As anyone who has tried knows, the central problem in thinking about race in Brazil is how to. The almost quantum-theory-like indeterminacy of the ways Brazilians of different skin colors interact has exercised imaginations for decades, from fin-de-siècle scientific racists, to the eugenists of the 1920s, to interwar modernists who promoted the idea of racial democracy, to the Brazilian and later North American revisionists of the 1950s and beyond. The complexities of Brazilian race have not always been in the forefront of these debates. For much of the period up to the 1970s, scholars focused on debunking Brazil's vaunted myth of racial democracy—the national ideology claiming Brazil to be free of racial prejudice (Costa 1985). The effort was roundly successful. From this literature, we learned not only how wide a gap there has been between the ideal of racial democracy and the reality of racial and color prejudice in Brazil, but also the role elites have played in manipulating the myth to defuse racial and other social tensions (Hanchard 1994). Recently, some scholars have suggested that it is high time to look beyond the debunking agenda and take up once again the complexities of the Brazilian situation. Anthropologists have led the way, seeking to reveal “the range of contemporary understandings” of racial democracy and to explain something of its persistence as a tangible “dream” in the face of ongoing discrimination and prejudice in everyday Brazilian life (Sheriff 2001:8).

This turn opens out to broad questions regarding the conceptualization of race in the Brazilian context. The central issue is how to understand what might be thought of as Brazil's distinctive cultural option with regard to race. According to anthropologist Peter Fry, the United States has “opted definitively for the particularism of the races,” the presumed naturalness and fixity of racial identities. In Brazil, by contrast, a sense of unity that transcends race, what Fry calls “universalism,” remains a vibrant part of everyday life. Failure to acknowledge the distinction, argues Fry, has often led scholars—especially, though not exclusively North Americans—to a careless use of language. Terms and expressions such as “nonwhites,” “whites,” “blacks,” “African Brazilians,” “racial groups,” and many others, says Fry, are not analytically neutral. They are, rather, “carriers of value in the context of North American ‘racial politics.’” Their uncritical use in the Brazilian context “elides” distinct racial situations, making racial democracy into a kind of “error or aberration” that must be explained. The question for Fry is whether race can be understood on Brazilian terms, terms than might diverge rather sharply from the ones North American (and some Brazilian) scholars have tended to rely on. His answer is clear: “I'm sorry,” he writes, “but I believe that a ‘racial politics’ does not have to follow the same paths as in the Anglo-Saxon world” (Fry 1995–1996:96, 126, 133–34). Ditto a scholarship of race.

What I take Fry to mean here is that in seeking an understanding of race in Brazil it will not do to assume that the language and sensibilities of U.S. racial theorizing will work. What might this mean with regard to the conceptualization of race in Brazil? The heart of the matter is that available notions of racial identity are rooted in “foundational concepts of self and identity” that have grounded notions of personhood in liberal societies. These notions of self and identity, argues philosopher Linda Alcoff, are premised on “purity, wholeness, and coherence” (1995:261). This is how, in the United States, a term like “racial groups” can make immediate intuitive sense. Once the purity and fixity of racial identities is presumed, it follows that there will be relatively clearly defined “racial groups.” There is no great ambiguity in the term, because the “objective” criteria that allow individuals to be assigned to a group match up with the felt—or attributed—identity of those who live in a regime of supposedly pure racial types. The so-called “one drop rule,” in which one drop of African blood makes one African American, is the clearest historical expression of this pre-theoretical commitment to purity. Or, as Robert Young has noted, the idea of race “only works when defined against potential intermixture” (1995:19). This is why miscegenation has had such a fraught history in the United States: race mixing is an acid that eats away at the idea of a stable and autonomous self.

In the context of Brazil's large, racially mixed population, a commitment to purity, wholeness, and coherence makes little sense. Brazilians, Fry argues, conduct their social lives according to a modo múltiplo, a multiple mode of color gradation, whereas North Americans live according to a bipolar division of society into sharply defined racial groups (1995–1996:130–34). By itself, this is not a new idea. Degler's famous “mulatto escape hatch” was an early North American attempt to open up a new line of inquiry regarding race in Brazil and the United States (1971). What Fry wants to get across, however, is that Brazilians—as Brazilians, not just as mulattoes or blacks—see and know race differently than North Americans as a consequence of this difference—less categorically and more contextually (Burdick 1998:112–17). Specifically, by not seeing only sharply divided groups, it remains possible for Brazilians, across color, to subscribe to a kind of “universalism” which, ideally, transcends race. This does not mean Brazil has vanquished racism. It means, rather, that alongside the divisive impulse to practice and recognize prejudice and discrimination there is a competing value, one that refuses to allow the perception of racial difference to subsume all other possibilities of human interaction. Thus does racial democracy remain so powerful an oneiric presence, a “dream” in the face of ongoing discrimination.

To understand this abiding presence requires a deep historical investigation that approaches Brazil from a critique of the intimate relationship between race and liberalism. The point of departure for such a project is to recognize that race and racism are not merely unhappy accidents of modern Western culture. They are immanent if veiled aspects of liberal development itself (Banton 1977; Hannaford 1996; Fredrickson 2002). This is what makes thinking about race so difficult—racism is the ghost in liberalism's ideological machinery.

Historically, race was the chief means by which liberal egalitarianism saved itself from fatal self-contradiction. Even as the notion of abstract human equality was being articulated, actual inequalities among human beings persisted and sharpened during the economic revolutions of the seventeenth and eighteenth centuries and during European colonialism in the nineteenth century (Malik 1996:70). Both in the colonies and at home, European and Anglo-American elites confronted forces—principal among them masses of “backward” people—that seemed to resist the unfolding of progress. Race was the always-tenuous effort to naturalize the inferiority of those who were felt to lie beyond the radius of liberalism's charmed circle of humanity—natives, colonials, and retrograde internal populations. What came to be seen as unignorable physical differences later took form as the scientific “fact” of racial difference, a development that successfully papered over the contradiction between the idea of racial innateness and the axiom of universal human equality. Liberalism, in short, can avoid neither the historicity of its universalizing premise, nor the role of race in that history.

What is at stake here is as much an experiential as an ideological point. In liberal societies people are split in two—the abstract individual denuded of all marks that would distinguish it from any other, and the concrete person a human being is in everyday life (Da Matta 1987). Whites in racialized, liberal societies can more or less integrate these two roles. They live with the constitutive tension of liberalism to the extent that no one actually is an abstract individual. But nonwhites face an additional problem: no matter how hard they try, they cannot project a public persona free of that which brands them as being of another race—their appearance as a public marker for a racialized descent. As a consequence, they can never succeed as fully as whites in pretending, under certain circumstances, to be the abstract individual of liberal society. In this situation, nonwhites have little alternative to asserting separate identities in an almost Quixotic quest for a kind of internal coherence, even as these identities are unavoidably organized around the contradiction that called them into existence in the first place. The demands of purity, wholeness, and coherence, in other words, impose heavy psychological and political burdens on nonwhites in social orders dominated by liberal categories. Thus the paradox, after a certain point, of trying to combat racism from a liberal perspective: the very limits of liberalism's ability to confront racism constructively are implicated in the prescriptive norms that mediate the confrontation. This, as Alcoff frames the issue, is why “[w]e need to reflect upon this premium put on internal coherence and racial purity and how this is manifested in Western concepts and practices of identity as a public persona as well as subjectivity as a foundational understanding of the self. We need to consider what role this preference for purity and racial separateness has had on dominant formulations of identity and subjectivity, and what the effects might be if this preference were no longer operative” (1995:261).

A groundbreaking survey of racial attitudes in Brazil may help us conduct a thought experiment along these lines and challenge the social epistemology underlying liberal thinking about race more generally. In 1995, the Folha de São Paulo, Brazil's largest newspaper, together with its associated polling agency, Datafolha, published a nationwide, scientific survey of racial attitudes in Brazil, Racismo cordial (Turra and Venturi, eds. 1995). The largest and most rigorous nongovernmental survey of its kind, Racismo Cordial (Cordial Racism), as it was called, sought to be an antidote to the official census, which had long treated race in strictly and reductively categorical terms. The study's power derives from the fact that it does not flinch from the equivocalities of the Brazilian situation, the central of which has been the gap between what ordinary Brazilians say about race and the hard realities of a society in which prejudice and discrimination are undeniable.

The key to the study is its title. The rich oxymoron of a “cordial racism” seeks to play up the fact that in Brazil race both seems to exist and not to exist, depending on the situation, or the angle from which it is being viewed. The word “cordial” is not casual. It is an allusion to Sérgio Buarque de Holanda's Raízes do Brasil (1948 [1936]), one of the canonical texts of Brazilian national identity. Holanda held that the authentic Brazilian was a “cordial man,” whose virtues and social being were rooted not in legal or institutional norms, but in a “rich and overflowing emotional source” fed by deep rivers of intimate and familial relations. This “cordiality,” he concluded, would be Brazil's contribution to “civilization” at large (1948:219). Of course, polemics have long swirled around Holanda's definition of Brazilian character—not least because he said virtually nothing about Brazil's African heritage. By juxtaposing the cordial Brazilian to the racist Brazilian, the Folha study sought to subvert received wisdom and change the frame within which race is understood.

A glimpse at the data reveals the conundrum investigators faced. Unsurprisingly, the study confirms recent scholarship showing that color prejudice is woven into the fabric of everyday life: nearly 90 percent of those polled answered “yes” to the question, “Are whites prejudiced against blacks?” (Racismo 1995:89–90). Beneath the surface of this disarmingly straightforward “yes,” however, roil deep currents of ambivalence on matters racial. According to the Datafolha survey, Brazilians admitting the existence of prejudice were evenly spread across color lines: roughly nine in ten browns, whites, and blacks agreed that whites were prejudiced against nonwhites. In virtually identical proportions browns, whites, and blacks agreed that prejudice was “considerable” (around 60 percent) or “minimal” (around 25 percent). And yet, equal proportions (almost 90 percent) of browns and whites said that they personally bore no prejudice against blacks, just as 90 percent of blacks denied prejudice against whites. Roughly equal proportions of browns, whites, and blacks (80–85 percent) agreed that there was no difference in intelligence between whites and blacks. Over 90 percent of respondents, regardless of skin color, said they had voted for or would vote for a black politician. Nearly 50 percent of browns (49 percent) and whites (46 percent)—both those admitting prejudice against blacks as well as those denying it—agreed with the idea of employment and educational affirmative action for blacks, a proportion quite close to blacks who did (55 percent). And, to round out this series of statistical snapshots, over 75 percent of browns and blacks said that they personally had never experienced discrimination because of their skin color in employment or educational settings, public places, private gatherings, or jokes (Racismo 1995:88, 96, 102, 163, 167).

The full meaning of these numbers can be made out only in light of the fact that, if the Folha study is to be believed, Brazil is a majority nonwhite nation. Fifty-two percent of respondents self-identified as either light or dark brown (moreno, moreno claro, moreno escuro, or pardo), 39 percent as white (branco), and 8 percent as black (preto or negro). This contrasts sharply with the government's official 2000 census, which claims that 54 percent of Brazilians are white, 38 percent pardo, and 6 percent black (www.ibge.net/home/presidencia/noticias/20122002censo). Divergent methodologies account for the difference: whereas in the official census respondents were prompted by five categories from which they had to choose one—branco, preto, pardo, amarelo (yellow) and indígena (indigenous), the Datafolha survey asked respondents to write in what color or race they considered themselves to be, without any specific prompt.

The tantalizing paradoxes of the survey might be dismissed as proof of how deeply entrenched racism is in Brazil—so deeply that Brazilians, even Brazilians of color, cannot even see, or at least will not talk about the obvious (Twine 1998:150–53). Silence becomes the irrefutable evidence of racism. This is not altogether false. Brazilians are aware of racial discrimination and very often do not talk about it. The question is whether it is possible to make sense of this silence without defaulting to North American notions of race. Perhaps silence is not merely instrumental and defensive, as some have suggested (Twine 1998), but tied up as well with a positive vision of Brazil's possibilities as a nation and Brazilians' hopes as a people. The apparent incongruities of the Folha survey represent an opportunity to pursue this line of thinking, to open up the conceptual space within which it might be possible to take Brazilian racial democracy on its own terms and look toward a history of how Brazilians came to acknowledge the racism of their country without surrendering the idea of a kind of unity. With all the contradictory richness of the survey as a backdrop, this article hints at how such a history might be told from the perspective of Brazil as a place where Western notions of self and identity are ambiguously operative and where people have responded to the dilemma of racism in relationship to, but not strictly from within, the framework of liberal formalism.

Thinking About “Cordial Racism”

What makes the survey's findings so arresting is the uniformity of opinion where one might reasonably have expected a clear split according to skin color. How to explain this?

The survey does not yield easily to a liberal analytical approach. A narrow focus on the denial of individual rights can account for the remarkable uniformity of opinion across class, race, color, gender, and region either by denying people an understanding of their situation—false consciousness—or by claiming that they are so oppressed as to lack the ability to speak their true hearts and minds. The first of these seems manifestly incorrect, given virtually universal acknowledgement of bias and prejudice. The second, though not without an odor of truth, presumes that when people express themselves ambiguously it must be because they are unable to express themselves unambiguously—a presumption for which there is no warrant and much contrary evidence.

The choice is clear: either we reject ambiguity as emblematic of an underlying problem or we allow ourselves to be led into the temptation of “legitimizing ambiguity” in the expression of race and color (Da Matta 1995:283; Fry 1995–1996:131). One way of coming to grips with this problem is to acknowledge the historical centrality of miscegenation in Brazilian conceptions of race and its link to Brazilian national identity. Anthropologist Renato Ortiz has argued that Brazil underwent a crucial transformation in the twentieth century when “[t]he ideology of mestiçagem [miscegenation, in the sense of biological and cultural mixing] . . . came to be propagated socially and became common sense, ritually celebrated in everyday relationships, or at grand events, such as carnival and soccer. What had been mestizo became national” (Ortiz 1985:41). This formulation addresses several critical points. First, by insisting that the idea of miscegenation operated—and operates—in the realm of “common sense” it directs attention to the making of “practical and popular forms of consciousness affecting the broad masses of society” (Hall 1986:20). Second, it highlights the indissoluble relationship between race and national belonging. Finally, and less obviously, it hints that an understanding of Brazilian race involves an epistemological shift away from the hard categories of race toward a concern for mediations in a plural system of social understanding (Da Matta 1995; Vianna 1999). From this perspective, the always-ambiguous historical process of miscegenation and mediation, in tension with categorical certainties, becomes the analytical and experiential core of racial understanding.

In its simplest terms, Ortiz's statement about the ideology of miscegenation restates the problem of “racial democracy”—one of the most exhaustively studied issues in modern Brazilian scholarship. The idea of racial democracy, which dates from the publication of Gilberto Freyre's Casa grande e senzala (Masters and Slaves) in 1933, can be heard in two different registers. The first is lyrical and asserts that Brazil is a land of racial harmony and equality. This strong version of the thesis has had a tough time standing up to the undeniable fact of discrimination in twentieth-century Brazil. The second version declares, less ambitiously—and for that reason more robustly—only that race has contributed little compared to class in the making of social inequality in modern Brazil. However understood, the idea of racial democracy became the “myth of racial democracy” in the 1950s when scholars noted the enormous gap between the stated ideal and the persistence of color prejudice. The revision that followed concentrated on laying bare the intellectual history of racial democracy and showing how far it was from describing Brazil's actual racial situation (Costa 1985).

Though salutary, the revision came at a price. The effort to show that the emperor had no clothes advanced through a series of analytical elisions and exaggerations, neglecting the cultural realities of race in Brazil and granting an all-pervasive manipulative capacity to elites. As a consequence, while much was discovered of how intellectuals and elites have thought about race and nationality, popular attitudes were scanted (Skidmore 1993:xi). Partly this is because among researchers the “myth” came to be so fused with “racial democracy” as to deny the legitimacy upon which any core idea of national identity relies. But merely declaring an idea mythical is no license to ignore it. Ideals acted on and believed in are no less a force in people's lives than other aspects of their everyday existence. The Folha survey, as well as recent scholarship, suggest that Brazilians of all colors understand quite well that racial democracy is a myth—in the sense that the ideal is contradicted by reality—and yet continue to hold it as an ideal to live by, a fact that cries out for understanding.

A further reason not to dismiss the myth of racial democracy is that the word myth does not have to be understood exclusively in terms that demand its rejection. A myth can also be a “true story” which, by locating itself outside of historical time, binds the sacred and the mundane to each other (Florescano 1994:176–83; Eliade 1963:1–20). If in our “secular” age the nation has represented a kind of sacred space premised on the incorporation and neutralization of historical time, then all national identities are myths of union in the face of the dissolving differences of modern subjectivity. Understood in this way, myth stands simultaneously as a privileged remembering and a principled forgetting in the service of common identity. From this perspective, the challenge before Brazilian nation-builders in the nineteenth and twentieth centuries was enormous: how to forge unity from a diverse population at a time when the very idea of nation was thought to require a kind of racial purity Brazil simply could not claim (Schwarz 1993; Skidmore 1990; Stepan 1991). For better and for worse, in other words, Brazilian national identity had to be built on the racial fault line that ran through what was supposed to be the unshifting bedrock of undifferentiated liberal individualism that made the idea of unified nations imaginable.

Alcoff's critique of Western concepts of self and identity offers a means of facing the problem. A half millennium of physical and cultural miscegenation in Latin America, including Brazil, argues Alcoff, have led to understandings and experiences of identity and selfhood very different from those in countries, such as the United States, where miscegenation has been ideologically bracketed. During this historical trajectory realities of mixture, variability, and fluidity have grown up alongside concepts of purity, wholeness, and coherence. Amidst this cultural welter, experiences of identity and self have been decidedly ambiguous and contradictory.

We have barely begun to ask about how this perspective might affect the way we think about race more generally. What, after all, could Brazil's cultural ambiguity on matters racial, its “weak” institutions, and the inconstancy of its liberalism (Mota 1977), have to say to the problem of race? Perhaps the answer is “much.” Could it be that Brazilians have developed resources for confronting racism beyond those of liberalism? To approach this question seriously in the Brazilian context is to treat the idea of racial democracy—a synonym for the relationship between race and nationality—not merely as a myth that elites have manipulated to racist ends but also as an historically accumulated fund of social understanding on matters of color, prejudice, race, and national belonging shared across class and color lines and deeply a part of what it has meant to be Brazilian since the 1930s.

From Racial Pessimism to Racial Democracy—Brazil's Elite, 1880–1940

Throughout Brazil's colonial period and much of the nineteenth century, Brazilian social relations pivoted on the distinction between slave and free. Alongside this fundamental divide, Brazilians of various skin tones and features recognized all manner of physical distinctions and everyday life was suffused with an idiom of color, ethnicity, and nationality that could be baroque in nomenclature. White or near-white elites not uncommonly held darker-skinned Brazilians in contempt and denied them privileges, though this domination was not yet expressed as an explicit ideology of race.

As slavery gradually declined over the nineteenth century, Brazilians of all conditions were forced to contemplate a world without masters and slaves. As authority and hierarchy eroded—and as slaves dropped to under 15 percent of total population by the 1870s—elites worried about how to secure social order against what a São Paulo newspaper characterized in 1880 as a “black wave . . . a horde of semibarbarous men lacking direction” (Azevedo 1987:68). The abolition movement, muted by comparison to its U.S. counterpart, succeeded in passing the Free Womb Law in 1871, sounding the death knell for Brazilian slavery (Conrad 1972). By 1872, three in four nonwhites were free, almost half the total population of 10 million. Immersed in the day-to-day realities of slavery and having relied so long on a legally sanctioned distinction between slaves and freemen, elites lacked a ready framework for making sense of the transformation they were living.

This is the context in which scientific racism, growing out of European alarm over racial differences between colonies and metropoles, led Brazilian elites, with elites all over Latin America, to a nervous fretting over the “degeneration” that was thought to follow from miscegenation (Schwarz 1993:44–70; Stepan 1991:137–39). Where in Europe “the appeal to superiority over other races . . . was projected over perceptions of social divisions at home,” in Brazil the inferior races were found at home and increasingly without any clearly defined social role (Pick 1989:38–39, 41). If “degeneration” among basically white populations explained everything from criminality in Italy, to neurasthenia in the United States, to the unruliness of crowds in France, to a fascination with mysticism and crude naturalism in Germany, what did it imply for Brazil, where skin color distinctly marked social difference and where miscegenation was ongoing (Borges 1993:236–37)?

Faced with this question, some sank into pessimism. Physician Raimundo Nina Rodrigues, himself a mulatto, despaired of Brazil's future because he considered blacks to be hopelessly inferior to whites and, with most European theorists, felt that miscegenation could only degrade the population (Eakin 1985:160). Others groped for alternative formulations. Sociologist Sílvio Romero sought to adjust racial theories to Brazil's heterogeneous reality. Specifically, he exploited ambiguities internal to those theories to deny the idea of unilinear evolution and stages of development that had come to dominate notions of national progress, insisting instead that Brazil could follow its own path of a miscegenation that ultimately would adapt Europeans to the tropical environment. For Romero, and others, European immigration became if not a panacea then at least a plausible solution to many of Brazil's woes.

By the turn of the century elites and intellectuals sensed dimly that Brazil's situation differed profoundly from the historical reality that had given rise to scientific racism in Europe and the United States. Rodrigues' pessimism and Romero's contortions reflected divergent attitudes of those locked within the conceptual prison of racial determinism. European racial theorists were able to condemn miscegenation in strong terms and insist on sharply exclusionary racial categories because, except in colonial peripheries, the relative somatic homogeneity of their populations allowed them the luxury of doing so. The presumed evils of large-scale miscegenation were always in the distant, almost mythical past, impinging on the present only in terms of the problem of ethnic minorities or in terms of regionally-specific or widely diffused traits that needed to be eliminated from the nation's organism (Blanckaert 1988:25). In the United States, the threat of racially dangerous populations and miscegenation were met more brutally—by exterminating the Native Americans and fencing African Americans off from wider society through legal segregation.

Through the first decades of the twentieth century Brazilian elites found no resolution to their dilemma. Many agreed that “whitening” through immigration was the only possible response, though intermarriage as policy ran counter to the assumption that miscegenation could produce only degeneration. “Whitening,” thus, might best be thought of less as a grand statement of national direction than as a position defaulted to from the irresolvable contradictions of racial pessimism. In effect, Romero's recasting of racial theory was preferable to Rodrigues' despair. While it remained deeply tied to the racializing assumptions of its day, the “whitening” thesis represented an opportunity for those who saw race as subordinate to the broader issues of national identity, national progress, and national unity. Between 1900 and 1930, a number of writers weakened the remaining props of scientific racism. They rejected as anti-Brazilian the idea of inferior mixed races and proposed health and educational reforms to benefit Brazil's black and mulatto population. These authors represented a shifting of attention away from the divisiveness of race toward a more forward-looking sense of shared identity among all Brazilians (Skidmore 1993:185–90; Needell 1995).

Gilberto Freyre's publication of Casa Grande & Senzala in 1933 brought an alternative vision of race into focus. His depiction of Brazil as a land where a hot climate, miscegenation, and an absence of violence tended to “dissolve” race prejudice is well known and widely criticized (Costa 1995). I have another concern. If Freyre's work is taken as a form of myth making, as it often has been, then the question becomes one of understanding how the myth came to be so broadly accepted and so enduring, despite—or perhaps because of—the undeniable persistence of actual prejudice.

It is a commonplace that Freyre's portrayal of the colonial fazenda (estate), where whites and blacks discovered ways of relating to each other across the boundaries of race and slavery, found wide appeal in 1930s and 1940s Brazil (Fonseca 1985; Amado 1962:31). The first three decades after 1900 had seen rapid urbanization, high rates of immigration, the creation of an assertive labor movement, the emergence of an urban middle class, and the fractioning of an oligarchic elite along various fault lines. Freyre's vision of a socially harmonious Brazil caused a great many heads to nod in agreement, or at least to hope that he might be right. But the appeal of his idyll ran deeper than political expediency.

The question of how to make Brazil's diverse population into a “people” capable of representing a modern nation dated back to nineteenth-century debates about immigration and race. Following abolition in 1888, growing numbers of Brazilians had begun to struggle with the issue of national unity and national identity. Indeed, elites' efforts to promote European immigration may have been as much “a desperate strategy to create a body politic” as an overt expression of racism (Morse 1982:97). The problem was similar to the one that had given rise in Europe to the very theories of race that ultimately proved so nettlesome to Brazilian elites: how to incorporate distinct regions and people or races into a single nation. As such, Freyre's Casa Grande represented far more than an ideological response to a specific political conjuncture—it proposed a radical reformulation of Brazilian culture.

If between 1900 and 1930 Brazilian intellectuals and writers had gained some wiggle room within the straitjacket of racial determinism, Freyre slipped free of its bonds (which is not to say he escaped racial conceptions altogether). Epistemologically, scientific racism was premised on an illusion of pure types. Centuries of miscegenation, however, had made Brazil a place where that illusion could not credibly be sustained. And since the idea of race is premised on antimiscegenation (Young 1995:19), Freyre's embrace of mestiçagem—sexual and cultural—as the basis of Brazilian social dynamics amounted to a repudiation of Eurocentric racist and racialist theorizing.

According to historian Ricardo Benzaquen de Araújo, this focus on line-crossing and mediations led Freyre to think of Brazil as a country that oscillated eternally between Europe and Africa, taking from both, belonging to neither. The result, in Benzaquen's characterization, was a “hybrid, syncretic, and almost polyphonic” society that had not undergone a straightforward process of Europeanization. In describing the outcome of the unequal coming together of Portuguese, Indians, and Africans, Freyre pioneered a new vocabulary that ran against the grain of accepted notions of national identity. Benzaquen argues that miscibility, flexibility, plasticity, and social adaptability became the watchwords of a Brazil that would not be bound by conventional notions of pure types. By using miscegenation as the defining principle and metaphor of Casa Grande, Freyre effectively enshrined difference, hybridism, ambiguity, and heterogeneity as the determining qualities of Brazilian life (Araújo 1994:43, 44, 47, 53). This allowed him to approach everyday life in terms of a “wealth of balanced antagonisms,” where violence and sexual congress, despotism and fraternization coexisted, and black, white, and brown lived side by side, always in tension but never apart (Freyre 1970:348).

Racial Democracy and Popular Culture in the Twentieth Century

From the vantage of Brazilian elites, Freyre's vision represented a remarkable about-face from racial pessimism. This was no grand ideological coup. Instead, racial democracy is best understood as the outcome of a series of never-fully resolved experiments groping toward a solution to the crisis of order that began with the decline of slavery from the mid-nineteenth century forward, the contingent and wholly unexpected response to the intellectual trap of European and American racial theorizing, which had consigned Brazil to backwardness because of its “inferior”—that is, miscegenated—population. Nor should the global context within which it took shape be forgotten: Brazil began to celebrate its mixed-race heritage, at least symbolically, at the zenith of Jim Crow in the United States and the heightening of racial and ethnic tensions in Europe. If Brazilian elites and intellectuals of the 1940s could make a sense of racial democracy that would have baffled their ancestors two generations earlier, it is largely because by the 1940s even Brazilian elites and intellectuals had had their fill of a theoretical construct that relegated them to second-class status in the concert of nations.

But an understanding of the ideological fashioning of racial democracy after the 1930s—and of the critiques leveled against it—does not bring us any closer to determining how the ideal came to be accepted in the “world according to ‘nonelites’” (Skidmore 1993:xi). Of course, different segments of the population have understood the promise of racial democracy differently. They have done so, I contend, within the constraints of a growing commitment to racial democracy as an ideal defining national belonging, a commitment broad and deep enough to account for the Folha results. How this happened is not a problem amenable to the text-oriented methods of intellectual history, where most of the work on the meaning of racial democracy has been done. Instead, I will turn to the arenas of popular culture, where race, nationality, and a social epistemology of ambiguity have been most lavishly performed.

The preface to this story involves the nearly invisible social mobility of light-skinned, successful mixed-race people beginning in the late nineteenth century. This is a complicated matter because the racial and color classifications that might be used to establish mobility are themselves tangled up with the process they would be used to describe. A critical reading of national censuses, however, provides powerful evidence to the proposition. Between 1872 and 1940, according to official government counts, Brazil's white population increased from 38 percent of total population to 63 percent. In absolute terms, this represented a jump from just under 4 million whites (of a total population of 10 million) to just over 26 million (of 41 million). During the same period, mulattoes dropped from 42 to 21 percent of total population, and blacks from 20 to 15 percent (Ribeiro 1995:229). At least according to the official censuses, Brazil was in fact whitening.

But as anthropologist Darcy Ribeiro notes, no combination of immigration rates and white birth rates can explain this “prodigiously large” increase of whites. The only reasonable explanation, he concludes, is that the census of 1940 reflected a tendency to classify as white all who had achieved a threshold respectability (1995:228–30). This implies a government-sponsored, though largely unacknowledged classificatory migration of light-skinned mulattoes to whiteness. “Whitening,” in other words, could be many things to many people. It suggests a widely shared and studiously unarticulated coincidence of interests between mulattoes and white or near-white middle-class people who did not want to be confused with blacks, and with political elites still captivated by the dream of a whitened Brazil. If this is right (and much more research is needed), then the ideal of whitening, freed from the constraints of scientific racism, linked in popular imagination and practice with social mobility, and rarely discussed openly, could have represented the spark of a will to unity and a collectively-imagined convivência (intimate sociability) among diverse segments of the Brazilian population.

Yet, given its exclusionary history and its appeal to individual fortune, a pragmatic and sub rosa whitening hardly seems the basis for a strong sense of national belonging (Fernandes 1978). This is the context in which racial democracy's inclusive insistence of a deracialized Brazilianness—which bubbled up from the racially mixed lower classes and percolated down from elites (Fry 1995; Castro 1995)—gained purchase in everyday life. Between 1920 and 1960, several different aspects of Brazilian culture and life—carnival, samba, soccer, and popular religiosity—came to be widely seen as expressing Brazilian national identity. The unity and convivência implied in these expressions of popular culture converged with the powerful integratory message of racial democracy, so that for elites and non-elites alike, some participation in and emotional connection with these activities, across racial lines, came to define a large part of what it meant to be Brazilian.

The history of samba and carnival, perhaps the richest vein of literature about Brazilian popular culture, provides some clues for approaching the problem. Although there is a long history of Afro-Brazilian street festivities, especially in Bahia (Lara 2002; Reis 2002), carnival was a relatively limited affair with little Afro-Brazilian influence through the colonial period and the first decades of the nineteenth century. After the mid-nineteenth century, when masked slaves participated in Lenten street celebrations, and especially toward the end of the Empire in 1889, carnival took on a new importance, particularly in Rio de Janeiro (Cunha 2002). Masked balls of elites became far more elaborate than earlier in an effort to emulate European celebrations. And with the street pranks of Lenten entrudo, which had long involved the throwing of perfume, or vinegar-filled bombs in a “war on the top hats,” outlawed in 1853 and successfully campaigned against in the press and repressed by the police during subsequent decades, others began enthusiastically to move organized festivities out of doors again. In the 1890s, new carnival societies pioneered street parades in which members sported allegorical costumes and performed skits and dances, some of which poked fun at the government. Worries about the menacing mass of blacks in the streets did not prevent Afro-Brazilians from coming out, but rather led them to devise alternative means of participating in carnival (Cunha 2002:95). By the turn of the century, some Afro-Brazilians had formed their own societies, which quickly became targets of police repression. Concerned that carnival was being “Africanized,” elites sought to prohibit African drumming and religious practices as “incompatible with our current civilized state” (Rodrigues 1977:177).

Nor was this fear limited to whites. More prosperous blacks and mulattoes—artisans, skilled workers, professionals, and government functionaries—looked with disdain on the antics of the poor, whatever their color. Many of these joined with other artisans and skilled workers in 1907, whites as well as mulattoes, to form a new type of carnival society known as a promenade. Seeking respectability among Rio's more established carnival societies, the Delightful Myrtle (Beija Flor), for instance, allowed singing but no African chanting, drumming but only as a gentle background to flute solos accompanying paraders who dressed in European fashion (Guillermoprieto 1991:25). Their allegories took on broad themes, from the planetary system to the brotherhood of nations, winning accolades from the press and politicians. Not to be outdone, poor carnival organizers in the mid-1920s decided to establish their own societies so that they might parade with better-off whites, blacks, and mulattoes who now dominated street carnival. In the following years, poor blacks and mulattoes formed numerous “samba schools”—escolas—across Rio's slums, which before slavery's abolition in 1888 had been the hiding places of Afro-Brazilian liberty (Chalhoub 1990:233–48). While Rio's established carnival societies sniffed at these upstarts, less than a decade later the new escolas had become so well entrenched that the city gave up trying to do away with them and decided to regulate them instead. Among the most telling of the new rules was one banning non-Brazilian carnival themes.

At around this same time in the early 1920s, a group of musicians were gathering at the home of Aunt Ciata on the outskirts of the old city in a part of Rio de Janeiro known as Little Africa. Born to a prominent and well to do black or mulatto Rio family, Aunt Ciata married a doctor. A lavish entertainer, she became a patron to many of the musicians who invented modern samba. She held formal ballroom dances but also African drumming sessions (Moura 1983). Composers, songwriters, and performers met at her house, creating a musical style that swept Rio. By 1926, samba was so widely admired that white intellectuals were trying to meet the men who were making it all happen. Profoundly dissatisfied with Brazil's cultural dependence on Europe, young writers such as Sérgio Buarque de Holanda and Gilberto Freyre, and classical composer Heitor Villa-Lobos, were casting about for a new sense of what it meant to be Brazilian. They wanted nothing more than to talk to these dynamic black and mulatto sambistas who had created a new form of authentically Brazilian music. Freyre, when he finally met three of the most prominent sambistas, saw them as representing “the great Brazil that is growing half-hidden by the phony and ridiculous official Brazil where mulattoes emulate Greeks . . . and caboclos try to appear Europeans and North Americans; everyone looking stupidly at things Brazilian . . . through the pince-nez of a Frenchified doctor of laws” (Vianna 1999:8–9). Ten years later radios were filled with samba and record companies could not produce new collections fast enough. Samba had arrived (McCann 2004).

By the mid-1930s, both carnival and samba had come to be seen as expressing aspects of Brazilian nationalism. President Getúlio Vargas saw samba as a tonic to regionalism and undertook to have it broadcast on short-wave radio, reaching the remotest corners of Brazil. Contests for carnival music and competitions among the samba schools, dating from the early 1930s, quickly became big business and were subsidized by the government by the late 1930s. Samba's popularity led government propagandists to try to regulate the content of samba lyrics, especially those connected with carnival. Starting in 1939, the Ministry of Propaganda sought to play down the satirical, drippingly sarcastic samba in favor of a samba that could serve the government's modernizing thrust. Where independent sambistas emphasized the character of the malandro, a kind of urban black or mulatto trickster who lived by his wits and charm at the margin of legality, the government pushed lyrics praising hard work, discipline, and sacrifice (Matos 1982:90–97). These efforts met with very little success, as poor sambistas resisted ideological incorporation. But the effort bespoke the cultural importance samba and carnival had attained just as Freyre's notion of racial democracy was blitzing across Brazil's collective consciousness.

Samba, carnival, and racial democracy came together so powerfully during this period because they offered something to almost every segment of a fractured social order. A new urban elite of industrialists, upper-echelon government bureaucrats, professionals, and intellectuals, many still new to their status, was less invested in traditional notions of respectability and lost no love on the pretensions of the old cultural guard. And even within the traditional elite, intellectuals and artists had expressed a fascination for popular music dating back to the colonial period (Vianna 1999:17–31). In the context of a loosened European cultural hegemony after World War I, samba seemed to many an expression of the Brazilian soul long covered up. Poorer, darker-skinned Brazilians also welcomed racial democracy's repudiation of scientific racism. In the context of a Brazil that now officially embraced aspects of Afro-Brazilian culture, the samba schools could see themselves as the culmination of a process that had made carnival more inclusive. The government was dishing out money, politicians were presiding over carnival parades, and officials were visiting Rio's poorest favelas to speak about the future of Brazil's racial democracy, a project in which the samba schools were promised an important role. Middle-class Brazilians, too, had reason to celebrate racial democracy. They were the mass reading audience for Freyre's book. They understood, perhaps more acutely than most, given their recent arrival to respectability, that close scrutiny of their family tree might reveal a “foot in the kitchen”—an African ancestor. Racial democracy legitimized, or at least softened the impact of that uncertainty. Moreover, long experience of carnival, through the societies of the late nineteenth century, and high rates of radio ownership, exposed them to a torrent of popular culture, including samba.

A sectoral analysis goes only so far. A fuller explanation of how and why race and nation came to be so closely linked in popular imagination may depend on a deeper understanding of the role intermediaries and cultural mediators played in opening up spaces where recognition, understanding, and even admiration could flourish. As anthropologist Hermano Vianna has noted, there has not been a period in Brazilian history since independence in the early nineteenth century without significant cultural contacts between elites and non-elites, especially in the area of music. Mulatto editor and poet Francisco de Paula Britto had brought together politicians, poets, playwrights, artists, and travelers at his printing shop during the 1830s and 1840s. Alberto Brandão, an elite Carioca, had invited folk musicians for performances in his home in the early twentieth century. And Aunt Ciata had midwifed samba's birth from her parlor in the 1920s. We know next to nothing of such figures, and the many less-storied ones who “shuttled back and forth among social groups, putting them in contact with each other, constantly redefining boundaries between them and remodeling the patterns of their collective life” (Vianna 1999:115). It does not seem at all implausible to think—and this is where research is required—that one of the reasons samba, carnival, and racial democracy came together as they did is that the institutionalization of carnival, the commercialization of samba, and the fact of government involvement with the samba schools multiplied the number of such mediators, increased the density of connections among them, and broadened the possibilities for meaningful contact (McCann 2004).

The bond forged between race and nationality during the middle decades of the twentieth century extended beyond carnival and samba to soccer. The basic outline of soccer's history in Brazil is clear. It was introduced in the 1890s by English employees of a railroad company. By the turn of the century it had spread to German and Italian immigrant communities in Rio and São Paulo. Up to 1930 it enjoyed an amateur phase. Brazilians, including blacks and mulattoes, played in the streets, using orange peels, rocks, or any rag filled with paper. Increasing numbers began to play on official pitches, as soccer clubs grew (Pereira 2000:110–34). As early as 1914, a Rio newspaper marveled over the “great happiness that had possessed the people of Rio” when a Brazilian national team defeated an English side 2 to 0, demonstrating Brazil's “unity,” as a journalist put it (Pereira 2000:141–42), even though blacks were excluded from the Brazil eleven. Despite this exclusion, blacks and mulattoes continued to root for the national selection, though underlying racial tensions remained (Pereira 2000:170–75). Throughout the late teens and 1920s, soccer became ever more popular a pastime among workers and within labor associations. In 1932, two black players instrumental in Brazil's victory over Uruguay, were carried through the streets of Rio and heralded as national heroes and the Flamengo club included several black footballers on its roster for a traveling team.

The years between 1933 and 1950 marked the professionalization of the sport and the institutionalization of the soccer clubs that became so important a focus of local identity in Brazil's largest cities (Levine 1980). As inter-club rivalries intensified, the informal prejudice against black and mulatto players began to fall to the demands of competition. Teams began to scour the slums for talent. Elites remained in firm control of the clubs, but membership expanded dramatically, especially among middle-class people. By 1940, radio broadcasts took the game to all corners of Brazil and hundreds of clubs produced thousands of aspiring stars. Newspapers reported the games in detail, creating “shared images and a new form of knowledge without connotations of class [and perhaps racial] privilege” (Rowe and Schelling 1991:139). As with samba and carnival, soccer came under government regulation by the end of World War II.

From the 1930s forward, soccer came to be deeply connected to Brazil's international self-image. World Cup play put Brazil on display before European teams, often to Brazil's advantage. Not surprisingly, the delicate fugue of race and national pride was played out through the national selection (Rodrigues Filho 2003). Nonwhite players were commonly kept out of important international matches during the 1930s and 1940s, particularly for games against European (white) teams. Yet, by the 1938 world championship, where the selection comprising several black players performed very well against Europe's best, a cartoonist from the middle-class magazine Careta depicted an Africanized mulatto striker, Domingos, towering above Hitler, saying that he did not believe in that story—“nessa historia”—of racial superiority (Pereira 2000:342). Even so, as late as 1958 there was some question whether the selection should appear at the World Cup with a mostly nonwhite side. But on this point the top hats of the Brazilian Sports Federation were out of step with the public at large: repeatedly during these decades, the Federation bowed to public pressure to field Brazil's best to play against the world. By the 1950s many, sometimes most of these players were black or mulatto.

It was at this time that a distinctively Brazilian style of soccer developed, one widely recognized as expressing the Brazilian soul. Fluidity, improvisation, elegance, and joyful glee were its hallmarks. Freyre argued that it represented Brazil's “Dionysian,” mestizo character (Freyre 1964). It might just as easily be seen to express the experience of improvisation and uncertainty of poor blacks, mulattoes, and whites who, like the malandro, had learned to struggle at the margins of Brazilian society (Rowe and Schelling 1991:139). Either way, this style came to be heralded as the Brazilian way of soccer, most fully expressed by Brazil's mulatto character, and most evidently a matter of pride in confrontations with other nations.

Like the samba schools and carnival in general, soccer and its clubs may have broadened the field of operation for cultural mediators. As clubs began to send scouts to working-class neighborhoods and as new players gravitated to the clubs the fascination for novel jogos de cintura, or the shot with the tightest curl, would have brought people together who might otherwise have remained divided by race and class. This sense of connection may well have been projected to a much higher level. In the context of a government officially promoting national unity through racial democracy and given the emergence of a uniquely Brazilian “mulatto” style (Freyre's introduction to Rodrigues 1964), it does not seem hard to believe that winning and losing took on a meaning far beyond that of a mere sporting event. Perhaps, as one scholar has suggested, losing came to be understood along the lines of “we lost because we lacked the unity as a nation to win” (Ceasar 1988:281). The desire to win and affirm a distinctively Brazilian identity, thus, might be seen as having promoted an always tenuous unity over the divisiveness of color prejudice—never far beneath the surface—for all who cared about the games' outcomes. At the level of collective subconscious, unity as expressed through soccer may have become so imperative that any threat to it from overt prejudice could be seen and experienced as an affront to the nation.

Another arena where racial democracy came to be expressed as “common sense” is Umbanda, Brazil's distinctive spiritist religion. As with carnival, Umbanda's roots can be traced to a convergence of European and African influences dating to the nineteenth century. Africans transported to Brazil as slaves brought with them a broad array of religious traditions. Where large numbers of slaves from diverse backgrounds congregated, especially in large cities such as Rio and Salvador, eclectic practices developed. By 1880, there were well-established sects combining the African orixás and Catholic saints. Many rituals involved spirit possession. At roughly the same time, white Brazilians were experimenting with a form of French spiritism known as Kardecism, after Léon Rivail, a Paris schoolteacher wrote a series of books he insisted were the psychographed communications of a Druid spirit that identified itself as Allan Kardec. Kardecist mediums claimed to incorporate famous historical figures, including Voltaire, Napoleon, Plato, Caesar, and Confucious.

According to some accounts, Umbanda can be traced to a 1908 séance among middle-class Kardecists in which the spirit of an Indian boy tried to embody a medium and was told to go away. Rebuffed but not discouraged, the spirit possessed an infirm teenage boy attending the séance and vowed to found a religion in which Indians and Old Black Folks (the spirits of Brazilian slaves) would have a place. A different myth of origin claims that Umbanda came to exist in 1920, when a middle-class Kardecist was possessed by the spirit of a Jesuit priest who directed him to found a new, truly Brazilian religion organized around the worship of Brazilian spirits—Caboclos (spirits of Brazilian Indians) and Pretos Velhos (Old Black Folks). The main point here is not to choose between these myths, but to notice that both involve a crossing of traditions and an appropriation of African themes by white spiritists, just as issues of race and national unity were becoming matters of deep concern for many Brazilians.

Through the 1930s Umbanda remained a small sect that suffered the same repression other Afro-Brazilian religions did, notwithstanding the number of whites involved. Its members hailed mostly from an emergent middle class of white or near-white commercial clerks, government bureaucrats, military officers, and professionals. Many were Kardecists who had begun to visit Afro-Brazilian religious centers in search of spiritual solace and cures more efficacious than the arid didacticism of dead European philosophers and generals. Attracted by the spirits of Afro-Brazilian religions, they were repulsed by such practices as animal sacrifice, demon possession, and heavy drinking. In 1941 they held their first congress in Rio and began to codify Umbanda, seeking to establish doctrinal orthodoxy and to blanch Umbanda of most of its African coloration, except for the Caboclos and Pretos Velhos. This last point was a sticky one, given the spirits' origins. New, mostly white adepts finessed the matter in their 1942 statement of principles by arguing that although the Caboclos and Pretos Velhos were often taken for simple, they “show[ed] a degree of cultural and spiritual evolution superior to that of Western civilization.” Here is evidence that even ordinary people felt weighed down by Brazil's cultural dependence on Europe and the United States. A further notable point is how closely Umbanda paralleled Freyre's hopes for a miscegenated Brazilian civilization. In 1940, an early Umbanda leader published a book entitled A Genuinely Brazilian Religion, in which he asserted that the Brazilian was the “son of three races: the White, the Negro, the Indian,” and insisted that “the Brazilian must be destined an eclectic religion, whose principal characteristics are charity, humility, and tolerance for the immense ignorance of mankind, and which will unify the experience of the White, the tradition of the Indian, and the magic of the Negro” (Brown 1994:42, 48).

By the early 1950s, Umbanda had achieved considerable renown in Rio. Its popularity led Afro-Brazilian religious leaders to claim, in direct response to white Umbandistas, that Umbanda was and always had been an African religion (Brown 1994:46–47). This campaign to claim or reclaim Umbanda can be understood in different ways. It might be seen as a moment when black leaders sought to preserve a space of Afro-Brazilian cultural autonomy from encroachment by whites. But it might also be that they were motivated less by a concern for cultural autonomy than by a desire to ensure that Umbanda's African roots not be forgotten at the moment it was becoming an expression of national identity. Against white leaders who wanted an etiolated Umbanda, they were asserting Umbanda's, and thus Brazil's, distinctively African heritage, in keeping with the government's welcoming of such an identification. The struggle, in other words, might be seen as having been premised on a strong desire to participate in a shared national identity.

This view of the matter may cast some light on Afro-Brazilian leaders' repudiation of what they referred to as “an ugly campaign” of “racial separatism” in the early 1950s. “Our historical traditions,” said one leader, “have always rejected . . . efforts to divide the population into whites and blacks” (Brown 1994:115). On this view, the disagreement over Umbanda's African roots could be read as a sign not of racial separation—though sharp tensions remained—but of a fundamental commitment to the idea of national unity and the possibility of convivência over and above racial divisions. Perhaps this is why in the late 1950s white and black Umbandistas gave up their feud and worked out a rapprochement. According to Diana DeGroat Brown, Afro-Brazilian Umbandistas may have reached this position in part because of “promises of protection, social legitimacy, and potential political advantages to be had from an alliance with middle sector leaders” (Brown 1994:155–57). It is worth noting that other forms of Afro-Brazilian religion, especially Candomblé, have long posited a less ambivalent relationship to their African pasts than has Umbanda (Ferretti 1999; Teixeira 1999).

As with samba, carnival, and soccer, a fuller understanding of cultural mediators is crucial to making sense of Umbanda as a lived emblem of racial democracy and national identity. Umbandista mediums are regularly possessed by spirits whose race differs from their own. White mediums incorporate Caboclos and Pretos Velhos, and may dispense spiritual advice to white, black, and mulatto petitioners. Black and mulatto mediums, whichever spirit possesses them, may find white supplicants asking for counsel or cure. Umbanda's cosmology and principal practice, in other words, take hierarchical inversions and variable identities for granted, and in so doing articulate an alternative to racial and racist discourses (Burdick 1993:15, 225). And after the 1950s reconciliation of Umbanda's white and black strains, it appears to have become far more likely for the faithful to approach their religious practice as an unstable amalgam of white and black Umbanda. It was not uncommon for practitioners to tack back and forth between spiritist disobssession and quimbanda, black magic, depending on what was called for (Hess 1995:195–99). Here mediums have been active agents of cross-racial contact in the context of a system inherently ambiguous in its understanding of race, though far from innocent in drawing racial distinctions.

Sources of Identity and Belonging

The idea that samba, carnival, soccer, and popular religiosity have played a special role in Brazilian cultural life during the twentieth century is hardly a novel one (Carvalho 1987:163). Anthropologist Roberto da Matta has argued that these cultural expressions are “so basic in Brazil [that] in contrast to certain European countries and North America, our sources of social identity are not [I would say ‘not exclusively’] institutions central to the social order, such as laws, the constitution, the university system, the financial order, etc., but rather certain activities which are taken as secondary sources of identity in the center and dominant countries” (1982:60). What might this mean for racial democracy as a source of identity and belonging? The answer may lie in the fact that studies of race in Brazil have tended to focus precisely on the arenas of social interaction—political groups, the state, social movements—that have been least important for understanding how ordinary people live their everyday lives.

The concern dominating many studies of Brazilian race, sometimes consciously, sometimes less so, is the mystery of why so large an Afro-Brazilian population with a history of slavery and undeniable discrimination has not created an American-style civil rights movement. Framed almost exclusively in terms of parties, movements, and citizenship, this question has tended to slight or delegitimize the role of popular culture in shaping attitudes toward race. This scholarship has shed light on the exclusionary nature of official institutions and elites' tremendous political and cultural power (Hanchard 1994). But such studies say little about why ordinary Brazilians, mulatto, black, and white, elites and non-elites, might have embraced the idea of racial democracy in the face of persistent discrimination (Goldstein 2003:107–8, 130). This is fundamentally an historical question. Samba, carnival, soccer, and Umbanda did not achieve their importance only because of a cultural predisposition: they developed as alternatives to official institutions that have tended to operate in an exclusionary fashion.

Conventional approaches to politics, which see citizenship narrowly in terms of the effective exercise of constitutionally guaranteed political and civil rights, are inadequate by themselves to exploring this issue. Before 1889, the Empire's political system embraced a larger proportion of voters than was true of the seventy years following (Graham 1990). As elsewhere in the Americas, ordinary people in the nineteenth century, including slaves and freed people in Brazil, used liberal ideology and electoral processes to secure greater freedoms and open political space for themselves (Castro 1995; Sabato 2001). Although balloting not infrequently involved violence and was as much a celebration of patronage and hierarchy as a recognition of an individual right to vote, the fact remains that a sizeable proportion of the nation's people—perhaps fifty percent of adult males, irrespective of skin color—participated in electoral politics (Graham 1990:103–9). Indeed, it was not uncommon for mulattoes to be elected to public office, including the national parliament.

This was the period, according to Hebe Castro, in which the free mulattoes and blacks of the rural Southeast, well before abolition in 1888, confronted elite efforts to articulate their vision of Brazilian reality to European racial theories with a discourse and practice of liberty as “essentially nonracial.” To an emerging ideology of racial exclusion and whitening, they did not counterpose a compensatory racial identity as blacks, but instead rejected the premise of separate racial identities (Castro 1995:403–5; Fernandes 1978:403). As historian José Murilo de Carvalho has shown, these were precisely the people, as well as the poor more generally, regardless of race, who were largely banished from politics by the Republic after 1889 (Carvalho 1987; Castro 1995:315), allowing elites to conclude, as Alberto Torres noted at the beginning of the twentieth century, that “this state is not a nationality, this country is not a society, the populace is not a people, our men are not citizens” (Luce 1998:34).

Brazil's political arc over most of the twentieth century did little to reward a deep or abiding faith in the possibilities of political participation among ordinary people. Political citizenship conventionally understood along liberal lines has always been in crisis, from the mid-1890s to 1930 when rarely more than 1.5 percent of the population voted and winning margins generally topped 90 percent, to the unrealized promise of the Revolution of 1930 to put an end to oligarchical rule and broaden the franchise, to the sharp battle between populists and antipopulists during the 1940s and 1950s, to the military dictatorship after 1964. If anything, this twisting historical trajectory could be thought of as having engendered a profound ambivalence about politics as such at the very core of modern Brazilian political culture (Owensby 1999:233–35, 241–43). It avails little simply to bemoan the denial of citizenship, a reflex that I believe makes it almost impossible to bring Brazil's twentieth-century political experience into focus.

It is against this background that it makes sense to think of expressions of popular culture as sites of “unofficial citizenship,” places where people could avoid entanglements with a politics that so often excluded them (Shirts 1988:107). Locally rooted, samba, carnival, soccer, and Umbanda were tightly tied into national identity. Though hardly free of racial tensions, they represented a kind of option against a politics that had so long opted against so many, blacks, morenos, and whites alike. This suggests the possibility that large numbers of Brazilians have not felt diminished in their sense of national belonging for having been excluded from effective political participation, or at least not more so than the vast majority of their compatriots. To be Brazilian, from this perspective, is to rise to a moral plane above the pettiness, corruptions, and exclusions of politics. Perhaps more precisely, it is to build a firewall between a positive sense of national identity, experienced as an always yearned-for unity, and the uncertainties of official citizenship, experienced in many different ways, including racially, as fragmentation and exclusion from political participation.

If this is right it could have broad implications for understanding the staying power of racial democracy in Brazil—and why it has proven such a puzzle for scholars, especially North Americans. U.S. political culture pairs official citizenship and national identity: to be an American is to consider oneself a citizen as defined by constitutional norms and a legal system. For Brazilians, whatever their color, this may not have been possible through much of the twentieth century. Rather, among poor morenos, whites, and blacks, and even many middle-class people, national identity may well have connected, among other things, a sense of limited participation and a shared consciousness of having constantly to negotiate between the “official” and “unofficial” sides of their lives.

Recent anthropological work has focused attention on this negotiation by looking at the jeitinho. The jeitinho has been recognized since the 1960s as a means by which individuals faced with a difficult bureaucratic situation, something like a catch-22, can obtain what they want by asking someone, usually a bureaucrat, to relax or break a rule impeding an obviously desirable or just result. The asker makes a personal appeal to the functionary, who is officially charged with uniformly enforcing a set of routinized norms, to bend the rules. Much depends on the attitude of the supplicant. Haughty petitioners are usually denied, with the reminder that it would be unfair to change the rules for a single person. Those whose mien is sympathetic, kind, and egalitarian, who project a sense of trust and helplessness, who can allow the bureaucrat to see the human side of the problem, more often succeed. The jeitinho's history is hazy. Suffice it to say that it appears to have entered everyday language and practice sometime after 1930 and to have become a recognized social phenomenon in the 1950s. By the 1960s it was being discussed in texts on law and administration (Barbosa 1992a: 11–29, 139–47).

What defines the jeitinho as performance is a deep sense of moral equivalency—tomorrow it could be me—for there is no immediate gain from granting a request. According to anthropologist Lívia Barbosa, the jeitinho triggers a “diffuse sense of reciprocity” without implying a permanent hierarchy between granter and taker (Barbosa 1992b). Status, money, personal relations, education, color, and race all matter to the very personal issue of whether the request will succeed, but they tend to be seen as secondary to the quality and circumstances of the request itself. The jeitinho, therefore, is defined by its “universally democratic character” precisely because of its ability to counteract the arbitrariness built into bureaucratic structures and because it can overcome differences rooted in the many social inequalities of modern Brazilian society. As such, the jeitinho is a ritual of egalitarianism that operates above (or perhaps beneath) and beyond the rules defined by law and bureaucracy: true democracy rests as much in the moral obligations of personal relationships as in impersonal norms (Barbosa 1992a; 1992b).

Like samba, carnival, soccer, and Umbanda, the jeitinho may have opened up spaces where the experiences of everyday life came to be linked to the idea of an inclusive national identity. In popular conception since roughly the 1960s, Brazilians across class and race have referred to the jeitinho brasileiro as an ambivalent manifestation of Brazilianness (Barbosa 1992a). On the one hand, it implies precisely what Brazilians find most disagreeable about their own country, the sense that “official” Brazil does not work as it is supposed to. On the other, the jeitinho is a source of pride in the very “unofficial” interpersonal cordiality that is seen as an expression of Brazil's deepest identity. This pride may be a reflection of a consciousness among Brazilians that tensions between the formal and the informal have created a situation in which “the people” have not been inclined to invest themselves exclusively in institutional politics as a source of identity because avenues of meaningful participation were so straitened throughout the twentieth century. Instead, samba, soccer, and carnival became the basis of a “community of sentiments” that, at its best, seeks to transcend Brazilian social differences, including racial ones (Carvalho 1987:160).

Virtually everyone is a potential mediator between these two facets of Brazilianness, both askers and granters: a low- or mid-level bureaucrat or clerk in a private or public bureaucracy—growing numbers of whom were nonwhite after 1950—have had endless opportunities to grant or deny requests for jeitinhos. And given the bureaucratization of social security, labor relations, and expansion of the banking system, most ordinary Brazilians have had endless reasons to request them.

Toward a History of Mediation and Beyond

The advantage of locating samba, carnival, soccer, Umbanda and the jeitinho in the period between 1920 and 1960 is that we can glimpse broader historical patterns that have eluded analyses exclusively concerned with the institutions of government, politics, and the state. Taken together, what these various forms of popular culture suggest is that mediation itself has been a shared experience that, at its best, has sought to transcend the divisions and limiting norms that might otherwise have led people to perceive difference as requiring separation. If in Brazil democratic sentiment is expressed at the capillary level in the inclusionary nature of carnival, in the joyous freedom of soccer, in the fleeting inversions of Umbanda, and in the equality and reciprocity of the jeitinho, rather than exclusively by the legal order, then to deny the possibility of mediation on which that sentiment is grounded by, say, admitting to racial prejudice, is to deny one's own Brazilianness. Racial democracy and prejudice can coexist because they operate according to different logics and because there is no social rule indicating how they should relate to one another. Popular culture represents one of the ways in which ordinary Brazilians have mediated the tensions between these logics without finally reconciling them, which would require a starker, perhaps exclusionary choice.

This points toward an explanation (beyond charges of hypocrisy) of how respondents to the Folha survey could recognize the fact of prejudice but insist that they personally were not prejudiced: not only does moral egalitarianism rooted in “a common biological membership in the human species” weigh against overt prejudice, but admitting prejudice denies Brazilianness, an identity defined by an integratory ideal that reaches across racial divides (Barbosa 1992b:44). It might even shed light on why so many blacks who recognize structural prejudice denied ever having been discriminated against personally: to admit to prejudice, even against oneself, would be tantamount to a symbolic self-exile from the nation. For them, as for so many Brazilians, the nation may not be conceived of or experienced solely in terms of the institutions of liberal citizenship. Their Brazil came to be the parallel universe of popular culture that has felt meaningful in the context of Brazil's political history in the twentieth century. And because Brazilians—poor, working-class, and middle-class people, morenos, blacks, whites, and even elites—have sambaed, joined carnival, rooted for the selection, sought the intercession of spirits, and relied on the jeitinho, this convergence of national belonging and the variegated identitary possibilities offered by a popular culture rooted in the experience of mediation has been broadly persuasive. In this way, Brazilians have faced the quotidian indignities of ongoing racial discrimination without abandoning the idea of a unity and convivência capable of transcending racial difference. In essence, they are people who, having lived in full view of liberal formalism and so often experienced its rebuff, found hope and part of an answer in human solidarities and participation expressed beyond the confines of liberal redress.

To tell the story of how this happened we must shift away from explicating race as an abstract category and undertake a broader inquiry into the history of how race came to be bound up with other elements of social life. Cultural and social mediators, largely invisible in a scholarship dominated by methodological individualism, would play a prominent role in this story. Further research may reveal that samba, carnival, soccer, Umbanda, and the jeitinho, always verging on being clichés of Brazilian social relations, are but outcroppings of a richer and deeper history of mediation that we have not yet begun to understand. Such a history would concern itself with capoeira (a Brazilian martial art), from its link to politics in the nineteenth century to contemporary middle-class enthusiasm for it, with the invention of Brazilian national cuisine and specifically of feijoada as the national dish, with the making of a dark-skinned Virgin Mary as a symbol of national identity (Almeida de Souza 1996), with labor unions, with the always tense relationship between regional realities and national ideals (Weinstein 2003), with various artistic movements, with the everyday centrality of fleeting social contacts, whether the moral reciprocity of the jeitinho (Barbosa 1992b) or the role of humor and laughter in interpersonal relations (Goldstein 2003), with the practices of patron-client relationships in a competitive social order, with the social meaning of the botequim or neighborhood bar (Fry 1995–1996:134), and with the reality of passions, affections, and sex across racial lines (Caulfield 2003). It would pay closer attention than we have to the world of the lower classes, how they have lived and came to embrace a social logic of multiplicity rather than one of strict bipolarity, and how that understanding became national “common sense.” Doing so might force open the analytical and imaginative space of the political, broadening our sense of what can count as meaningful participation, political and otherwise, in the life of the nation. Such an approach does not imply an out-of-hand rejection of liberalism as long-term political project; it calls simply for a willingness not to be trapped within its gilded ideological cage. The promise of seeking a new path is that we may be able to confront Brazilian racism on its own historical terms without having to deny the powerful integratory sensibility or will to unity that runs parallel to it.

From a certain perspective it may be tempting to conclude that the Brazilian experience has no lessons to offer the United States. After all, there is no obvious program implicit in the history of racial democracy, and discrimination in Brazil persists and may even be worsening (Andrews 1992; Guimarães 1999). And yet, some in the United States worry that theorizing about race has come to be “marked by a deep discontent with liberalism” (Delgado 1995:1). Incrementalism and tokenism bespeak liberalism's limits in confronting racism. As a result, writes legal scholar Richard Delgado “[i]n the short run liberalism will continue to decline, and nothing coherent will replace it” as an organizing framework for thinking about race and racism (Delgado 1995:469). Whether or not the sky is actually falling, there is no denying a sense of frustration with regard to the struggle against racism in the United States.

Perhaps this is how we come to understand that it is time to broaden our sources of response to racism, in Brazil, the United States, and beyond. Liberal values and norms are indispensable. But they are not the only ones available. Nor need they be understood strictly in terms of a theoretical purity extruded from European and North American historical experiences. Indeed, there is no a priori reason to exclude the possibility of a liberalism loosened enough from its tight categorical moorings to float more freely with the reality of racial democracy's aspirational unity. For if the Folha survey is right, Brazilians across class, color, race, gender, and region have eschewed a strict “preference for purity and racial separateness” (Alcoff 1995:261), opting for compensating solidarities expressed as national common sense.

This is not to say racism and racial thinking have vanished in Brazil. Nor is it to say that Brazilians have abjured liberal categories root and branch—as contemporary Brazilian politics shows. But the Folha responses remind us that human beings are capable of quite diverse notions of identity and self beyond hard-edged exclusionary ones. And Brazilian racial democracy, for all its failings in practice, represents a coherent experiential and identitary response to the fear of race mixing, one that might help in thinking about a U.S. “racial ecology” far more heterogeneous than has generally been admitted (Root 1992:4). After all, if the Brazilian “dream” of racial democracy is unrealized, and to that extent a myth, so too is the United States' “dream” of liberal egalitarianism, given the great gulfs of inequality in contemporary American life. Yet it rarely occurs to anyone to ask why Americans should continue to believe in the “dream” of that egalitarianism, or why those most critical of actual inequalities, and especially racial ones, so often fail to see that the liberal framework within which they articulate their critiques implicates the very problem they are trying to solve. At the very least, the Brazilian experience invites us to consider that a heartfelt and culturally pronounced desire for unity and convivência through social mediation can be a counterpoise to forces that might otherwise drive people to poles of racial separation—the originary impulse that has defined racism and racial thinking all along.