblum, Waldron, and Issacharoff argue that context and complexity militate against a single regulatory principle for defining, excluding, or limiting antidemocratic parties, Kirschner (2014, p. 25) provides one that applies to all political opposition, not parties specifically: "self-limiting" militant democracy. Antidemocrats have indefeasible rights to participate, on his view, and any exclusionary rule or ban should be limited to thwarting activities that violate others' core interests in participation.
The matter of legitimate opposition and measures to ban parties or otherwise constrain their activities is no longer restricted to antidemocratic parties or bound by the concept of militant democracy. Existential challenges have come to the fore. From this standpoint, political pluralism does not translate into legitimate opposition if parties use elections to transform the established idea of national identity (e.g., a Jewish state) or national political identity (e.g., secularism). However, the core idea behind the political justification of parties-the value of pluralism and its nonviolent institutionalization in regulated rivalry-argues that essentialist questions of identity too must be open to party contention (Rosenblum 2007, p. 67).
This political justification of parties (nonviolence, regulated rivalry, and legitimate opposition) has resonance today, when the fundamental value of pluralism and commitment to regulated rivalry-"that the losers of today may unseat the victors in a new round of electoral challenge"-is under attack, practically and ideologically (Issacharoff 2018, p. 485).

Public Reason
Theorists working within the framework of public reason defend parties in different terms. They value parties insofar as they take up the business of justifying the exercise of coercive authority in conformity with the requirements of public reason. Public reason is based on the elemental notion that to regard others as free and equal requires that the exercise of state power be justified with www.annualreviews.org • The Political Theory of Parties reasons open to the understanding of all and that others can accept. Philosophical accounts of public reason by Habermas (1995) and Rawls (1993Rawls ( , 1999 are frequent touchstones. This literature in defense of parties challenges the prevalent assumption that parties are identified with coercion, bargaining, and rhetorical manipulation and are fundamentally incompatible with public reason or, in a related defense discussed below, with deliberation (Christiano 1996).
Thus, Bonotti (2017) argues that Rawls' political liberalism nourishes parties by leaving many issues open to democratic contention. His "wide" concept of public reason allows the introduction of comprehensive doctrines in public discussion-provided that in due course public reasons are offered. On Bonotti's (2017, p. 123) interpretation, it falls to parties to identify those public reasons, and "an internal connection" exists between the demands of public reason and partisanship.
Working with a Habermasian conception of justification, White & Ypi (2016, p. 57), too, see "a deep structural affinity between the practices of partisanship and political justification." Partisanship is a politics of principle that speaks to the whole and is concerned to justify its commitments with general reasons. Moreover, insofar as justification is adversarial, party competition makes competing reasons accessible: It exposes citizens to political justification by demanding that reasons be engaged rather than censored, ignored, or dismissed.
For these theorists, parties are what make possible a politics of public reason. This is a view of parties from the outside; it revolves around an independent concept of justification and stands apart from existing political institutions. This work argues that parties can contribute to the circumstances of justification, but a lot is lost if parties are defensible only in these terms. To the extent that parties are valued mainly as agents of justification, other democratic purposes such as representation, political mobilization, and defining the terms of political division are eclipsed. To see parties in this way is to understate their role in linking government and pluralist civil society. Muirhead & Rosenblum (2006, p. 99) take this up in their discussion of Rawls' express disdain for parties and "the great game of politics," which "betrays the marks of warfare." They argue that political liberalism nonetheless accommodates parties. Rawls admits that his view of the most reasonable conception of justice as fairness stands on a political spectrum and can be seen in partisan terms. Parties also unite what Rawls calls the background culture of society, where associations rooted in religious, moral, and philosophic ideas that do not comply with public reason have full play, and they bring the content of politics into the public political forum of constitutional democracy. Their value is not justification in terms of public reason but rather their "bilingual" translation between civil society and the constitutional sphere (Muirhead & Rosenblum 2006).

Parties and Deliberation
The thrust of deliberative democratic theory is its contrast to classic democratic theory's emphasis on voting in terms of pre-existing preferences that are "tainted by their causal history" (Beitz 1990, p. 179). Deliberation helps shape informed preferences and offers justifications of these preferences to others. Until recently, parties have been ignored or depreciated both as deliberating actors and as forums of deliberation. Instead, political theorists focus on designing special deliberative institutions: deliberative polls, citizen juries, and mini-publics. These are enhancedinformation environments, with moderators guiding discussion, which yield decisions that are nonbinding, typically on a single issue. These deliberative settings are divorced from elections, and in some accounts partisans are explicitly excluded from participation. But there are exceptions. Hendriks et al. (2007, p. 362) ask, "Given that partisanship is an inevitable part of political life. . .how can it best be accommodated in deliberative practice?" For many theorists, the goal of deliberation is arriving at impartial judgments of the common public interest, and the desired outcome is consensus. Yet not all theorists insist that deliberation reduces the scope or intensity of disagreement or aims at informed consensus or requires disinterestedness; clarifying aspects of commonality and disagreement is also a valued outcome (Gutmann & Thompson 2004). As accounts of democratic deliberation became more expansive, theorists began to consider not just specially designed deliberative settings but deliberative democracy writ large, in which parties have a role. Manin (1997, p. 15) proposes citizen deliberation informed by experts and moral authorities but allows participation by politicians provided it is decoupled from elections and campaigns. Parekh (2000, p. 306) prescribes public forums for cultivating "dialogically constituted multicultural society" and for making representations to the legislature, though he excludes parties in the legislature from serving this purpose. Ackerman & Fishkin (2004) prescribe "Deliberation Day"-nationwide citizen forums deliberating issues in anticipation of elections. They assign major parties a role in selecting issues and include debates between party representatives. That said, these deliberative settings are designed to stand outside of existing political institutions, and their decisions are not binding.
The agents of deliberation and its meaning and purpose are inseparable from institutional context, and normative theory lags behind empirical political science and positive political theory when it comes to the purpose of parties and party leadership in legislatures. There is little on the role of parties in areas such as agenda setting, committees, institutional rules and incentives, or the privileges of a minority party. Democratic theorists do consider whether and how partisan legislatures might function not only to publicize the reasons that ground these commitments, but also to prod "the parties to engage with the positions and the reasons of their adversaries in a way that informs citizens about the facts, the issues, and the options at hand" (Leydet 2015, p. 236; see also Mansbridge et al. 2012). Others question whether the party discipline that characterizes parliamentary government is consistent with the ethos of deliberation (Bhatia 2018). Work in this vein also challenges the sharp dichotomy that deliberative theorists sometime draw between political arguing and political bargaining (Habermas 1995(Habermas , 1996Elster 2000). Mansbridge et al. (2012) introduce the notion of "deliberative negotiation" based on mutual justification, respect, and reciprocal fairness, although the authors do not discuss partisan negotiation specifically. Finally, democratic theorists could do more work on the interbranch relations between parties and administrative agencies and courts (Schwartzberg & Knight 2020). Hofstadter (1969, p. 72) emphasizes that deliberation "goes on not merely in the legislative process. . .but in the internal processes of the great political parties." Political theorists such as Wolkenstein and Invernizzi-Accetti attend to intraparty deliberation. Standard views see the internal working of candidate selection and agenda setting in terms of elites and already-formed preferences. In contrast, the argument for parties as deliberative forums emphasizes preference formation in local branches of the party. Intraparty deliberation from the bottom up connects the party base and government, on this view, so that parties mediate between society and the state (Wolkenstein 2016, Invernizzi-Accetti & Wolkenstein 2017. The argument is that parties' claims to be representative have declined because they have lost their deliberative and participatory qualities. A broader argument about deliberation concerns parties in the open public sphere, engaged in shaping opinion and garnering support. At this level, parties do what philosophy cannot do, and what no other political organization can do as consistently or comprehensively: embrace and organize pluralism for democratic politics. Yet theorists of deliberative democracy have "abandoned mass democracy" (Chambers 2009, p. 323). Parties' creative political role has been given short shrift (Herman 2017, p. 741). Some theorists, however, have attended to the importance of parties for democratic deliberation in the wide sense. Rosenblum (2008) argues that they shape the range of matters for discussion and decision; parties create and articulate lines of division. Party rivalry focuses attention on specific problems, brings out information and interpretations, and delineates what is at stake. Without parties selecting and excluding, organizing and articulating issues, democratic theory's "trial by discussion" cannot be meaningful (Rosenblum 2008, p. 307). In institutional contexts like legislatures but also in the general public sphere, Muirhead (2019, p. 83) argues, "parties are necessary to deliberative democracy." These normative defenses of parties argue that they can contribute to the circumstances of justification and to deliberation, but much of what makes parties defensible and valuable is left out. Some of what is left out, we have said, is fundamental to democracy: parties' organization of political pluralism and commitment to regulated rivalry and legitimate opposition. We now take up the question of parties' role in democratic citizenship and participation.

CITIZENSHIP, PARTISANSHIP, AND INDEPENDENCE
Political participation is central to almost every conception of good democratic citizenship, but too little attention is paid to the critical connection between participation and representation as it is facilitated by parties (Urbinati & Warren 2008). In recent literature, parties are seen as essential for participation. A collective "we" is necessary to sustain and enhance political commitment, and parties keep the project visible and give it continuity over time. For Efthymiou (2018), for instance, partisanship is valuable because it supports political commitment and contestation. Parties inform, mobilize, and organize participation, and they connect government to citizens via "the partisan connection" (Muirhead & Rosenblum 2012). Their role includes education about political issues, "epistemic resiliency" to resist propaganda and misinformation (White & Ypi 2016, p. 93), and "cognitive mobilization" (Invernizzi-Accetti & Wolkenstein 2017, p. 98). Parties are a "useful training ground" for citizens to relate their comprehensive doctrines to liberal democracy, thus contributing to the overlapping consensus that guarantees stability (Bonotti 2017, p. 100). Another claim is that parties are "collective epistemic agents" that transform abstract value judgments into coherent conceptions of justice, reducing complexity, lowering the threshold of political competence, and creating epistemic equality (Ebeling 2016, pp. 650, 640). White & Ypi (2016, pp. 96, 210, 212-16) identify principles of party structure that enable a "democratic ethos." For Urbinati (2019b, p. 99), "[p]artisanship is in fact a term for thinking politically." A stronger version asserts that to be a good citizen, one needs to be a partisan. Partisanship itself, not just partisanship as a spur to participation, is a civic ideal (White & Ypi 2016). What makes partisanship a defining element of good citizenship is that it enacts a commitment to pluralism: recognizing the legitimacy of opposition, complying with the rules of regulated rivalry, and accepting the partiality of partisan claims. The "moral distinctiveness" of partisanship lies in "commitment to the provisional nature of political authority, its periodic recreation" (Kateb 1981, p. 358). The usual view is that, to serve their valuable purposes, parties require partisans, but it is also the case that parties are necessary to realize the value of partisanship (Rosenblum 2008, p. 367).
Today, the normative value of partisanship is challenged by claims that Independents have superior moral standing. In this view, parties may be a systemic sine qua non of democracy, but partisan citizens are not. Rather, democracy requires free-thinking, independent-minded citizens and representatives who are not swayed by party loyalty, who weigh rival claims, and who go where the facts and evidence lead them. The logic of this view is that Independent status is not adequately characterized as nonpartisanship but, in the United States at least, as a distinct political identity that entails antipartisanship (Rosenblum 2008). Independents are defined in contrast to partisans, who are said to be bought, or vulnerable to the undue influence of activists, or who are themselves extremists who contribute to the skewing of representation and the polarization of politics. The claim is that Independents constitute a more deliberative public and arrive at an unbiased estimate of the public good (Dalton et al. 2000, p. 60). The normative ideal of independence has institutional consequences, among them nonpartisan primary elections that reflect antiparty animus and whose constitutionality has been the subject of US Supreme Court cases (Thompson 2002).
Standing for principles has integrity; this is what the normative ideal of independence is about. But for defenders of partisanship, antipartisan voters are "not self-reliant but weightless" (Wolfe 2000, p. 15). Politics is not only about standing for. It is always also about standing with. Thisstanding with others-is the challenge of the ethics of partisanship.

THE ETHICS OF PARTISANSHIP
The ethics of partisanship has two aspects. From one standpoint, it consists of constraints on the partisan's pursuit of actions that might be tempting or effective, but are wrong judged by standards external to partisanship. For instance, White & Ypi (2016, p. 2) emphasize that standards of political justification should constrain partisanship. The second aspect of the ethics of partisanship denotes virtues distinctive to partisans that are cultivated and expressed through partisan activity. The most fundamental of these is the willingness to stand with others in a group sufficiently capacious that it can act with democratic legitimacy (Muirhead 2006, White & Ypi 2016. Within this framework, we find the characteristics that political theorists most often identify with ethical partisanship: inclusivity, comprehensiveness, compromise, and loyalty. Inclusivity is the desire to win office (and power) on the most democratic terms possible-with a popular mandate. As an empirical matter, some might see the inclination to inclusivity as solely the result of electoral incentives established by the electoral rules of the game, and the necessity to build a large electoral coalition will depend on whether a system is first past the post or proportional. Inclusivity implies that parties in democracies will not pursue disenfranchisement because such antidemocratic strategies will never be in their interest. When Schattschneider (1942, p. 1) wrote that "political parties created democracy," he tacitly invoked this idea that democratization works in only one direction-the franchise is extended but not retracted. But retractions of the franchise do occur. Some parties make democracy, as Schattschneider said; others can diminish it-or destroy it (L.E. Herman & R. Muirhead, unpublished manuscript). With respect to inclusivity, electoral incentives matter enormously, yet are insufficient. Partisans also need an ethical commitment to inclusivity, understood as winning on democratic terms.
Comprehensiveness is what definitionally separates a party from a faction, or what Sartori (1976, p. 25) calls "only a part for itself" (Bonotti 2017, p. 10). Comprehensiveness means that partisans take a view on the full range of issues that constitute the public interest-from the economy to national security. In contrast to single-issue advocacy groups, parties address the public good in the widest sense. This is why parties have platforms that do not claim to benefit just one group at the expense of others, but to benefit the nation. At their best, partisans attempt to address the common good, even though they do not presume to speak for the whole. In White & Ypi's (2016, p. 53) terms, partisans should be "nonpartial." One might argue that comprehensiveness is more evident in large coalitions, but there is no correlation between the raw size of a party and the comprehensiveness of its arguments and platform. Small parties and parties of conscientious dissent might address questions of justice and the national interest more comprehensively-and, for many, more compellingly-than the large parties that dominate a party system.
Inclusivity and comprehensiveness lead to the third element of ethical partisanship, a disposition to seek compromise. For some thinkers, compromise is the signal virtue of parties. This was the view, for instance, of Hans Kelsen, whose 1929 On the Essence and Value of Democracy justified parties mainly in terms of their role in "creating the organizational conditions" for compromise and "moving the common will toward a median" [Kelsen 2013[Kelsen (1929, quoted by Mersel (2006), p. 163]. Kelsen's arguments and his theory of "party democracy" command the attention of a new generation of democratic theorists working on partisanship (Ragazzoni 2017). While the possibility that parties might compromise with each other was central to Kelsen's thought, it is important to recall that the disposition to compromise is located within parties as well: Without an ability to compromise with copartisans, there is no party, no group large enough to legitimately govern.
The Spirit of Compromise by Gutmann & Thompson (2012) is a comprehensive discussion of the political need for a "compromising mindset" and the inseparability of the question from institutional context. The authors distinguish compromise, which entails mutual sacrifice and likely internal incoherence in the result, from "common ground." They reject the interest/principle distinction as a way of evaluating political compromise. Adapting one's principles ("principled prudence") and respecting the opposition are set against standing on principle and against the mistrust of opponents that characterize the uncompromising mindset. Rosenblum (2008, p. 42), too, argues in this vein: Uncompromisingness inhibits the public business from getting done, and therefore represents an abdication of responsibility for governing.
Consistent with this, Beerbohm (2018) locates compromise as a democratic practice peculiar to the legislator's role. He argues from the vantage point of what he calls "co-ownership." Compromise is defensible insofar as concessions "honor the procedural value of deciding together." Compromise "can raise a law's democratic credentials," and it can increase the "force of bipartisanship in a two-party system" (Beerbohm 2018, pp. 6, 30).
Whether compromise is part of the ethics of partisanship is disputed. Some think that the epistemic significance of political disagreement argues for "conciliation" rather than "compromise and consensus" (Ebeling 2016). White & Ypi (2016, pp. 142-63), too, challenge the importance of compromise. Although they concede that particular compromises may be advisable, "as an ideal, political justification constrained by norms of public reason is always to be preferred over the politics of compromise" (White & Ypi 2016, p. 163). A different standard connects the limits of compromise "to the life of the partisan project itself"; to compromise too much is to risk losing the soul of the party (Urbinati 2019b, p. 100).
The counterpoint to compromise is loyalty-the quintessential partisan virtue, without which any standing with others is impossible. Loyalty has its hazards: It can render one blind to all facts and events that seem bad for one's party and one's cause. But accomplishing anything ambitious in politics requires loyalty because policies and programs take time to be worked out, and those opposed to the policy or program will not relax their opposition while things unfold. Loyal partisans are patient in their support for their candidates, their plans, and their programs (Muirhead 2013). Bonotti (2017, pp. 16-17), by contrast, argues that partisans should not be loyal to their party. They should rather think of themselves as having obligations to civility and fair play to other partisans, even rival partisans. The ethical obligations of fair play are stringent side constraints on the partisan pursuit of power-not least constraining what they can do out of loyalty to their own party.
Both with respect to the ethics of partisanship (e.g., the quality of inclusiveness) and the side constraints that restrict what partisans qua partisans should do in order to win, partisanship entails a commitment to constitutional democracy. In particular, it entails a commitment to pluralism, to parties as a vital institutionalization of pluralism, and to the idea of legitimate opposition. Without this commitment, the ethics of partisanship dissolve. Not even the best constitutional design or electoral system can create an incentive structure that reliably motivates partisans to act in ways that sustain democracy even as they seek to advance themselves.

CONCLUSION: THE DELEGITIMATION OF PARTIES AND PARTISANSHIP, AGAIN
Political theorists have come to parties and partisanship late in the day. The qualities that enlisted parties and partisans in the project of making democracy (Schattschneider 1942) may be disappearing; certainly representative democracy is under more stress than at any time since the 1930s. Understanding and navigating the challenges and threats of the moment requires political theory that engages political science-and vice versa (Van Biezen & Saward 2008, Herman 2017. Two specific considerations are needed in theorists' thinking about parties today. One is attention to new forces permeating the public sphere and undermining deliberation in the broadest sense-including the regulated rivalry of political parties. In addition to the usual menu of obstructions (self-interest, insufficient information, partisan bias, the distortion of dark money, and more), the forces of conspiracism and disinformation have taken center stage. They are creating an epistemic chasm among citizens as significant as partisan polarization, making it impossible not only to compromise and negotiate, but even to disagree (Muirhead & Rosenblum 2019). What is at stake is the norm of a legitimate opposition. The practices associated with this norm are always delicate, if only because the idea itself is so subtle. How can we view an opponent whose program we take to be mistaken, even catastrophically so, as also legitimate? The answer depends less on manners of respect than on the rituals and actions that signal loyalty to rules of fair competition and to the democratic constitution itself. But if epistemic polarization makes opponents impossible to understand and impossible not to perceive as a threat, rituals like concession speeches after an election loss will not be sufficient to fortify the norm.
The second, even more disturbing consideration is that the incentives to ethical partisanshipinclusiveness, comprehensiveness, and the disposition to compromise-are under stress, and in the absence of ethical partisans, they may be insufficient to sustain these qualities in the future. Partisans today often give up on inclusivity and pursue strategies that depend on disenfranchising opposition voters. Rather than offer a comprehensive account of the common good, they settle for activating their base. Instead of building a legislative coalition through compromise, they impede and obstruct "on principle," seeming to prefer symbolic stands to governing.
Partisans at their best avoid claiming the kind of legitimacy that comes from representing the people as a whole. At the same time, they strive to rule with democratic legitimacy by forging a large and lasting coalition. Is modern society so fragmented that the partisan quest to build a lasting (but never permanent) majority or coalition is out of reach? Or is it rather that the political virtues of coalition-building-inclusivity and comprehensiveness-have been displaced by the moral virtues of independence or uncompromising integrity, giving rise to a politics of selfrighteous postures? In the future as in the past, the challenge of making democracy work will fall to partisans who take on the task of making parties work.
Over the past 15 years, political theorists have identified and defended the normative terms of parties and partisanship. This has improved democratic theory. Our concern is whether these defenses can also improve our capacity to counter new forms of antipartyism that buffet democratic institutions.

DISCLOSURE STATEMENT
The authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.