Why Do We Need Another Conception of Partisanship?

I aim to illuminate a basic conceptual question: what is the nature of the relationship between individuals and political parties? Yet this is far from an unasked question. Because of the importance of parties and partisanship within representative democracies, constructing valid measures of this relationship is some of the most heavily contested ground in political science. I must therefore begin by explaining why we need another account of this relationship.

The treatment of party attachment I offer here is needed because its conceptualization receives surprisingly little attention from the partisanship theorists I examine and because existing operationalizations from political scientists do not take into account the lived experience of partisanship, creating risks to democracy. A natural division of labor within the discipline of political science might recommend that theorists have something to contribute to the formulation of basic concepts, yet this has largely not been a feature of existing work on partisanship in political science. White and Ypi are intensely concerned about the normative–empirical balance of their rational reconstruction of partisanship and provide a learned account of how political science has conceived of parties, yet they offer little discussion of how political scientists have constructed partisanship itself (White and Ypi Reference White and Ypi2016, 9–14). Rosenblum and Muirhead, meanwhile, adopt wholesale the conception of partisanship introduced by the Michigan school of political science that has become hegemonic in large segments of the discipline (Rosenblum Reference Rosenblum2008, 323–25; Muirhead Reference Muirhead2014, 43). Rosenblum says that “political science deserves the first word” on how to conceive of partisan attachment due to its long history as a central focus of the field, but leaves it practically as the last word when it comes to the fundamental conceptualization of partisanship (Rosenblum Reference Rosenblum2008, 323). Though these authors offer compelling rational reconstructions of what partisanship can be at its best, as I discuss later, they contribute little to a basic understanding of what precisely is being studied.

This has ceded the field of conceptual definition to empirical political scientists. Yet they have characteristic concerns, methods, and training that affect how they formulate and use concepts. In particular, many political scientists are content with “good enough” concepts that will serve as workable operationalizations for their specific purposes, which, in the case of party attachment, are often limited to explaining or predicting mass voter behavior and attitudes. For example, Keith et al. (Reference Keith, Magleby, Nelson, Orr, Westlye and Wolfinger1992, 193) assert—in an otherwise wide-ranging exploration of those who do not identify with parties—that “what matters is the role of attachment to a party in reaching voting decisions and forming opinions about policies and the performance of the incumbent administration.” Similarly, Kollman and Jackson (Reference Kollman and Jackson2021, 2) identify those for whom their novel operationalization of party attachment will be useful as “analysts of electoral politics,” whose understanding of partisanship’s dynamics will be enhanced, and “party strategists,” who can use it to improve their party’s strategic positioning and win elections.

Conspicuous in its absence here are citizens. Both analysts and strategists adopt the perspective of observers interested in understanding and perhaps manipulating party attachment, reflecting the deep behaviorist assumptions of much political science. Not so citizens. They view parties at least in part from the agent or actor’s perspective, as a participant in politics who is faced with taking actions and making choices, including how to relate to parties. Though Kollman and Jackson (Reference Kollman and Jackson2021, 4) acknowledge the important role of partisanship in helping citizens navigate the political information environment, they do not otherwise elaborate on citizens’ lived experience of partisanship. This raises acutely the question: who are our concepts for? The answer does not ordinarily extend much further than analysts and strategists. But changes in political journalism over the past decade or so ought to make citizens a much more salient audience.

Recent years have seen the emergence of a new, more data-driven and scientifically engaged form of political journalism that makes extensive use of political science. Political science blogs like the Monkey Cage, the Upshot, and Mischiefs of Faction (associated with the Washington Post, New York Times, and Vox.com, respectively) have directly published political scientists’ work and journalists like Ezra Klein (now at the Times), Matthew Yglesias (co-founder of Vox), and Nate Silver (at fivethirtyeight.com) extensively deploy political science concepts to explain current events to massive popular audiences, arguably helping transform political journalism more broadly (Klein Reference Klein2014; Drezner Reference Drezner2017). This is to say nothing of the burgeoning ecosystem of political science podcasts featuring scholars talking about their research and current events, some of which aim at niche audiences of influential policy makers (Grossmann Reference Grossmann2020).

One effect of this transformation has been to spread the dominant identity conception of partisanship from the American politics subfield into actual public discourse as the primary frame for understanding partisan attachment and, indeed, political conflict generally. I will argue later that this falsifies the experience of millions of citizens, encouraging the feeling that there is no place in politics for people like them. I also argue that the dominant approaches to party attachment occlude possibilities for relating to parties—and indeed for doing politics—differently, blocking up the imaginations of citizens and leaders to coalitional innovations that can unlock solutions to difficult political problems. Before I can elaborate these arguments, however, I must first discuss how I suggest we rethink party attachment.

Conceptualizing Partisanship

So, to take up the challenge: what is partisanship, stripped to its basics? It is an account of the relationship between the individual and the social object or group of the party. What I’m going to do now is apply the political theoretic tool of conceptual analysis to two widely used survey instruments and excavate two distinct conceptualizations of the relation between individuals and parties implicit in them. This analysis will not necessarily reflect how researchers have interpreted these instruments—indeed, I suggest an approach that is at odds with how many researchers think of party attachment. I address these concerns in the next section.

I begin by analyzing the American National Election Study’s (ANES) three-item party identification battery and how it is conventionally translated into a unidimensional scale (American National Election Study 2021). It asks respondents:

1. 1. Generally speaking, do you usually think of yourself as a Democrat, a Republican, an independent, or what?

2. 2a. (If Respondent considers self a Democrat/Republican): Would you call yourself a strong [Democrat/Republican] or a not very strong [Democrat/Republican]?

3. 2b. (If Respondent is Independent, No Preference, Other, Don’t Know): Do you think of yourself as closer to the Republican Party or to the Democratic Party?

These questions are usually understood to generate a unidimensional measurement of partisan attachment on a 7-point scale such that determining partisan attachment is simply a matter of locating where citizens belong along this one continuum. Importantly, this continuum is one of identification since it is scaled as if all points along it are degrees of identification with the parties; see figure 1. This scale is the measure of party identification used in nearly every study of American politics. Indeed, Richard Johnston says it is “probably the most highly leveraged measure in all political science” (Johnston Reference Johnston2006, 347).

Figure 1. The 7-point scale derived from the ANES 3-item partisanship survey instrument

Figure 2. An illustration of the plurality of partisanships

On this scale, partisan attachment is a matter of stronger or weaker self-identification with a party, indicating that one’s self-conception includes belonging to a party (or not—I address independent voters later). It is thus about identity, and this is why the main term for party attachment in the United States is party or partisan identification—that one identifies oneself with the party. One description explains that it is “entirely a matter of self-definition” (Green, Palmquist, and Schickler Reference Green, Palmquist and Schickler2002, 26). The relationship between the individual and the party here is thus one of coincidence—the individual absorbs the party into their sense of self and uses it as a basic building block of their social identity. ^{Footnote 1}

Coexisting with this way of relating to a party is another suggested in the ANES’s follow-up question asking whether those who do not identify with a party nonetheless “think of themselves as closer” to one. This wording suggests that one’s psychological relation to the party is one of proximity or “closeness” rather than identity (Barnes et al. Reference Barnes, Jennings, Inglehart and Farah1988). This concept of closeness is the central one in most comparative surveys of individuals’ party attachment. The Comparative Study of Electoral Systems (CSES), for example, uses the following question battery to gauge party attachment (Comparative Study of Electoral Systems 2022):

1. 1. Do you usually think of yourself as close to any particular party?

2. 2a. (If Respondent answers “yes” that they think of themselves as close to a party): Which party do you feel closest to?

3. 2b. (If Respondent answers “no” that they do not think of themselves as close to a party): Do you feel yourself a little closer to one of the political parties than the others?

In these questions, the relationship between the respondent and the party is exclusively conceived in terms of degrees of psychological closeness. On this “closeness” view, far from being identical, the individual and the party are seen as distinct but set at some metaphorical distance from each other, and this distance might be greater or less. This differs fundamentally from identification, where the metaphorical distance is zero, since the individual and the party coincide—they are identical. In the closeness conception, the individual’s self remains distinct from the object that is the party; see figure 2. Their relation is one of affinity or affect (Iyengar and Westwood Reference Iyengar and Westwood2015, 704). ^{Footnote 2}

I am suggesting that these survey instruments imply two categorically distinct ways we might conceive of the relationship between citizens and parties. One sees it as identification while the other constructs it as closeness, giving rise to two distinct kinds of partisans: identity partisans and closeness partisans. When speaking of the party, identity partisans characteristically say, “we;” we Democrats, we Republicans, here’s what we’re doing, here’s what we’re about. Closeness partisans by contrast will tend to say, “they” about the parties and their adherents. To grasp the significance of this difference, consider that to identify with something is a unique relation likely to influence a subject powerfully. When a subject considers an object identical with itself, the subject will find it difficult to treat the object impartially and may perceive threats or triumphs associated with the object as applying to themselves as well, triggering emotional reactions and motivating actions with power and immediacy. For identity partisans, politics feels personal. If an attack on your party feels like an attack on you, you might be an identity partisan. Closeness partisans will not have these reactions. The distance between their identity and the party means politics will seem rather more impersonal to closeness partisans, as if it is not about them in a direct way. They are not personally implicated in politics and lack the emotional investment of identity partisans. Though they may care about the fate of their party and its policies, this will lack the energy and immediacy found among identity partisans.

This gap in personal psychic involvement with parties is essential to understanding citizens’ different experiences of partisanship. Those on one side of this elemental divide often find the behavior of those on the other bizarre (Krupnikov and Ryan Reference Krupnikov and Ryan2022). Identity partisans may find closeness partisans inexplicably blasé about important issues or events, for example, while closeness partisans may see identity partisans as pathologically obsessed with politics. If we want our conceptualization of party attachment to capture the diversity of what it is like to be a partisan, from inside the participants’ perspectives, I argue we should make this distinction between identity and closeness partisans—between those who say “we” and those who say “they” about the parties—central to it.

Empirical scholars have conceptualized partisanship in three main ways: as a heuristic running tally of parties’ issue positions and performance; as itself a social identity; and as a channel for other social identities (Huddy and Bankert Reference Huddy and Bankert2017; Orr and Huber Reference Orr and Huber2020). How do identity and closeness partisanship relate to these? There appears to be considerable overlap in the use of the term “identity”, yet this is misleading because the distinction here is mainly tracking the attachment style—what is it like to experience partisanship?—whereas these conceptions focus on what, exactly, partisans are attaching to—is it the parties’ policies, or its brand, or the social groups associated with it? So, one might either identify with or feel close to a policy position (anti-abortion, etc.), or a partisan group as such, or an associated social group. The distinction is thus orthogonal to much of the existing empirical debate yet illuminates an important dimension of partisanship all the same.

In this section, I suggested that there is a plurality of partisanships implicit in existing survey instruments that is constituted by at least two distinct attachment styles between individuals and parties: closeness partisanship and identity partisanship. In the next section, I explain why this distinction is likely to matter for political science, public debate, and the practice of partisanship.

The Difference a Plurality of Partisanships Makes

As mentioned earlier, measuring and conceptualizing party attachment is among the most contested ground in political science, so there are numerous objections to this approach. All agree that there are some who hold tightly to political parties and some who do not, so the real question is how we make sense of—and talk about—this difference. I am suggesting that at least part of it is a categorical difference between closeness and identity partisanship. Yet most operationalizations generate a continuous variable, such as the 7-point scale presented earlier, which suggests the metaphor of “strength” to discuss differences in party attachment. Moreover, some scholars construct social identity in politics as itself a measure of closeness to various social groups, collapsing the distinction I draw (e.g., Shayo Reference Shayo2009). In the former case, identity effaces the concept of closeness by replacing it with the language of strength, and in the latter, closeness replaces identity itself. Both eliminate any categorical difference of the kind I suggest.

The problem with both of these approaches is that, while they formally recognize the difference among partisans, they effectively homogenize it in ways that falsify the diversity of citizen experience and impoverish our imaginations about how politics can be done. Consider first the longstanding coexistence of closeness survey measures of partisanship like the CSES’s and identity ones like the ANES’s. If either closeness or identity were sufficient on its own, one would expect scholars to herd toward whichever presented the best measure. Instead, we see the persistence of both. Why? There might be many reasons, but I suggest it is in part because each presents a compelling account of how some citizens relate to parties and can be plausibly fudged to apply to the rest. So, although the ANES battery reveals that many citizens identify with a party, it remains the case that others don’t. Yet they can be approximately captured with the language of Independence,Footnote ³ “leaning”’ or being less strong (or secret) partisans. In the CSES case, all are asked whether they feel close to a party, and those relating to parties that way are correctly denoted as such. Meanwhile those who identify with a party can just be scaled as if they are very, very close to it, even though that does not accurately reflect how they feel. What both approaches do therefore is represent a fundamental difference as homogeneity, which is why I say they homogenize the difference between partisans.

One might respond that I am begging the question by assuming there is a meaningful difference that cannot be captured on a continuous scale For example, Kollman and Jackson (Reference Kollman and Jackson2021) find that even after major shifts on core policy positions, some members of parties persist in their attachment while others migrate to other parties. This pattern suggests to me that those who remain attached to the party after major policy transformation are likely to be identity partisans while those who depart will tend to be closeness partisans. But of course one could also say remainers were closer to the party than departers, or more strongly identified with it. Something similar could be said about American independents—I’d say independents are mostly closeness partisans, while others could say they’re weak identifiers (Keith et al. Reference Keith, Magleby, Nelson, Orr, Westlye and Wolfinger1992). Most pieces of evidence that I could highlight to support my point admit of such alternate interpretations; Orr and Huber (Reference Orr and Huber2020, 571–72) identify this observational equivalence as a major challenge in the conceptualization of partisanship. It is therefore tempting to soften my claim to the ecumenical one that different purposes call for different conceptualizations; a continuous one might make sense for understanding mass behavior while a categorical one might be needed to reflect how citizens experience partisanship. Yet insofar as observational equivalence gives us discretion in choosing how we operationalize partisanship, I insist that we should distinguish closeness and identity partisanship.

This is because the homogenizing approaches to party attachment obscure possibilities for how citizens might relate to parties, failing to recognize their agency as citizens to reflect and make choices that do not fit the schema. The party identification scale, for example, appears to be unidimensional but actually shoehorns together many disparate considerations, suppressing possibilities such as a principled affirmation of nonpartisanship, general anti-party preferences, or having only a negative view of a party and no corresponding positive one to a different party (Rosenblum Reference Rosenblum2008, 319–68; Barnes et al. Reference Barnes, Jennings, Inglehart and Farah1988; Johnston Reference Johnston2006, 347; cf. Klar and Krupnikov Reference Klar and Krupnikov2016). Moreover, some non-identifiers might be cross-pressured by having negative judgments of all the existing parties, either due to particular issue positions or social affinities—something possibly true of those who identify as “not very strong” partisans as well. Finally, we have already seen how the closeness approach forecloses the possibility of relating to parties as identities (at least, in a non-tautological way), but it also omits from view principled anti-partisanship and nonpartisanship. Thus, not only do these interpretations of party attachment not tell us about these perhaps idiosyncratic ways of relating to parties, they actively obscure many such possibilities.

Recall that political science has entered and come to shape the public discourse in dramatic ways in recent years. As journalists and other news sources have increasingly turned to political scientists to help them understand U.S. politics, the concepts the latter use have infiltrated news sources, helping to frame popular understandings of what is happening. Such framing effects powerfully shape public opinion (Chong and Druckman Reference Chong and Druckman2007; Baumgartner, De Boef, and Boydstun Reference Baumgartner, De Boef and Boydstun2008). Party identification has been foremost among these disseminated concepts, arguably spreading throughout American public political culture. I claim that party identification has, in the form it has entered American political discourse, helped frame the practice and understanding of party attachment in the United States in democratically harmful ways.

Consider first that those who do not relate to parties in terms of identity are left out of the political drama as it is reported in the news every day. That drama features a “cast of characters” and, on every channel, the cast is the same—(strong) Democrats and (strong) Republicans—occluding anyone not firmly within one of those camps. I suggest that this false bifurcation can lead to alienation, as closeness partisans come to feel that there is no place for partisans like them in American politics. Due especially to the lack of viable parties beyond the two-party duopoly, this exclusionary understanding of their place (or lack thereof) in politics can help demobilize them and attenuate their attachment to politics itself as democratic citizens, harming fundamental democratic values of equality and inclusion (Elliott Reference Elliott2023).

Moreover, by representing partisanship in Manichean terms and occluding other possibilities for how one might relate to parties, party identification may encourage bad practices of partisanship and block up imaginations regarding strategies and coalitions that might help escape negative or excessive patterns of partisan conflict. Partisanship is, among other things, a performance; it is a collection of habits, including habits of speech, that partisans enact in daily social interactions. They learn how to perform their partisanship by looking at the people around them, especially co-partisan leaders. When the most socially available way to enact partisanship reflects an exclusively identitarian conception, citizens may inadvertently lock in ways of behaving toward each other that reflect the zero-sum conflict that comes naturally to Manichean struggles, such as affective and social polarization, which consist in intensified feelings of hatred and contempt for partisan opponents (Iyengar and Westwood Reference Iyengar and Westwood2015; Mason Reference Mason2018).

These destructive dynamics are likely to be harder to get off the ground when the plurality of partisanships is widely recognized because they seem to be driven by the specific psychological mechanisms triggered by identity. Mason shows how social group identities, once mapped onto politics, generate escalating identity-based conflict that subsumes all substantive political agreement and compromise while triggering cognitive bias, prejudice, and emotional volatility (Mason Reference Mason2018). Political sectarianism in this sense has been rising dramatically in recent years in the United States (Finkel et al. Reference Finkel, Bail, Cikara, Ditto, Iyengar, Klar and Mason2020). This is alarming because it transforms partisan competition into scorched earth, zero-sum politics, leading to constitutional hardball and democratic erosion, as partisans prioritize defeating the other side above all else (Levitsky and Ziblatt Reference Levitsky and Ziblatt2018).

Highlighting the possibility of closeness partisanship in the public discourse could blunt these forces, if slightly, by expanding the public imagination regarding the nature of the partisan tie and so providing an alternative to mentally dividing the political world into starkly divided camps. Instead, it is rendered as a tapestry of different ways people relate to the parties. One would no longer have cognitively available just one way to think about partisanship, avoiding a narrow stereotype of it and diversifying the imagination in inclusive ways about what partisanship can be. This opening of conceptual space might operate as a break to political sectarianism since it could generate multiple political cleavages—making it harder to imbue any with existential threat—thereby reconfiguring and diversifying expectations about what it means to be a partisan.

Whereas the danger of alienation is largely a problem afflicting the mass public, the channeling of our ways of thinking and behaving in zero-sum ways affects both the mass public and elites. When elites adopt this way of thinking about competing partisans, it doesn’t just threaten to escalate political conflict—it can blind them to off-ramps from such intensification spirals by occluding possibilities for strategic cooperation or coalition. For instance, recent changes in the Republican Party have pushed a non-negligible group of former party officials and intellectuals out of the party’s orbit, potentially opening the way for coalition with Democrats. Yet the past associations of these former Republicans can make this difficult, particularly for Democrats whose partisanship formed partly in opposition to the administrations these former officials served. In light of the January 6, 2021, insurrection, however, saving American democracy might necessitate forging coalitions between a variety of kinds of citizens, including those who formerly aided what one regards as oppression, in order to divide such unintentional or well-meaning oppressors from those who revel in it (Cheng Reference Cheng2022).

Recognizing a plurality of partisanships might help provide both conceptual and social space for such negotiations. Though closeness and identity partisanship will not perfectly reflect the complexities I’ve been discussing, they would help provide greater conceptual space within which to define one’s relationship to parties, as discussed earlier. Recognizing the distinction may also open shared social spaces wherein we can relate to the parties and their adherents using tools other than those of identity. Whether these are online or in-person spaces, knowing that the person sharing your party affinity is merely close to the party, whereas you identify with it, can help reorient internal coalitional discussions by pluralizing the party’s understanding of itself, as I discuss further later.

Readers of an empirical bent may wonder what evidence I have for the claims of the last paragraphs. I posit them in part as suggestions informed by theory, ready to be falsified or confirmed insofar as possible. Yet I also intend them partly as ethical suggestions for how citizens might occupy the role of partisans—there might be reasons to identify with a party just as there might be ones merely to be close to one. Exploring these ethical and more thickly normative dimensions of the plurality of partisanships is the task of the next sections examining the political theory of partisanship.

I do not intend to blame political scientists for the dramatic rise of partisan conflict in American politics; that would ascribe them far too much influence. My point is rather that, due to the enhanced influence of political science in public discourse, political scientists should strive to make their contributions to that discourse with some consideration as to its effects. When their concepts become public and influence the actions of those whom they study, they should think about how those concepts might shape the choices of actors in politics.

Thus clarified, my claim raises the powerful objection that it is not the responsibility of scholars how others use their concepts—they are not to blame for misuse. Not only is it not their problem, but it is impossible to anticipate all the ways that one’s ideas might be abused by others. The task of scholars should therefore be the narrow one of trying to understand the phenomena they investigate. They thus ought to proceed as if the only thing that concerned them is the immediate subject of their investigation.

This is a sound default, and, properly understood, my contention does not contradict it. By no means am I suggesting that concerns about public uptake should always or decisively shape how scholars construct concepts, particularly not when doing so would involve misrepresenting the object of study. That line of argument would turn political science into ideology. My claim is much more limited and applies specifically to the case of partisanship, where there is a clear case on the merits for reconceptualization due to the dominant approaches’ falsification of citizens’ lived experience, with potentially deleterious consequences for democracy, and the presence of a ready alternative that is not obviously worse than existing approaches.

My point is that when the political stakes are high and the choice is between equally plausible conceptualizations, political scientists should then consider how their concepts are used in political debate. It is thus not about speculation regarding how they might be misused, nor is it about sacrificing fidelity for political utility. It is about adding a consideration into the balance to break an apparent tie between different approaches.

In saying this, I do not mean to suggest that changing our words about partisanship can alone heal the divisions afflicting politically sectarian environments. Such naïveté ignores the powerful structural incentives that reward political actors for furthering it (Drutman Reference Drutman2020). Enduring amelioration of political sectarianism likely requires tackling those structural features through reform. Nonetheless, words matter. Regardless of the fate of American democracy, we would do well to remember that, with respect to partisanship, speaking of the plurality of partisanships is likely to achieve for democracy what is best in partisanship, as we shall see.

The Value of Partisanship in Political Theory

Political theorists have recently come to a new appreciation for the importance of political parties and partisanship in democracy. Works exploring the democratic functions of parties (Disch Reference Disch2002; Rosenblum Reference Rosenblum2008; Goodin Reference Goodin2008, 204–23; Muirhead Reference Muirhead2014; Rosenbluth and Shapiro Reference Rosenbluth and Shapiro2018; Landis Reference Landis2018), and of partisanship (Rosenblum Reference Rosenblum2008; Muirhead Reference Muirhead2006; White and Ypi Reference White and Ypi2016; Efthymiou Reference Efthymiou2018), have been joined by others showing how parties and partisanship can help manage pluralism (Bonotti Reference Bonotti2017; Bellamy et al. Reference Bellamy, Bonotti, Castiglione, Lacey, Näsström, Owen and White2019) as well as a burgeoning literature investigating whether parties should be internally democratic (Wolkenstein Reference Wolkenstein2016; Invernizzi-Accetti and Wolkenstein Reference Invernizzi-Accetti and Wolkenstein2017; Wolkenstein Reference Wolkenstein2020; Bagg and Bhatia Reference Bagg and Bhatia2021). So how have political theorists conceived of the value of partisanship for democracy? This section outlines three distinctive views of partisanship’s value found in the work of Rosenblum, Muirhead, and White and Ypi, as well as a set of systemic virtues highlighted in all three of these otherwise distinct approaches.

These three accounts can be seen as rational reconstructions of partisanship that seek to demonstrate its value in democratic regimes (e.g., White and Ypi Reference White and Ypi2016, 3–4). Yet these reconstructions have neglected to theorize the fundamental relationship at the heart of the concept of partisanship—the nature of the relationship between party and individual. This is not a small point, but the theorists can be excused for neglecting it. Their interest is not the conceptual analysis of partisanship, but rather its value. Overlooking the conceptual question, however, involves the assumption that partisanship is homogenous—consisting of just one kind of relationship rather than as a class of more or less defensible kinds of attachment. I aim to question this assumption because it turns out that the plurality of partisanships has enormous consequences for these theorists’ rational reconstructions.

For Rosenblum, the core of partisanship’s moral distinctiveness is its institutionalization of pluralism. Pluralism, and the variety of ways it is manifested and accommodated, is the central theme of Rosenblum’s scholarship, and here she is interested in how partisanship gives it institutional form. Partisanship is a way, she says, to make manifest that the group that loses in an election does not disappear or go underground but remains publicly visible, palpably instantiating the loyal opposition (Rosenblum Reference Rosenblum2008, 357). Partisanship institutionalizes this persistent pluralism by promoting three things: inclusiveness, comprehensiveness, and the propensity to compromise. In these ways, partisanship generates patterns of behavior and carves out institutional spaces in which democratic pluralism can flourish, making it an invaluable part of democracy.

The first way is through inclusiveness. No other kind of political belonging includes as many different people from as many different segments of the population as partisanship. This internal diversity extends to differences in ideology, as well as to disagreements over which issues should receive priority in the party’s platform and governing, making the party internally heterogeneous. This inclusiveness is generated by the political incentives of electoral competition. Parties must “tolerate or welcome diversity” wherever they “are ambitious to be in the majority” (Rosenblum Reference Rosenblum2008, 357). Majority coalitions must be inclusive, and thus, so must partisanship. ^{Footnote 4}

The second way partisanship promotes pluralism is through encouraging the telling of comprehensive public stories. Partisanship helps constitute partisan groups that have a responsibility to tell a comprehensive public story about the polity and what policies it should pursue. This does not mean the party has a comprehensive theory of justice or anything as philosophically ambitious as that. Partly as a function of its inclusiveness, the party must instead tell a story about where the polity is and where it should go that could persuade anyone—not just a narrow segment of the population—to support it. When individuals identify with a group putting such a story forward, they become invested in that story, to improving it or holding the group to it when necessary.

The drive to tell a comprehensive public story is, under other descriptions, emphasized by the other two approaches as well. According to Muirhead, the drive to win contributes to deliberative democracy by moving partisans to offer reasons that appeal broadly (Muirhead Reference Muirhead2014, 89). These reasons are a primary supply of material for political debate and so they constitute the bread and butter of deliberative democracy in practice. White and Ypi also make much of this feature of partisanship, yet I discuss it later with the rest of their approach due to its distinct centrality for them.

The final way that partisanship institutionalizes pluralism according to Rosenblum is that it entails a “disposition to compromise”—but only with fellow partisans (Rosenblum Reference Rosenblum2008, 374). She argues that there is ongoing deliberation within the party as to its future direction and regarding who is inside and who outside the bounds of the party. Settling these questions requires compromise among the diverse groups comprising the party, so parties end up cultivating the ability to compromise among their members. Rosenblum notes, moreover, that the other two attractions of partisanship, inclusiveness and comprehensiveness, are only possible with this disposition since it enables the toleration of diversity and stimulates the articulation and re-articulation of the party’s public story.

The second major account of partisanship’s value is that of Muirhead. For him, the value of partisanship comes from its blending of two different tendencies, both of which are essential for electoral democracy. These are ambition for power, on the one hand, and constraining principle, on the other. Each tendency unalloyed is fatally problematic. Ambition without principle makes politics into zero-sum conflict, undermining democratic norms which are cut down in pursuit of victory, while principle without ambition is politically sterile. The partisan thus combines the power seeker’s ambition with the purist’s limited flexibility on principle, blending “the pragmatic and the principled,” while checking the former’s endless flexibility and the latter’s unbending perseverance even in the face of defeat (Muirhead Reference Muirhead2014, 42). Partisanship at its best, then, is an ongoing balancing act between ambition and principle, defining the narrow middle path of political virtue.

Muirhead thinks that the way partisanship shapes politics will therefore have the function, by turns, of blunting the cutthroat tendencies of ambition and dampening rigid adherence to principle. Since these tendencies are a permanently recurring problem, partisanship becomes an invaluable addition to the political landscape. Indeed, politics may be impossible without it.

Partisanship also provides a necessary feature for deliberative democracy: a “negative capacity” to distance oneself from one’s commitments (Muirhead Reference Muirhead2014, 106). This quality is essential for deliberation because it allows one to reformulate one’s argument to better persuade one’s audience (as well as to reconsider one’s own view). It is counterintuitive, however, for Muirhead to claim that partisans have this quality, since an important part of the stereotype of partisans is that they hew to their party’s views and cannot be budged from them. Yet this ignores the dual nature of partisanship—that it blends principle and ambition. Because partisans are driven by the ambition to win, they must be able to distance themselves from their principles at times to reconceive them. When their principles are unpopular in their current form, flexibility or re-articulation is often necessary to attract political support. Partisanship at its best, then, strongly incentivizes the capability to step away from one’s views and consider them from another perspective, even if this is in service of victory rather than finding the truest view.

This negative capacity is also key because it allows for compromise with non-co-partisans to achieve a portion of their idea of the common good (Muirhead Reference Muirhead2014, 53). Note the distinction with Rosenblum’s argument. Rosenblum argues that partisanship promotes habits of compromise, but only with co-partisans. Muirhead is claiming that the ability to abstract from partisan commitments enables compromises with opponents. The negative quality makes such partial victories possible by weakening the intensity of the partisan drive toward total victory.

Perhaps the most telling characterization of Muirhead’s idea of partisanship, and the one most salient for the discussion later, is that it should be “worn lightly” (Muirhead Reference Muirhead2014, 17). This suggests the metaphor of a garment, one that can either be pulled tight to—or draped loosely around—oneself. A garment worn lightly is easily shed, and so Muirhead is suggesting that partisanship at its best is defined by being held loosely.

The third distinctive account of partisanship’s democratic value is that of White and Ypi. For them, the value of partisanship lies in its enabling collective self-rule. Partisanship is fundamental, they argue, to sustaining collective commitments, without which there can be no democratic self-rule. Shared commitments, and sustaining them, is at the core of their account. Partisanship does this because, for them, it is best conceived as “an associative practice with the purpose of promoting and sustaining principled projects” (White and Ypi Reference White and Ypi2016, 143). Such projects are the outcome of interpretations about how public power should be used (White and Ypi Reference White and Ypi2016, 21). The practice of partisanship thus consists in sharing a set of commitments to advance principled projects that are pursued with others. These commitments, in turn, ground exercises of both individual and collective political agency by providing the vital directional guidance that is necessary for any purposive action. Relatedly, the commitments also provide the material on which citizens exercise their autonomy, affording them “authorship” over their lives (White and Ypi Reference White and Ypi2016, 82). Because of this connection to commitment, agency, and autonomy, partisanship is essential for collective self-rule, which is for White and Ypi the core democratic ideal.

Articulating a comprehensive public story, or in their terminology, “generalizable principles and aims” is an especially central part of partisanship for White and Ypi. This is partly because the commitments around which their account of partisanship is organized are constructivist—they are the outcome of deliberative processes aimed at deciding what is to be done in one’s polity should one’s party come to power. In determining for themselves what the party is about, partisans formulate the generalizable principles and aims that the party then uses in electoral messaging and persuasive appeals. This makes the principles and aims publicized by the party the nexus of collective self-rule since it constitutes the party’s collectively-determined account of how the polity should rule itself. Though White and Ypi do not emphasize intra-party democracy as have other theorists (Invernizzi-Accetti and Wolkenstein Reference Invernizzi-Accetti and Wolkenstein2017; Wolkenstein Reference Wolkenstein2020; Bagg and Bhatia Reference Bagg and Bhatia2021), the importance of internal deliberative processes to determining how the party should rule makes clear the special importance of generalizable principles and aims to their conception of partisanship.

Another reason generalizable principles and aims are central to White and Ypi’s account is because (they assume) voters choose which party to vote for based on them. Through this choice, voters convey their judgment about how the polity should be ruled. It is therefore through the parties’ principles and aims that the people collectively rules itself. It represents both the internal deliberative processes of partisans and the external processes by which the wider polity adopts or rejects parties’ public projects (White and Ypi Reference White and Ypi2016, 149–51). Again, this importance is unique to White and Ypi because they make collective self-rule the central democratic principle.

That exhausts the three distinctive theoretical perspectives on partisanship’s value. Yet in the overlaps of these accounts are found three other functions served by partisanship that make it systemically valuable for democracy. The first of these systemic virtues is that partisanship stimulates electoral participation, providing an animating spirit to this vital but easily undervalued form of participation. According to Rosenblum, partisanship helps us vote by making voting come alive (Rosenblum Reference Rosenblum2008, 354). It expands our imagination regarding what we are doing when we vote by situating us among a group of like-minded people, acting together to win an election. Both Muirhead and White and Ypi also praise this mobilizing function of partisanship (Muirhead Reference Muirhead2014, 111; White and Ypi Reference White and Ypi2016, 89).

The second systemic virtue of partisanship is that it provides stability and continuity to parties and electoral competition. It does this because party elites are often so strongly motivated to win elections that they are willing to ignore the party’s principles to achieve short-term victory (Rosenblum Reference Rosenblum2008, 355). Partisans in the mass public, however, are in it for the long term and bear memories of what the party has been about in the past (Muirhead Reference Muirhead2014, 128–30). Since their support is essential for electoral victory, this long-term orientation makes them a force for integrity in the party, limiting the flexibility of party elites to follow short-term political considerations. Partisans carry the flame of the party, even when electoral precarity causes it to flicker. White and Ypi emphasize this virtue in terms of stabilizing the commitments that unite partisans over time (White and Ypi Reference White and Ypi2016, 122–41). For all three approaches, this stability is a core feature—and virtue—of partisanship.

Partisanship also enhances the collective nature of democratic politics by constituting groups of actual persons with whom one stands, rather than for solely abstract principles, promoting what we might call, adapting Muirhead, “standing-withness” (Muirhead Reference Muirhead2014, 90). Standing with others goes to the heart of political motivation, interest, and engagement. Some might even call it the most fundamental part of any democratic politics. Partisanship is imbued with loyalty to a concrete group of people, even if it is mediated by principles or commitments (White and Ypi Reference White and Ypi2016, 76). This loyalty is also an important part of the reason that partisanship stimulates electoral participation, since loyalty is the spirit that invigorates it (Rosenblum Reference Rosenblum2008, 354). We vote not only to further our principles but also to better our group. We are not responsible only to our conscience to further our commitments through participation, but also to other actual, identifiable persons, who may importune us to follow through. This provides a concrete and external source of accountability for, and spur to, political action. The loyalty of “standing with” others generated by partisanship is thus a powerful source of selective solidarity and political motivation, further boosting the endurance of parties as well as democratic engagement.

We can think of these final three functions of partisanship specifically as systemic virtues that support the regulated rivalry that is the core of electoral representative democracy (Rosenblum Reference Rosenblum2008, 362). They thus accomplish things that any theory of electoral representative democracy would require, including, as their overlapping endorsement suggests, all three distinctive theories of partisanship.

The Plurality of Partisanships and Democratic Value

How does thinking about partisanship as a plural set of partisan attachments, including closeness partisanship, impact the value of partisanship? We shall see in this section that it has distinct and dramatic effects on the three conceptions of partisanship’s value. For Rosenblum, recognizing the plurality of partisanships promotes all its core values. Identity partisanship is largely incompatible with White and Ypi’s account because identity partisans do not relate to parties via commitment. For Muirhead, closeness partisanship makes excellent sense of what it would mean to wear partisanship lightly, but identity partisanship stands in marked tension with such ease. Identity partisanship is also likely to disrupt Muirhead’s core balance between ambition and principle. Overall, the plurality of partisanships—and closeness partisanship in particular—seem to illustrate what these theorists cherish most about partisanship.

Let us begin by considering the impact on the three systemic virtues of partisanship: participation, stability, and “standing-withness.” A diversity of partisan attachments would likely weaken the power of partisanship to motivate electoral participation since the attachment to the party in closeness partisanship is attenuated and less imbued with threat than partisan identification. This might seem a serious concern. Yet we can also see the reduction in feelings of threat and danger as meliorative because such feelings surely have a deleterious effect on both the tenor and substance of political conflict. Moreover, there are other ways to boost electoral participation. These include institutional means, from cost-reducing initiatives such as same-day voter registration (Grumbach and Hill Reference Grumbach and Hill2022) to turnout-maximizing institutions like mandatory voting (Hill Reference Hill2004). They also include increasing electoral competition through reforms that encourage multi-party competition, such as fusionism and multi-member districts (Disch Reference Disch2002; Drutman Reference Drutman2020). By giving citizens a wider variety of choices, democracy would provide citizens with options closer to their ideal points and stimulate participation by those who currently feel shut out of the two-party system (Elliott Reference Elliott2023, 176–77). There are, in other words, alternative ways to promote participation that are both compatible with the diversity of partisanship and that allow us to reap benefits from that differentiation without sacrificing participation.

What of stability? As discussed earlier, stability here is stability within the party, regarding what it stands for, as well as in wider politics, since internally stable parties make for stable political choices. Stability in this sense might be reduced insofar as closeness partisans would be more open to changing what the party stands for since they are less closely aligned with it in its current form. Yet it is not clear that this kind of stability is always valuable. In times of political turbulence, the inability of the parties to change creates stresses on the political system as demand for change seeks outlets outside the system, empowering anti-system forces. Sometimes the conflicts that have long defined a polity’s everyday politics need to be displaced.

In these periods, it is vitally important to have a way out of the settled party arrangement. This could happen in at least two ways. New parties can help re-articulate conflicts, as can party realignments. Yet these are made more difficult by partisans sticking tenaciously to their party lines and the articulations of political conflict baked into them. A more relaxed connection with the party would enable change to occur more easily. Closeness partisanship thus presents a sort of insurance against the risks of political turbulence by suggesting a more flexible relationship to one’s party. Thus, stability is actually best to have in moderated form, as the mix of partisanships likely provides. ^{Footnote 5}

The plurality of partisanships would not, on balance, much affect the practice of standing with others in politics. On the one hand, for the same reasons that closeness partisanship would likely weaken the motivation to participate, it could also reduce the sense of standing with others. On the other hand, closeness partisanship is still partisanship, and so entails feeling some degree of solidarity with others, even if not of the same type as in partisan identification. Standing with others does not seem like an especially demanding type of belonging, barring special circumstances. It seems compatible with many different motivations, degrees of commitment, and types of relationship to the party, just as we would find in any crowd of actual protesters who are literally standing together. Such crowds reflect many different motives and types of commitment, from none at all to life-shaping. Insofar as the feeling of standing with others can be said to apply even when the connection with those others does not reach identity, then, it is compatible with closeness partisanship. It may not even be attenuated by wider adoption of a closeness relation among partisans. In fact, insofar as inclusiveness is bolstered by recognizing both types of partisanship, as I argue later, it may enhance the collective capacity to stand with others, by diversifying our idea of who qualifies as a true ally.

Moving on now to Rosenblum’s account, recall that for her, the core moral value of partisanship lies in its institutionalization of pluralism through the promotion of inclusiveness, comprehensiveness, and the disposition to compromise. The introduction of a plurality of partisanships seems promising for an account like Rosenblum’s which prizes pluralism, and this proves to be the case when we consider each of these three values in turn.

A plurality of partisanships aids inclusiveness because it forces the parties to be more inclusive of a wider variety of ways of relating to the party. Instead of treating all partisans homogeneously, as if all of them relate to the party by way of self-identification, partisan plurality requires parties to take a more inclusive approach to its own members. Closeness partisanship in particular provides a non-pejorative name for a group essential for the party’s electoral success. I say non-pejorative because a common way to refer to those whose dedication to a party is less marked is often in terms of being soft, “squishes”, or, in the Republican party, Republican-In-Name-Only (RINO). Even the terminology of “leaners,” used widely in political science, denotes a lack or failing insofar as they are not represented as standing for something—with positive connotations of integrity and commitment—but rather as wishy-washy, like they can be shifted by the breeze. Providing a neutral nomenclature of closeness opens the door to an inclusive plurality within party coalitions, one in which none are second class.

This plurality also aids comprehensiveness because it is the variety of partisan attachments that move parties to formulate comprehensive public stories. Rosenblum, Muirhead, and White and Ypi all emphasize that partisanship encourages the telling of a comprehensive or generalizable public story about how power should be used. But this practice makes the most sense in a world composed of both identity partisans and closeness partisans. Such a story is unnecessary to entice identity partisans, who are already as committed to the party as can be. This commitment also renders identity partisans largely impervious to the appeal of other parties’ stories, reducing the stories’ utility as persuasive tools in a world composed solely of this kind of partisan. Instead, comprehensive public stories are told to attract and hold rapt those already in the orbit of the party, but not among the hard core of supporters. White and Ypi explicitly recognize the necessity of an audience of less connected, sympathetic citizens as “an important source of members-to-be,” the prospect of which “gives members reason to maintain the party’s ideational focus” or comprehensive story (White and Ypi Reference White and Ypi2016, 28). They do not, however, conceptualize this sympathy in terms of closeness partisanship. Yet this type of partisan attachment provides a clear conceptualization of what they are talking about here, and so makes for a good way to think and talk about this necessary reservoir of potential support that motivates practices of partisan justification.

Lastly, the plurality of partisanships aids the disposition to compromise with other partisans by further diversifying the party internally. This requires compromise to span a wider, more internally heterogeneous group. It is, in particular, the liminal quality of closeness partisans’ connection to the party that enhances the need for compromise. Because of it, closeness partisans constantly threaten the party with exit (Hirschman Reference Hirschman1970). We would expect this exit to take different forms in different systems; in multi-party systems, exit could take the form of transferring support to another party, for instance. In a two-party system, exit in the form of switching support is unlikely due to the social, psychological, and ideological distance it would involve. Yet this does not tie down closeness partisans. They can still exit, but in the two-party context this exit would often take the form of demobilization. Closeness partisans who feel the party moving away from them can opt out of political activity altogether, joining, in the US context, the nearly half of the electorate that does not turn out to vote even in presidential election years. Indeed, it is well known that turnout is lower in two-party systems than in multi-party ones (Powell Reference Powell1986), and I suggest this is in part because closeness partisans who become sufficiently distant from the party exit by opting out of active support.

There are serious costs to this form of demobilization, however. In particular, it leaves the party to be captured by identity partisans whose ability to think critically is diminished compared to the exiting closeness partisans or, as I argue later, who are susceptible to being trapped in downward spirals of either ambition or ideological rigidity. The coexistence of the two types of partisans within the party thus helps promote compromise by enhancing the outside option—exit—of a substantial enough part of the party’s coalition to force members to compromise with each other.

Switching tracks now to consider the impact of the plurality of partisanships on White and Ypi’s account, we find a dramatically different story. Identity partisanship is problematic for White and Ypi’s account because it is inconsistent with party attachment being based on commitment, which constitutes the core of their account of party attachment’s value. Partisan identification is a matter of social identity, and so is learned as part of unconscious processes of socialization and establishing group affinity (Green, Palmquist, and Schickler Reference Green, Palmquist and Schickler2002, 138–39). This implies that it is not, for the most part, reflectively embraced. Moreover, when identity partisans adopt their party’s commitments, it is usually merely a matter of following the party’s lead (Levendusky Reference Levendusky2009, 113), not reflective endorsement. It thus reaches no deeper into their hearts than someone donning the characteristic dress of an ethnic group, and is discarded just as easily. The point is that commitment follows identity, rather than the other way around, and so is epiphenomenal and easily discarded should the party change its mind. This cannot be consistent with White and Ypi’s valuation of democracy as a means of collective self-rule since commitments here are not chosen in the spirit of autonomy. Witness, for example, the overnight desertion of postwar conservatism by the Trump-era Republican Party. Republican partisans promptly dropped a whole host of longstanding commitments—from issues as disparate as free trade and free markets, to an aggressive foreign policy, to caring about leaders’ personal moral conduct, to constitutionalism itself—as fundamental to the party just because the new leader effectively told them to. This illustrates how, when push comes to shove and social identity clashes with commitment, it is commitment that is rejected, in precisely the opposite way to that desired by White and Ypi.

This is not offered as a criticism of White and Ypi’s account, of course, since they explicitly offer it as a rational reconstruction of a defensible and normatively attractive partisanship (White and Ypi Reference White and Ypi2016, 3–4). Rather, the point is to observe that any party that consists partly of identity partisans ought to be disfavorably evaluated by White and Ypi. If, as seems likely in light of the empirical evidence, the core of most parties consists of identity partisans, then White and Ypi’s account of partisanship may defend something that does not exist and is not likely to. Their account may end up, therefore, telling us that partisanship as it exists in the world is usually indefensible because it is moored to identity rather than commitment.

Closeness partisanship is much more consistent with their account because its hold on the individual is less thoroughgoing. Even if closeness partisanship is also a product of unreflective processes like socialization or social sorting (Mason Reference Mason2018), as seems likely, the critical distance between individual and party preserved by closeness partisanship allows for greater influence from commitment. This is to say, then, that democracies made up predominantly of closeness partisans would be promising settings for White and Ypi’s preferred species of partisanship. It is not clear if such democracies exist, however, but if they did, they are most likely to be found amidst multi-party systems which systematically discourage identity partisanship through their greater variety of more nearly attractive options for citizens. Thus, White and Ypi’s account seems best suited to a particular type of democratic arrangement—a multi-party representative democracy—and perhaps travels less well to two-party systems.

Muirhead’s account of partisanship’s value is in a similar situation to that of White and Ypi in that identity partisanship is less compatible with his vision of partisan value than closeness partisanship. The most straightforward reason is that closeness partisanship better embodies Muirhead’s vision of a partisanship that is worn lightly and that balances the ambition for power with a dedication to principle. The critical edge, or negative capacity, that partisans should be able to adopt toward their own party is blunted, however, by identity partisanship since it may render the adoption of critical distance from one’s party nearly impossible for many people much of the time. This is due to familiar biases and epistemic limitations that bar accurate perceptions about ourselves. Relations of identity invite strong emotional responses when those identities become politically salient, responses that can block effective communication and reflection. Likewise, through their psychic distance from the party, closeness partisans are in a better position to balance ambition and principle since they are not as subject to getting locked into one side or the other of this balance, as I explain in a moment.

Closeness partisanship supplies a desperately needed psychological explanation of how Muirhead’s account could plausibly work within the individual. Because Muirhead adopts partisan identification as a form of social identity, his theory puts itself in the position of positing behavior that is at odds with, as he puts it, “the partisanship that defines the moment in American politics” (Muirhead Reference Muirhead2014, 22). He frames his theory as a plea not for the false idol of non-partisanship but for a “better partisanship,” one that is self-consciously set against the facts of contemporary partisan practice.

Like White and Ypi, then, Muirhead is offering a rational reconstruction of an attractive partisanship, and such a normative project cannot be dismissed simply because it is at odds with the facts. Yet the internal coherence of the account is nonetheless compromised if it cannot offer a plausible psychological story of how an individual could inhabit it. Indeed, as mentioned earlier, identity partisanship sits uneasily with one of the core features of Muirhead’s account, the negative capacity to “see ourselves from a distance” and which involves turning a skeptical eye on “the commitments and loyalties that define our political orientation” (Muirhead Reference Muirhead2014, 106). These are precisely the things that political science tells us partisans fail to do, particularly when they are “strong” or identity partisans (Shapiro and Bloch-Elkon Reference Shapiro and Bloch-Elkon2008; Hannon Reference Hannon2022). This is where closeness partisanship can do significant work filling a gap in Muirhead’s theory. It supplies a plausible psychological connection between the party and the individual that naturally embodies this feature of his “lightly worn” partisanship. His invocation of the spatial metaphor—seeing oneself “from a distance”—is particularly telling here, since it is precisely how closeness partisanship represents the relationship.

Identity partisanship also creates problems for Muirhead’s conception because it is likely to disrupt the balance that partisanship is meant to strike between, on the one hand, ambition for office, power, or victory and fundamental, constraining principle on the other. Yet this is not because party identification is going to magnify the importance of one or the other in general but rather because identity rigidifies partisan tendencies, making them litmus tests for belonging, and so will tend to destabilize the balance wherever it is struck. In other words, identity locks in whichever tendency happens to predominate at a particular time, whether it is ambition or principle. This means that identity partisanship is not just uncongenial but rather an active danger to Muirhead’s account. Yet it is an indeterminate danger—it cannot be known in advance whether it will render a partisan group unprincipled power-seekers or inflexible ideologues. By bringing the party within the self, party identification makes both (or rather, either) more likely since, if the party tacks in either direction, anyone pulling the other way is more likely to be rendered a traitor and ejected. This leaves an even purer group to continue down the path to destruction either of democracy (if the party is power-seeking) or of the party’s electoral chances (if it becomes inflexible) in precisely the ways Muirhead hopes partisanship would forestall. Once more, closeness partisanship serves Muirhead’s purposes in ways that identity partisanship does not.