I want to begin this discussion of On the Side of Angels by complimenting Nancy Rosenblum for writing this book. She focuses attention on the importance of political parties to democracy and to political science scholarship, and criticizes the limited recognition of these points in contemporary research. Even though political parties are central actors in the democratic process, the fashions of political research in this area have recently emphasized civil society activity, new social movements, and other developments. Research on political parties apparently lacks this glamour or novelty—and the book seeks to correct this by giving research on parties its due.

In addition, this book signals a welcome trend among some political theorists to engage theory more directly in the empirical study of contemporary politics. In my view, too often political theory becomes an abstract exercise in imagination or historical reconstruction. Most theorists address important social and political issues, but too often through introspection, subjective observation, or the exegesis of historical texts, rather than through the intersubjective analysis of empirical evidence. Given the vast storehouse of political theory, there is great potential to apply existing theories in order to determine which are really useful in developing a generalized understanding and interpretation of the world. Rosenblum and a new wave of other political theorists are applying their insights and analytic skills to contemporary issues where they can enrich the research debate. My late colleagues Harry Eckstein used to say that political theory should not stand alone as a subfield, but should be engaged as part of each of the substantive fields of political science. Rosenblum's book is an excellent example of that dictum at work. If theorists and behavioralists can discuss party identification in meaningful ways, then there should be more possible bridges between these two camps within our discipline.

The book's subtitle, An Appreciation of Parties and Partisanship, summarizes its general theme. In 500-plus pages, the book covers much territory, as discussed in the other essays in this symposium. For my contribution, I would like to focus on Chapter 7, which presents Rosenblum's perspective on the concept of party identification. For those of who study political behavior, party identification (PID) is the holy grail of electoral research. These early-learned, enduring, affective, psychological identifications with a political party form the core of many citizens' belief systems. Angus Campbell and his colleagues initially described the concept's importance:

Electoral scholars generally maintain that PID is the most important invention/discovery of electoral research (Weisberg and Greene Reference Weisberg, Greene, MacKuen and Rabinowitz2003, 115; Green, Palmquist, and Schickler Reference Green, Palmquist and Schickler2002; Dalton and Wattenberg Reference Dalton and Wattenberg2000, chap. 2).

Rosenblum's review of the party identification literature begins by trying to locate this concept within the framework of partisan and antipartisan traditions in American politics. Partisanship has had a mixed heritage, such as Tocqueville's discussion of parties as a necessary evil and contemporary debates about excessive partisanship in Washington. Party identifiers are, by her accounts, “on the side of the angels,” as she describes the partisan tradition in American politics. This chapter thus echoes some of the claims by Campbell and his colleagues about the importance of partisanship for contemporary democratic publics.

Rosenblum then asks whether independents are the modern successors to earlier populist antiparty sentiments. She offers several interesting observations, arguing that independence is more than just the lack of party attachment; it also can reflect an identification with an alternative view of politics. But independence lacks the core of most social identities because it is a loyalty to a nongroup. She also maintains that independents are not antiparty in their views and values but antipartisanship. That is, independents do not accept partisan identities as a basis of political action.

This is a stimulating discussion, but it struggles to reconcile several pieces of conflicting evidence. First, there is a conventional discourse in American politics and political thought on the nature of partisanship and independents. In this literature, partisanship is often treated as an invective, while independence is viewed as having civic virtue. Independents are supposedly rational thinkers and decide their votes on the merits of the candidates, rather than blind partisan loyalties. Second, there is the empirical literature from electoral research in which these descriptions are nearly reversed. The American Voter (Reference Campbell, Converse, Miller and Stokes1960) and most of the early party identification literature described partisans as more informed about politics, more politically engaged, and more likely to vote. Independents, in contrast, existed at the boundaries of politics as disinterested and disengaged citizens.

A more recent revisionist literature, however, argues that partisanship and independence are changing in important ways. I think Rosenblum misinterprets this part of the empirical literature. Throughout the chapter, partisans and independents are treated as homogenous groups that can be characterized in simple terms. One is sophisticated and the other is unsophisticated; one embodies civic idealism and the other is the opposite. Yet partisan identifiers and independents are not homogeneous in their level of political sophistication, engagement, political interests, and political behavior. Some partisans use the party labels they inherited from parents or social context as a blind guide to make political choices; other partisans are well-informed and active participants in democratic debate and so their identities include rich political content. Similarly, the traditional independents in The American Voter were at the margins of electoral politics, almost prepolitical in their orientations. Since then, the percentage of independents has grown substantially.Footnote ¹ Moreover, a larger share of these new independents are politically aware and engaged—the antipartisanship label is less applicable to this group than a postpartisanship label.

By not fully recognizing this complication, the discussion of the literature on partisan identifiers and independents appears to be making contrasting claims. However, these constrasting images arise because identifications and sophistication interact, and simple descriptions miss this variability. Even more important, this shift in the characteristics of partisans and independents is affecting electoral processes and democratic politics. For example, the nature of issue voting and the role of candidate image have differential affects as a function of partisanship and sophistication levels. The frequency of stable voters and swing voters also varies with these two traits. The old labels and descriptions are insufficient for understanding the current implications of party identification and partisan independence.

The second major theme is the linkage of party identification to the substantial literature on social identity theory. This is an area where substantial cross-fertilization is possible. Rosenblum describes partisanship as the political identity in a representative democracy. Yet it is an assumed identity without the ascriptive characteristics of other social identities, such as race or gender. Social identity theory can add to our understanding of how partisanship develops. Similarly, party identification can illustrate how identities interact with interests and sophistication. The author discusses the separation of identity and interests, but how these interact in party attachments is not fully specified. At the same time, some of the interpretations of party identification are too focused on the unusual two-party system of the United States, and a comparativist would see PID as more complex in a multiparty system where there are multiple competing identities for the same political interests. There are hints of the potential linkage of identity politics and party identification in the chapter, but more development would be welcomed.

This fusion of political theory and empirical research also raises a set of methodological questions. Presumably, the audience for this research has divided loyalties. I looked for empirical evidence to support Rosenblum's claims, and there is a good grounding in the empirical literature. But I was perplexed when I checked certain footnotes and discovered references to Henry Adams's opinions about independents in the election of 1872 or to President William Howard Taft's description of parties during his college days. In short, there is first a question of how we value evidence in supporting our assertions. Second, there is a question of the relevance of even accurate historical observation to present partisan politics. I, at least, suspect that partisans in 1872 lived in a very different world of limited information, limited voter education, different meanings of political parties, and elections without secret ballots—so that observations on the 1872 election seem digressions, rather than evidence relevant to contemporary party identification.

The chapter also raises new perspectives on many other issues in electoral behavior research. For instance, there is an insightful discussion of how party activists use social identities and shape them into partisan attachments that have winning potential in elections. In the end, I think that behavioralists will benefit by engaging the fresh view of the literature that Rosenblum offers, whether they agree or disagree with the specifics of the review. Similarly, theorists might see how long-term issues of philosophical debate have relevance and worth through their connections to contemporary political phenomenon. As Rosenblum persuasively argues—and as her book clearly demonstrates—the topic of political parties and their role in the democratic process is central to both normative theorizing and empirical analysis, and our understanding of this topic is likely to be enhanced by further attempts to fruitfully integrate these modes of inquiry.