What Is “Political Identity” and Why Study It Now?

James Fearon has rightly noted that the Oxford English Dictionary definition of “identity,” as the “quality or condition of being the same in substance, composition, nature, properties, or in particular qualities under consideration,” is insufficient to capture the range of identity concerns visible in modern academic analyses. The American Heritage definitions are more promising: identity is the “collective aspect of the set of characteristics by which a thing is definitively recognizable or known… The set of behavioral or personal characteristics by which a person is recognizable as a member of a group.”⁵

By extension, a political identity can be conceived as the collective label for a set of characteristics by which persons are recognized by political actors as members of a political group. There are many sources of such recognition, such as party affiliation, nation-state membership, ethnicity, economic status, language, and others. All these possible sources, however, are only political identities when political actors treat them as such. We can define “political actors” and “political groups,” in turn, as those people who determine how governing power will be created, distributed, exercised, and ended, in ways that partly decide, among other things, who gets what, when, and how.⁶ A person's political identities—since individuals have many characteristics and group memberships, they usually possess more than one identity—indicate the populations with which political actors expect that person to be affiliated in contests over governing power and its use.

This definition of political identity is agnostic on the question of whether the characteristics that political actors recognize as defining a particular political identity result chiefly from some sort of intrinsic affinity among those who share that identity. In other words, these identities might be seen as somehow natural or primordial or as forms of comprehensible but still-contingent identity constructions. Furthermore, the definition also allows political identities to form either when those who share those identities in ways others recognize voluntarily and actively come together, or when outsiders ascribe a common group identity to certain populations, even if many members of those populations resist. The fact that the definition does not predetermine whether particular political identities are primordial or constructed, voluntary or ascribed, is desirable, because the definition does not bias investigations into whether those things are true. At the same time, this definition does imply that political identities are important realities in human experience. They are labels that indicate how political actors are likely to view various groups of people. This in turn influences how people think about themselves politically and, even more importantly, how they act toward each other politically.

Although it is wise to use a definition of political identities that does not prejudge their origin, I believe investigation will show that such identities are politically constructed to a much greater degree than many scholars in many disciplines now acknowledge, and that such constructions are among the most normatively significant and behaviorally consequential aspects of politics. I therefore think that political identity topics should always have been hardier perennials on the political science agenda than they actually have been. In any case, today's historical circumstances plainly require political scientists to address these topics more than we have in the past.

It is understandable that until recently, other questions loomed larger. The Cold War prompted most American political analysts, in particular, to believe that the big issues were all questions of democratic capitalism versus authoritarian socialism—fascism having been crushed after its brief destructive burgeoning. Political scientists analyzed the many dimensions of that opposition on the assumption that democratic capitalism and authoritarian socialism were political and economic systems constructed within particular states, usually thought of as nation-states. If questions of political identity were raised at all, therefore, the attention generally focused on national membership, class status, and ideological camp.

But the political world changed during the Cold War years in ways that have pushed new problems onto intellectual as well as political agendas. The era from the end of World War II through the 1990s witnessed the dismantling of most Western European empires and the rise of “new nations” in Africa and Asia. The end of empires eventually included the collapse of the Soviet Union and, with it, of many other Communist regimes, precipitating a further wave of nation-building and other, more novel forms of political reconstruction in many parts of the world. The post–Cold War years have also seen the proliferation of transnational economic entities, such as the old European Common Market, the North American Free Trade Agreement, and the World Trade Organization; the development of regional political bodies, notably the European Union and now, perhaps, the African Union; and the development of many transnational corporations as well as movement organizations, such as environmental, labor, and human rights groups. All of these have generated new forms of political community and some have created novel political identities.

The end of formal European imperial rule was, in many regimes, also linked in complex ways to the demise of formal domestic systems of racial hierarchy. Officially renounced systems include: Jim Crow segregation and the related legal denials of equal status to Asians, Latinos, and Native Americans in the U.S.; many Latin American governmental measures to promote the “whitening” of their societies; the “white Australia” policy and many legal doctrines that denied rights to aboriginal peoples there and in other British Commonwealth nations; and eventually, even South African apartheid. The fall of all those forms of domestic legal discrimination sparked efforts to change other unequal statuses. The successes of the civil rights movement in the 1950s and early 1960s, in particular, helped inspire a range of other liberation movements in the U.S. and elsewhere: for women; for cultural, linguistic, and ancestral ethnic groups and national minorities; for gays, lesbians, and others with non-traditional lifestyles; for religious minorities; for the disabled. Most readers can continue this list.

These events have gradually added a great range of previously absent or much more marginal topics to the agenda of political science. Such developments wrought fundamental changes in existing forms of political membership, status, and identity that are tremendously important for billions of human beings. They transformed the basic economic and political opportunities, resources, and statuses available to many political groups, and they also often fostered new senses of who people were politically and what their political identities signified for their lives. And they have set in motion sweeping changes that are far from complete.

Stirred by these historic events, popular writers, political activists, and scholars in many disciplines, eventually including our own, have in the last two decades devoted increased attention, both normative and empirical, to issues that are widely termed identity politics or (if that suggests purely expressive rather than constitutive processes) the politics of identity. Works on racial and ethnic politics; women's politics; politics and religion; immigrant politics; indigenous peoples' politics; gay, lesbian, bisexual, and transgender (GLBT) politics; along with globalization; cosmopolitan citizenship; transnational social movements; and so on, were not entirely absent before recent decades. Moreover, most such work is still remote from the pinnacle of prestigious political science endeavors today. Nevertheless, most of these topics have become far more visible since the 1980s than they were before.⁷

However one assesses their place now, the scholarly attention they receive clearly has been and almost certainly still is increasing.⁸

Increasing, but still neither central nor secure. Though scholars from across our ideological and methodological spectra now engage in such work, an equally varied group is deeply disturbed by these developments. Two scholars whom one might have considered sympathetic to identity studies, sociologist Rogers Brubaker and historian Frederick Cooper, have instead called for scholars to abandon the term as too ambiguous to be useful.⁹

To some critics, concerns about identity are merely a politically fashionable academic fad, one that scholars too often pursue with humanities-derived methods that lack the rigor of true social science. Many also worry that attention to such “faddish” topics can distract attention from earlier sets of political, economic, and social issues that remain extremely important. After all, topics of democratic and authoritarian systems, economic regulation and distribution, and international conflict matter today as much as ever.

In political science, as in other disciplines, a number of lively debates are therefore underway over whether, for example, globalization and transnationalism are, as recent writers aver, really making nation-states less central to political life, including political identities, or whether they are only changing the contexts within which nations and nationalism still play dominant roles.¹⁰

Probably even more scholars are concerned that, analytically, an undue focus on racial, ethnic, religious, cultural and gender identities misses deeper, often economic causes of political action. Some argue normatively that concentration on political identities deflects attention from persistent material inequalities that need to be addressed. Many more fear that such work strengthens false and politically divisive essentialist notions of the character of political actors. Hence one can find many sharp critiques, on the left as well as the right, of scholarly turns to identity politics.¹¹ Most such critics contend that these turns are at best intellectually and politically undesirable, at worst deeply immoral.

Though I too worry about many aspects of scholarship on the politics of identity, I believe the events of the latter part of the twentieth century show that greater focus on both empirical and normative questions about the politics of memberships and identities is inescapable today. The problem is not that political scientists are turning to such issues; it is that we as a profession still have not taken them seriously enough. We have long devoted scholarship to nation-building and nationalism, never more so than in the past two decades.¹²

Post-World War II political science also revived “group theory,” giving some attention to group formation. The political upheavals of the 1960s similarly generated valuable and varied analyses of the construction of gender, racial, and ethnic identities and social movements, far more than I canvass here.

But much of the latter work has followed paths laid by other disciplines—sociology, social psychology, anthropology, economics, as well as literary, linguistic, and postmodernist philosophic theorizing—and has stressed the social construction of identities, rather than advancing explicitly political accounts of political identity formation.¹³

Many of the best political scientists studying identity, such as Henry E. Brady and Cynthia Kaplan, follow the great social psychologist Henri Tajfel and his student John C. Turner. They argue that “ethnicity is a social phenomenon that at the group level is a product of individual choices influenced by contextual actors such as elite influence, available information, contact between individuals, and socialization. At the level of the individual, ethnicity also reflects psychological processes described by social identity and self-categorization theories.”¹⁴ Once ethnic identities are formed, these authors argue, they go on to shape political attitudes and choices. This is a framework that certainly has room for political factors such as the impact of political elites and socialization by political institutions. Yet unless those factors are explicitly highlighted and their special importance considered, these formulations, typical of much political science literature, are likely to suggest instead that a whole range of social elites and institutions are equal to or more important than anything distinctively political in constructing ethnic identities. Those basically social identities may then seem to determine what political actors merely recognize as political identities, as well as the political outlooks and conduct those identities promote. Though some such social determinism may in fact exist, political scientists should question, not presume, such a possibility.

Furthermore, while some cross-fertilization has occurred, most of these social constructivist accounts of identity have been identity-specific. There have been few comprehensive efforts to unify theories of national identity formation with accounts of group creation, social movements, and gender, ethnic, racial, cultural, and religious identities, though there have been many persuasive arguments for specific linkages of one sort or another.

That may be sensible. The many kinds of political identities may be so different from one another that trying to generalize across them could represent a hopeless apples and oranges exercise. A more general theory of political identity formation may be too general to be useful.¹⁵

And if we are talking about a general theory of phenomena that should not centrally occupy our attention in the first place, we might well conclude that the discipline's priorities should not shift at all.

On the contrary, I believe that the breakdown of older imperial systems, the waves of new nation-making, the various movements of identity politics, and the emergence of new international and transnational political associations show that nationalism and class cannot be our only or even our primary identity themes. These events have brought to the surface a whole dimension of political life that should have been a basic concern of political analysis all along—as basic as whether people should see themselves as governed by God and his deputies or by themselves; whether rule should be by the one, the few, or the many; whether economic interests drive politics; and whether governance should favor the rich, the middle class, or the poor, those who control capital or those who provide labor. How human beings acquire certain political identities is just as foundational because these identities often generate and usually substantially affect issues over governing power.

Probably the most politically important of such identities are those that define a person's trumping allegiances in cases where the demands of some memberships conflict with those of others (whether those memberships are in territorial nations, regions, provinces, or cities; religious bodies; racial or ethnic communities; corporate, worker, or other class organizations; or other groups). People's beliefs that they owe primary allegiance to some political memberships, along with the conviction of others that they are likely to hold such beliefs, have major consequences for how people understand their political interests, how they act, and how others act toward them on a range of politically significant matters. It is doubtful that those understandings and conduct can be explained or predicted unless such senses of political identity are understood.

And though it may prove to be the case that such identities actually do emerge seamlessly from extra-political social sources—perhaps inherited languages, ancestral and kinship groups, geographical clusters, initially apolitical religious and cultural associations, or economic structures—it is unlikely that any realm of human collective life operates so automatically. It seems more plausible to assume that, out of the multiple possible identities that human existence presents to most people, political activities of various sorts play important roles first in creating many of those identities, and then subsequently in determining which established senses of membership become salient political identities that command allegiance and shape values.¹⁶

Certainly, political scientists ought to explore political explanations for such identities thoroughly before concluding that their origins lie outside our disciplinary domain.

Some say the dearth of attempts to generate a unified field theory of political identities, which would explain their creation and transformations, is due to the impossibility of achieving such a grand theory. I take that objection seriously. Any all-encompassing account of such diverse phenomena may be so abstract and thin as to be useless. It may turn out that what we need is a set of explanations of different sorts of political identities that cannot be tightly connected to one another. This is more or less what we are now developing. Even if this is so, I still contend that we ought to treat the development of such explanations as one of the most valuable endeavors in contemporary political science, and give it much more emphasis than we now do.

And, we should not give up too easily. Different political identities are undoubtedly crafted through contrasting processes in some cases, but there may also be illuminating commonalities. Political scientists have an intellectual duty to take that possibility seriously. For if it is true, and if political identities do play the seminal roles many now assign to them, then these elements are surely fundamental for both empirical and normative political analyses.¹⁷

For though the relationship of political identities to human interests is complex, we have reason to think that it is reciprocal—that just as economic interests influence our affiliations, for instance, so do those affiliations shape our senses of economic interests. The same person might conceive of her most politically salient identity as a worker, or as a white worker, or as a female worker, or as an American worker, or as a global work force member. She is likely to define her economic interests differently and to pursue distinct political courses depending on which conception she favors. The same sort of list could obviously be compiled for capitalists and many other socio-economic identities.¹⁸

And it is equally plausible to think that our memberships reflect but also help define our interests in personal physical security and political power. The same person might seek protection and representation primarily as a Jew, or as a Brooklyn resident, or as a member of a radical socialist party. Again, these conceptions have very different implications for political conduct. If it is credible to think that political identities can play such decisive roles, then issues of their making and remaking cannot be marginalized in any defensible political science.

The Limited Utility of Formal Theory and Ahistorical Behavioralism

If we are to make questions of political identity formation and transformation as central to our discipline as issues such as constitutional design, electoral representation, legislative behavior, global security, and political economy long have been, such an expansion of our agenda has implications for the methods we employ. Many important aspects of the politics of identity cannot be adequately probed without richly interpretive methods that involve discursively grasping the consciousnesses and senses of value and meaning that identities give to people. And especially if political identities are socially, indeed politically, constructed, we need to attend to the historical and contextual processes through which identities are constructed differently among different groups in different times and places.

We cannot fully understand the sources and character of changes in political identities by looking only at timeless behavioral constants. Similarly, more formal methods that reduce identity choices to points on hypothetical preference functions are also proving valuable for many kinds of empirical identity studies. Yet abstract connotations of identity preferences cannot go very far in helping us comprehend the substantive appeal and normative significance of particular identities, both of which are important to understanding how those identities are likely to shape conduct. Both behavioral and formal models that treat senses of identity and interest as exogenously given, moreover, simply do not seek to shed any light on how identities are formed and changed.

We are likely to gain more insight on these topics through interpretive textual analyses; ethnographic fieldwork; biographical studies; in-depth interviews; individual and comparative case studies, both historical and contemporary; participant observation research; narrative historical institutional analyses; and other methods rather crudely termed “qualitative.” Though I cannot summarize what all these varied methods involve, each has a growing literature analyzing their distinctive ways and means, and their strengths and weaknesses.¹⁹

No method, however, provides an infallible cook book for grasping the phenomena of political identities. As I suggested above, a combination of contextual immersion, psychological empathy, and creative imagination, in a variety of research techniques, enables researchers like James Scott, Cathy J. Cohen, Claire Kim, and Courtney Jung to offer alternative conceptions of the identities they study, whether their subjects have previously been conceived as passive peasants, politically homogeneous African Americans, irrational racial nationalists, or ethnic tribespeople.²⁰ Again, I think even the partly ineffable process of concept formation always involves at least a rough method of inference testing. Researchers like these invariably check preliminary hypotheses formed on the basis of a few observations against many others made in the course of their fieldwork before elaborating positive arguments. Researchers can, in turn, use such conceptions to construct better large-n surveys that are more easily subjected to systematic, if always imperfect, statistical analyses to test particular claims. That work is valuable, but it depends on the strength of concepts that are generally formed by other means.

Not only must the means through which concepts are generated be to some degree contextually and historically sensitive, but so must all the methods used to formalize, operationalize, and test them. Because the political identities that prevail in one location in one era may well be very different than in another place and time, they can prompt local patterns of political behavior that we can only understand by grasping the substantive characteristics and distinctions among ways of life in those contrasting contexts. A large team of anthropologists and economists has recently provided support for this claim. In a set of field experiments, they asked members of many different societies on several continents to participate in rational-choice games involving decisions either to offer or to accept cash drawn from a researcher-provided pie. These researchers contend that the behavior of most of these societies did not match standard economic models of wealth-maximization, and that the sorts of rationality the societies displayed varied widely. Both to conduct the games successfully and to explain these variations, the research team relied heavily on ethnographic studies of the social and economic practices of each society, and they argued that these distinctive patterns of group behavior greatly explained individual choices.²¹

Their work shows how essential such ethnographic analysis is to insightful investigation of how distinct identities and values are formed and maintained. But like most research in the other social sciences, their study does not much explore whether political struggles, and the social policies and practices thus established, played a substantial role in constructing the different group behaviors and identities that characterized the different societies. Nor should such scholars have to do so. Analyzing the contributions of politics is our job.

But are we not doing it well enough? Let me consider two prime counter-examples. First, David Laitin's justly celebrated Identity in Formation: The Russian-Speaking Populations in the Near Abroad offers a theory of political identities centered on an abstract rational choice “tipping model.”²²

Laitin suggests that this tipping model is the key to illuminating the question of whether Russian-speaking people in ex–Soviet Union states will learn the dominant language of those newly independent nations or remain Russian, at least linguistically.²³

Though his theory features rational choice, Laitin's book is an impressive example of multi-method analysis. He presents an informative, analytical recounting of historical and institutional developments that help explain why so many areas never became Russian-speaking under either centralized Tsarist or Soviet rule. Laitin draws as well on four ethnographies conducted in the newly independent republics he studies, and on his own experiences living in Estonia for seven months, to dissect recent trends. He and his colleagues also conducted an extensive large-n survey of Russian and “native language” speakers in those republics, and he uses the survey in various quantitative analyses. Additionally, Laitin adds both interviews and documentary accounts to identify pertinent state policies in each republic, along with a Russian-language discourse analysis based on articles in the Russian press and the presses of the republics.

When all is said and done, Laitin's rational choice model of identity formation contributes surprisingly little to understanding the processes of identity formation. Most of the work's explanatory power comes from features he treats as theoretically less significant, especially his analysis of historical institutional patterns and current state policies, as well as from his evidence of group attitudes, for which he gives no real theoretical account. In the end, the book does more to demonstrate the limitations of purely formal rational choice models as responses to the questions he addresses than to show the strengths of such models.

Laitin seeks to build on Thomas Schelling's renowned “tipping model” analysis, which clarified, among other things, why discriminatory real estate practices do not have to be massive to produce sharply segregated residential patterns. I strongly affirm the usefulness of Schelling's formal model for depicting the consequences of racially inflected (rather than strictly wealth-maximizing) economic choices. Laitin, however, uses his “tipping model” to support a more obvious claim. Russian speakers in societies where Russian is not the dominant language will learn the native language, he suggests, when they find it advantageous to do so. But when is that? The “tipping model” suggests that one important element is their perception that most other Russian speakers in the region are shifting to the local language. If this is true, then presumably it will be increasingly harder to remain one of the last few monolingual Russian speakers.²⁴

We thus need to know what factors cause first some, then many, to decide to pay the costs of learning the local language, the initial impetus that eventually creates a cascade in which it becomes understandable why most of those remaining take that step. We also want to know why Russian speakers in some republics make the shift more readily and massively than in others.

To answer, Laitin elaborates his choice model. People might learn a new language under three conditions—when it is economically advantageous to do so, when members of their own “in-group” do not punish, but even encourage, such adaptation, and when speakers of the dominant language do not punish, but even reward, learning their language.²⁵

Though Laitin presents these elaborations in abstract terms, they quickly move the bulk of his analysis to non-formal, extra-model elements. If choices are greatly shaped by the attitudes and behavior of “in-group” and “out-group” members, then we need to know how those group identities formed, and why those groups punish or encourage members and outsiders as they do. The tipping model says nothing about those key issues.

The model might seem more useful if economic calculations told most of the story in explaining personal choices, as in Schelling's real estate example.²⁶

But when Laitin tests his model against data gathered in his surveys of the four republics, he concludes that if “the tipping model relied solely on economic returns and probabilities for occupational mobility, these data present an insurmountable challenge.”²⁷ The expected economic returns for language acquisition in the four republics just do not match the willingness to learn that Russian speakers in the different regions expressed. The economic returns of change are, for example, worst in Latvia, yet that is where Russian speakers are most eager to undertake linguistic assimilation. Sadly, the economic results do not display a nice pattern of inverse correlation, either. They do not have much explanatory power at all.

What works best is a sense of social distance. In republics where Russians share a similar language and religion with the dominant group, as in Latvia (Indo-European language and Christianity), the Russians are most open to change. Where that social distance is greatest, as in Kazakhstan (non-Indo-European language and Islam), they are least open. But how do we judge whether a language or religion is distant or close to another? We can do so only by learning about the particularities of the different languages and religions in question and determining how far, on the whole, they resemble each other in decisive respects. We must make those determinations prior to constructing a data set measuring social distance and they must be persuasive if any argument from that data set is to be convincing. On Laitin's own evidence, then, what we grasp by interpretively characterizing the identities in question—our judgments of substantive religious and linguistic similarities and differences—is far more crucial to explaining change than the tipping model is.

Laitin says this evidence applies only to conditions during the Soviet era, when there were few reasons to learn the regional language other than personal preference. He argues that now, “Russian-speakers need to calculate more consciously the potential payoffs for learning” the local language, and so the tipping model will work better in the future.²⁸

Laitin has to promise future vindication, because his other data do not provide much corroboration for the tipping model, either. For example, Russian speakers in Kazakhstan who learn the dominant language face low friendship losses among their in-group, and also have the best prospects for friendships among the Kazakh out-group (though the prospects are still negative). Nevertheless, they are the least open to doing so.²⁹

The few rather weak results that do conform to the model, moreover, do little to explain either the power of in-group and out-group identities or the reasons for their variations in disapproval and receptivity. For those questions, Laitin turns in part to the politically shaped historical patterns he identifies in the first part of his book, and in part to current state policies to which those patterns may have contributed. His analysis suggests that in those republics whose local elites had wide opportunities for mobility within the Soviet Union, such as Ukraine, the rewards Russians receive for learning the local language are not so high, partly because those elites are long accustomed to intermingling with Russian speakers and often speak Russian themselves. Conversely, in republics such as Kazakhstan, where the Soviets governed on a more colonial model and local elites held relatively little power, many Russians remain hostile to embracing what they still see as an inferior Kazakh identity. The rewards for learning the local language are not great there either, perhaps because resentment toward Russian assimilation and mobility remain strong (though the data do not clearly support that claim). In countries like Latvia and Estonia, where local elites had real internal power but not many prospects to rise in the larger Soviet Union, there tend to be more benefits to learning the dominant language and otherwise assimilating. Elites are relatively receptive to Russians who do so, and they expect business to be conducted in the local tongue. These different patterns of receptivity, traceable to identity-inflected power structures under Soviet and Tsarist rule, help explain why in some places enough Russians learn the local language to launch a cascade (though Laitin has no data showing any actual cascades).³⁰

Whatever weight we give such explanations—again, social distance variables fare best—any real explanatory power they possess does not come from the tipping model or any other formal theory. Rather, it rests on different historically shaped patterns of political institutions and state policies, patterns that have both expressed and reinforced distinctive senses of political identity and allegiance. Reasoning from these patterns, given supportive empirical evidence, scholars could explain the prevailing patterns of linguistic assimilation without any tipping model. If state institutions and policies and the behavior they foster make it sensible for most people to change, then most people will change. The intervening variable—how many others are changing—will probably prove more intercorrelated than independently causal. Laitin's invocation of a rational choice model does not in the end, provide much “value added” to his rich historically, institutionally, ethnographically and survey-based analysis.³¹

Laitin's landmark study thus suggests that grasping identity formation and change requires interpretive understandings of different identities, which enable us to comprehend phenomena like “social distance.” In addition, analyses of identity formation may also advance by understanding the historical processes of institution-building and power-structuring that have strengthened and modified certain identities, fostered new ones, and played strong roles defining the relationships of those identities to others.

Similar lessons can be drawn from an excellent recent example of behavioral political science, Brady and Kaplan's quantitative analysis of ethnic identities. Though they agree with other analysts that such identities have great explanatory power for political attitudes, behavior, and change, they contend that it is a “categorical” mistake always to measure ethnic identity in nominal terms, as either present or absent. Sometimes ethnicity should be measured as graded, varying depending on how strictly confined to co-ethnics are a person's social contacts, exposure to media, and positive group evaluations. Brady and Kaplan maintain that nominal measures of ethnic identity are fine when such identities are highly salient to those who possess them. Such consciousness tends to homogenize their political attitudes, at least on issues related to ethnic identity. Gradations in the degree to which these ethnic group members are immersed in their ethnic communities are then not associated with any significant variance in their political attitudes, so nothing is lost if ethnicity is measured nominally. But when many members of an ethnic community do not find their ethnic identity especially salient, then gradations in how extensively their social worlds are confined to co-ethnics correlate strongly with variations in political attitudes. In those circumstances a graded measure of ethnic identity should be used.³²

Brady and Kaplan support these claims with surveys of political attitudes toward discrimination, civil rights, and Estonia's relations with Russia that are held by ethnic Estonians or ethnic Slavs (overwhelmingly Russians) in Estonia.³³

The authors find that in the case of ethnic Estonians, for whom they say ethnic identity has long been and remains highly salient, nominal measures of ethnic identity predict political attitudes about as well as do graded ones; that is, ethnic Estonians who interact extensively with Slavs view discrimination, civil rights, and Russian relations in roughly the same way as do ethnic Estonians who live far more insular lives. The political attitudes of ethnic Slavs, however, vary greatly depending on whether their social worlds are exclusively Slav or more intermingled with those of ethnic Estonians. Brady and Kaplan maintain that this is because their Slav or Russian ethnic identity has long been much less salient to such Estonian residents. Hence they conclude generally, as a law-like regularity in political behavior, that the salience of ethnic identities determines whether nominal or graded measures of ethnic identity are most appropriate when we use such identities to explain political attitudes.³⁴

That claim is convincing, and I do not denigrate its importance. Yet though this work advances the science of measuring identities, it glides lightly over some of the identity issues that I think political scientists most need to stress—questions of whether and how political identities are politically constructed. The authors might justly respond that these are not the questions they wished to address. But problems remain, because their analysis can easily be read as implying answers that point away from politics.

It is likely, in the first place, that Estonian and Slav/Russian ethnic identities can best be viewed as historical products of past political struggles, though exploring that history would admittedly be beyond Brady and Kaplan's scope. It is fairer to fault the authors for failing to highlight how their evidence shows that politics determines the role of these ethnic identities in political attitudes and conduct today. As they present it, their study culminates in a rather apolitical lesson about how best to do social science—when to use nominal and graded measures—grounded on what they suggest is a general behavioral regularity: the salience of ethnicity always determines which measure should be used. They do not go on to explore the logical next question: what determines whether ethnic identity is salient to certain groups? Instead, they leave that issue to be handled by “social identity and self-categorization theories” that give no special prominence or attention to political factors.³⁵

Yet their essay includes ample support, consonant with Laitin's, for the view that, at least in the former Soviet territories, one can often trace the salience of ethnicity to historical and contemporary governmental policies, institutions, and political struggles. They note that the two ethnic groups they study in Estonia long “experienced the Soviet policies of ethnic engineering,” but in very different ways. For decades, the Soviet Union sought “to subvert ethnic identification,” replacing Estonian identity with the supra-national political identity of Soviet Socialist Republican citizenship. In fact, these efforts were simultaneously a hypocritical program of Russification that privileged Russians above other ethnicities, like Estonians. The result was that for Russians, ethnicity did not appear salient because it quietly blended with socialist and Soviet Union political identities. For ethnic Estonians, however, it was overwhelmingly salient, due to the concerted, long-term, discriminatory “attack upon Estonian ethnic identification” that limited their opportunities in the regime as a whole.³⁶

These experiences were far from forgotten when the Soviet Union fell.

One can therefore imagine counterfactually that if the Soviet Union had adhered rigorously to its socialist ideology and not simultaneously promoted Russification and discrimination against other ethnic groups, many Estonians might have come to identify more fully with their Soviet citizenship, instead of becoming ripe for “ethnopolitics.”³⁷

Their ethnic identity might have become as little salient to them as Russian (or Slav) ethnicity was to its bearers. Admittedly, this may be too counterfactual to be conceivable: it is unclear that the Soviet political project could have been as successful as it was without its alliance with Russian nationalism. Either way, however, we must conclude that prior politically constructed identities, including ethnic and national, often shape people's senses of interest and allegiance in ways that political actors and state policies must take into account to be effective. Furthermore, those state policies and institutions in turn frequently play leading roles in determining whether ethnic or other identities become and remain salient in political outlooks and choices.

In sum, by focusing only on the relationship between different senses of ethnicity and political attitudes, and by relying on relatively apolitical social psychological accounts of the origins and salience of those ethnic identities, Brady and Kaplan do not fully illuminate how the different political roles of ethnic identities are themselves products of politics. And in this case too, the most reliable illumination they do provide comes not from their quantitative analyses of degrees of ethnic identity, although those are valuable, but rather from the historical institutional dimensions of their argument.

At this point, my fellow interpretivists and historical institutionalists may cheer, but all political scientists may also feel anxious. We may not excel at such qualitative work. Sociological and anthropological ethnographers, literary critics, and narrative historians may do it better. And even doing it as well as possible may not satisfy our aspirations to social scientific explanatory theories. Those theories are supposed to be, if not universal, at least falsifiable, replicable, and generalizable over some middle range of cases. They should not be simply thickly descriptive case studies or more literary narrative accounts of unique historical events. Although I believe such works are valuable sources of insight, I do not in fact think that political scientists need to despair the loss of their social scientific theoretical ambitions.

The Role of Social Science Explanatory Theories

Let me reiterate that, though questions of political identity were not central to the American political science profession's agenda through much of the twentieth century, they have come to the fore much more in recent decades. We have valuable political analyses of the formation of nations, races, ethnic social movements, gender identities, religious affiliations, and other identities. These works show that focusing on topics in politics of identity need not be limited to ethnographies or narrative histories.³⁸

Cathy Cohen's Boundaries of Blackness, for example, uses a variety of research methods, including participant observation research, to develop a novel theory of how modern politics has reconstructed the identities of African-Americans. Other researchers are finding her theory applicable to different long-suppressed groups. She argues that in the wake of victories in the civil rights era, black Americans find themselves in a position of advanced marginalization: they are included as civic equals more than ever before, yet are still confronted by a variety of more subtle forms of discrimination. Those circumstances create lamentable incentives for more prosperous and socially conforming members of African-American communities to distance themselves from blacks who, because of their poverty, drug use, sexual orientation, or other traits, are seen as threats to the still-fragile civic status of African-Americans generally. The better off still define themselves as black Americans, but as ones who are now closer to being normal middle-class Americans. The latter identity heightens pressure to limit the inclusiveness of the former. Such attention to the complex processes of political identity construction enacted both upon and by different groups of black Americans remains an infant enterprise in political science, but it is essential if we are to understand central dynamics in contemporary American politics.³⁹

Although Cohen's advanced marginalization theory may well illuminate processes affecting many groups in the U.S. and elsewhere, it is still focused on those who have long faced official discrimination. I suggest that we seek to do even more. Grandiose as the notion of a “unified field theory of political identities” may be, it seems worth exploring whether we can in fact generate such a theory or, better yet, rival theories that we can test against each other and then improve. I make this recommendation because my own studies of the power of racism and nativism in American history have moved me to higher levels of theoretical generality in the quest for satisfying explanations of that power.⁴⁰

A sketch of the framework I now use to think about political identities may help to clarify not the “right” theoretical approach to such questions, but rather the sorts of theoretical endeavors I urge. In various writings, I contend that senses of membership in a political community, a political people, to which one is commonly understood to owe allegiance, are indeed political creations. They do not emerge semi-automatically from economic, demographic, sociological, geographic, linguistic, ancestral, biological, religious, or cultural characteristics. Rather, drawing on and constrained by such features of human life, aspiring elites craft many forms of political peoplehood by winning the support of a critical mass of constituents for their visions of political identities and memberships.

Such crafting of senses of political identity takes place through coercive force and persuasive stories. Many Native Americans are Americans because some of their ancestors were conquered, many African Americans are Americans because some of their ancestors were enslaved, and all Americans are Americans because a critical mass of colonists exercised sufficient force to make the more powerful British imperial authorities decide that it was easier to let them go than to coerce them into continued subjectship. Political science equips us well to grasp the role of coercive force in crafting political identities: that seems solid, measurable stuff. So I have concentrated instead on stories of peoplehood and the role they play in winning voluntary constituent allegiances.⁴¹

That role is inescapable, because no leaders have enough force to make all members of their community subscribe to their vision through coercion alone.

What kinds of stories inspire people to embrace membership in a particular political community? Like so many in social science, my model has three components. Economic stories offer material benefits for membership to current and/or future generations—both capitalist and socialist ideologies are stories of this sort.⁴²

Political power stories promise personal protection and a share in great collective power—both republicanism and fascism are instances. Ethically constitutive stories claim that membership in a particular people is somehow inherent in who the members truly are, in ways that are ethically valuable. Most racial, religious, ethnic, cultural, linguistic, historical, and gendered senses of community membership, among others, are ethically constitutive accounts.

On this basis I first hypothesize an empirically falsifiable behavioral regularity: all “real world” visions of political membership capable of attracting significant support, from Americanism to radical Islam to those offered by leaders of communal eco-villages, always blend together versions of all three types of stories.⁴³

The emphases differ—some are more economic, some more ethically constitutive, et cetera—and the substantive particulars vary, but the three types are always present. Additionally, a political actor, party, or movement rarely succeeds in institutionalizing their pure vision of peoplehood without making concessions to opposing views. Instead, political societies are always composed through the compromised, aggregated results of contests among proponents of different societal visions, each vision offering accounts that include all three story types, in ways that are always being contested and changed. This, in a nutshell, is the politics of political identity formation. Or so I claim.

My aim here is not to demonstrate that this framework is correct or even helpful, though I hope it is. Rather, I am suggesting the sort of theory building regarding political identities that now seems advisable for scholars to undertake. This is a framework or theory that meets social science aspirations both to heuristically useful simplification, since it involves elaboration of only a few basic ideas, and also to universality, since it purports to identify basic ingredients with which political identities are constructed in all times and places. It is necessarily abstract, but not as abstract as models of pure instrumental rationality. Rather, it presumes that all people do have basic, recognizable types of substantive interests: material well being, some forms of political protection and political power, and senses of ethically constitutive identity. It also presumes that these needs can be met in a great many ways.

In other writings, I elaborate why my category of ethically constitutive stories in particular highlights discourses that play vital roles in political life in ways that the other types of stories cannot do as well. Ethically constitutive stories are best equipped to confer an aura of moral worth on memberships, and they are also harder to discredit via empirical evidence than economic or power stories. These features enable them to sustain loyalty in materially bad times more effectively than the others can. For me, the framework helps explain why racial and nativist conceptions have so often been politically potent in the U.S., even when they were not helping to rationalize economic exploitation and were costly to enforce. It is appalling yet true that traditional American racial and ethnic stories have long supplied many native-born whites with a sense of superior moral worth that no evidence could dispel, but that racial equality and easy immigration could threaten. In so arguing, I generalize one view of W. E. B. DuBois's famous assertion of the existence of a “public and psychological wage” for whiteness.⁴⁴

This simple framework for analyzing the politics of people-making is a kind of general, trans-historical theory about some important dimensions of political identity formation. Because it has more specific substantive content, because it promises to help us understand the distinct roles of different types of political identity, and because it can spark ideas about their strengths and limitations, both politically and normatively, it seems to me to do more work than purely formal models. It also represents an effort to develop a theory of political identities that is more thoroughly political than many prevalent in the quantitative empirical literature. But it remains limited in its contribution to full understandings of the politics of identity making and remaking.

On its own terms, this framework requires us to engage in historical, empirical, and often ethnographic and interpretive work to grasp the kinds of economic arrangements that render certain sorts of economic stories more probable; the contextual traditions and structures of power that make specific political power stories capable of inspiring allegiance; and the existing array of demographic, cultural, linguistic and other identities that make the formulation of certain sorts of resonant ethically constitutive stories, and not others, possible. Like the similarly ethnographically-informed work of the researchers on game-playing, and like the historically- and institutionally-informed arguments of Laitin, Brady, and Kaplan, these investigations are often likely to confirm the belief that people conceive of their economic interests, political power interests, and ethically constitutive interests very differently in different times and places, and that different populations do so in the same times and places. That means, again, that though the same general underlying processes may be operating, the specific explanations for and predictions about political conduct that work in those contexts will vary from one to another. No general theory is likely to enable us to eschew many kinds of detailed contextual investigations. Such theories can only aid that work.

In short, we need to study processes of political identity formation largely through interpretive, ethnographic, and historical methods. Yet we can also hope to develop theoretical frameworks that will enable us to knit those studies together to a greater degree than we are now doing. A final example: my colleague Ian Lustick has been exploring agent-based modeling, which seeks to wed the logical precision of formal models with more contextual knowledge to illuminate processes of identity change in much the way I am urging. He seeks to show, for example, how political actors draw on repertoires of identity available to them in particular contexts to remake themselves in more circumstantially optimal ways.⁴⁵

Here, too, I doubt that most of the work is being done by the formal elements of the analysis. It is instead accomplished by documenting the available identities and the circumstances that make embracing some new sense of self attractive. But this work is a clear instance of the somewhat less thin but still theoretically ambitious endeavors that seem appropriate today.

It remains to be seen how much trans-historical insight and explanatory theorizing can be developed from such efforts. Perhaps we will be able to generate plausible trans-historical theories of patterns in the historical evolution of all political identities. Perhaps we will only be able to make sense of political identity formation at more middle-range theoretical levels. But through such work we will, I think, advance significantly our insights into these key features of political life.

If, that is, we try to do so. My aim here has been to provide some ideas that may aid those already predisposed to work in these directions and that may move some who are not as disposed a bit closer toward thinking that these tasks are worthwhile. If I have had any success in modifying anyone's sense of professional identity and purpose, then I further contend that this essay is actually part of an admittedly non-randomized field experiment—one that has, through the impact of its professional story on its readers, in contrast to unaltered masses of non-readers, transformed some identities and not others and thereby provided empirical support for its claims.

And if it turns out that no one is persuaded in the least by any of this, then, in the even more identity-reconstituting words of the late Gilda Radner, never mind.