Book Search:  


Google full text of our books:


Race in Another America:
The Significance of Skin Color in Brazil
Edward E. Telles

Book Description | Reviews

COPYRIGHT NOTICE: Published by Princeton University Press and copyrighted, © 2004, by Princeton University Press. All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher, except for reading and browsing via the World Wide Web. Users are not permitted to mount this file on any network servers. Follow links for Class Use and other Permissions. For more information, send e-mail to

This file is also available in Adobe Acrobat PDF format

Chapter 1


RECENTLY, the president of the United States asked the president of Brazil, "Do you have blacks, too?"1 Unbeknownst to President Bush and many other North Americans, that South American country currently has more than three times as many inhabitants of at least partial African origin as the United States. Both the United States and Brazil were colonized by a European power that dominated militarily weaker indigenous populations and eventually instituted systems of slavery that relied on Africans. In the Brazilian case, European colonists and their descendants enslaved and imported eleven times as many Africans as their North American counterparts. In the late nineteenth and early twentieth centuries, both countries also received millions of immigrants from Europe as they sought to industrialize. Since then, the light-skinned descendants in the United States and Brazil have come to dominate their darker-skinned compatriots through discriminatory practices that derive from a racial ideology, creating what sociologists call racially stratified societies. Both societies have experimented with affirmative-action policies to promote blacks and members of other disadvantaged groups, beginning in the 1960s in the United States and only recently in Brazil. However, the major similarities between these two large multiracial countries regarding race may end there. For one, the vast majority of persons in the United States with any African origin are categorized as black. In Brazil, large numbers of persons who are classified and identify themselves as white (branco) have African ancestors, not to mention the brown (pardo, moreno), mixed race (mestiço, mulato), and black (preto, negro) populations. Unlike in the United States, race in Brazil refers mostly to skin color or physical appearance rather than to ancestry. This difference, and many others regarding race matters, between the two countries derives from two distinct ideologies and systems of modern-day race relations. Although both racial systems are rooted in the ideology of white supremacy, their respective racial ideologies and patterns of race relations evolved in radically different ways as they responded to distinct historical, political, and cultural forces.

W.E.B. Du Bois arguably set the stage for the study of race relations in the first decade of the twentieth century when he declared the color line as the problem of the century. However, that assertion was clearly based on the bifurcated U.S. model, where blacks and whites were understood to be clearly separate groups. Had Du Bois witnessed the Brazilian case, he may have perceived that racism and discrimination were important social problems there, but he is unlikely to have identified the color line as the central problem. Also, Du Bois noted that blacks were exceptionally excluded from North American democracy; but for most of the twentieth century, there was no democracy in Brazil. Most of the population, including many whites, was excluded from access to even basic rights and subject to authoritarian domination.

Since Du Bois, the relation of blacks and whites in the United States has continued to serve as the paradigmatic case for the sociological understanding of race. Theories derived from the U.S. case are often then illegitimately applied to interpret other cases. In particular, mechanisms affecting race relations in the United States are often assumed to exist in other places like Brazil. But that is clearly not the case, as I will demonstrate in this book. Race is an important organizing principle in both Brazil and the United States but in very different ways. In the interest of building a universal sociology of race relations, I hope that this study will encourage a reexamination of sociologists' common conceptions of race relations, which too easily get translated into general knowledge despite their narrow empirical base.

In the last several decades, race relations have become a central area of sociological study which has uncovered a considerable body of evidence for understanding them. However, comparable evidence for Brazil continues to be relatively weak, largely because the small Brazilian social-science community considered the subject unimportant for that country. While a history of blatant and legal racism has undoubtedly contributed to making race an important area of study in the United States, racism in Brazil has generally been more subtle, and legal racial segregation has not existed since slavery. Indeed, the dominant assumption from Du Bois' time until recent years has been that race does not really matter in Brazil.

Such differences and similarities about race in the two countries have become common knowledge, but analysts are less certain of how other features of the two race systems compare. For example, analysts often note the existence of racial inequalities in Brazil as in the United States, but these are too easily explained as simply a product of racist practices that exist despite the absence of formal segregation. On the surface, that may be true, but there is much more to it than that. While it is becoming increasingly clear that racism is a universal phenomenon, it is less accepted that its manifestations may vary widely. Are the nature and levels of racial inequalities the same? Surely, history, politics, class structure, culture, and ideology are very distinct between the Brazil and the United States. Should these not have also affected the development of a distinct system of race relations?

Most notably, racial ideologies between the two countries contrast sharply. How did such distinct ideologies come about? Do they affect the social manifestations of race relations or merely their interpretations? A special problem in comparisons of race in Brazil and the United States has been the disentangling of ideology from social analysis. To what extent is research on race simply a reflection of the ideology? Do ideologies not have elements of truth? How much do they distort reality? Ideology also affects interpretations of sociological analysis. In other words, how do analysts present comparisons in ways that are compelling and make sense to both Brazilian and North American readers?

North American sociology has developed evidence-based theories for explaining the persistence of racism and racial inequality despite the end of formal segregation. For example, a key sociological text argues that racial residential segregation, which continues today in practice despite civil-rights reforms, forms the major basis for contemporary black disadvantage and other dimensions of race relations in the United States.2 It posits that the physical and social distance between blacks and whites, along with the strong social norms that maintain that distance, accounts for high levels of racial inequality. Conversely, it hypothesizes that without extreme segregation, racial discrimination and inequality will subside. Segregation is thus thought to be the linchpin of U.S. racial domination.3 The same may not be true for Brazil, at least if we are to believe its racial ideology. According to that ideology and to most of the research on the subject, residential segregation in Brazil is believed to be simply class based, and race is simply not an independent factor.


Segregation between blacks and whites is a well-known fact in the United States. Segregation was long formalized through the legal and policy apparatus, and as many scholars have pointed out, urban residential segregation continues to demarcate rigid boundaries between blacks and whites. At least prior to the civil-rights reforms, segregation was the dominant ideology behind race relations. Whites dealt with blacks largely by maintaining considerable social distance from them, whether through avoidance in residence, marriage, friendships, or elsewhere. Just as importantly, the practice and ideology of racial segregation came to be known in Latin America as a defining feature of North American culture. Latin Americans--especially Brazilians--thought that their culture made them morally superior, at least regarding issues of race.

Rather than segregation, race mixture or miscegenation (in Portuguese, mestiçagem or miscegenação) forms the foundational concept of Brazilian racial ideology. Race mixture represents a set of beliefs that Brazilians hold about race, including the belief that Brazilians have long mixed across racial lines, more so than in any other society, and that nonwhites are included in the Brazilian nation. Miscegenation has long been a defining metaphor of the Brazilian nation, although it initially provoked anxiety and fear among the elite, as in the United States. Although race mixture may not necessarily reflect the reality of Brazilian social behavior, the concept has been fundamental for understanding Brazil's race relations, on Brazilian terms. As Da Matta (1991) claims, understanding Brazil requires U.S. or other non-Brazilian readers to dismiss notions that Western societies are generally guided by ideas of purity. For him, Brazilians celebrate ambiguity, whereas North Americans seek to define clearly. In this sense, miscegenation represents the former and segregation the latter.

Like Brazil, many other Latin American countries hold dearly to their ideologies of mestizaje, the Spanish equivalent of race mixture. Those nations have melded racial differences into a single homogenous entity, creating an improved hybrid race of Mexicans, Dominicans, Venezuelans, and so on. However, accounts of Latin American race mixture tend to be romanticized versions that often became widely accepted as state-sanctioned visions of nationality or peoplehood in Latin America. Latin American elites have long prescribed their form of mestizaje as the formula for a positive system of human relations, free of the racial cleavages found in North American society. Even well-known Latin American scholars have been known to proclaim the virtues of presumed miscegenation in the region. In the United States, Latino scholars have also prided themselves on their racial mestizaje, as if their own histories provide a positive example for U.S. race relations. However, these supporters of mestizaje often fail to note that throughout Latin America it was built on white supremacist ideologies and has been unable to prevent the racial injustices that are increasingly uncovered throughout Latin America. Today, many sociologists have come to a consensus that race mixture represents little more than metaphor.

Brazilian academics and journalists have increasingly used the term "exclusion" to refer to the status of blacks and poor persons in their society. Exclusion is a well-known term in Latin America, with origins in Europe where it is also widely used. Exclusion, or social exclusion, refers to the "lack of social integration which is manifested in rules constraining the access of particular groups or persons to resources or limiting their access to citizenship rights."4 Social exclusion is thought to be particularly appropriate for describing Brazilian society because one-third of all Brazilians live in poverty, and most are not white.

The exclusion of blacks has thus become an important counterideology to the positive interpretation given to race mixture. Like Brazil's black social movement, which has long promoted the counterideology, a new generation of scholars largely holds that racism is pervasive throughout Brazilian society. Like the race-mixture ideology, that counterideology is dangerous to social analysis because it may also blind analysts to reality. Some have wholly accepted the counterideology and go as far as to say that segregation is similar to that in the United States in practice, despite the lack of any postslavery history of its formal manifestation. However, rejecting the ideology hypothesis does not require us to accept the counterideology. Accepting ideology or counterideology is especially tempting where the evidence about race relations is weak.

Like ideologies and counterideologies generally, racial ideologies often reduce our understanding of race relations to simple unidimensional assumptions. According to ideology, at least, exclusion is the antithesis of miscegenation. Rather, miscegenation in Brazil connotes racial inclusion, not exclusion. Latin American concepts of race mixture hold that blacks, Indians, and whites socialize, reside together, and biologically mix to the point that racial distinctions become unimportant. But is there any truth to this? If so, how can there be both exclusion and miscegenation? Exclusion and inclusion refer to extreme points on a continuum of bad versus good societies; in the case of race, bad versus good race relations. But it is common to hear Brazilians speak of their country as being the world's most miscegenated country and the world's most unequal country, in the same breath. Does that imply that there has been so much mixture that only class is important, whereas race no longer makes a difference? Or does it mean that Brazilian society is racist and stratifies by race, and miscegenation is merely ideological or characteristic of an earlier historical period? What about those white Brazilians who claim to find blacks and mulattos in their family albums? How common is this? Are such ancestors merely historical remnants? Or are such findings overstated to project a culturally desirable pedigree of miscegenation?

Contemporary analysts of Brazilian race relations seem to have discarded the possibility that race mixture and racial exclusion can coexist. If white Brazilians are so racist, then why would they mix with non-whites? Scholars argue that racial inequality and racism are so ubiquitous that they pervade all dimensions of Brazilian life. Miscegenation, some argue, occurred only among social unequals during slavery, and today occurs only for the sexual pleasure of whites but not in serious relationships. But what of all that common wisdom that miscegenation is widespread? Does it have no basis in fact? What of the earlier academic literature based on careful fieldwork which argued that Brazilian society was clearly more inclusive than the United States? Were those scholars completely wrong? Or did any inclusiveness that existed then disappear? Why would they make such an argument? Is there any evidence to support the existence of social inclusion for nonwhites anywhere, or were those scholars merely overtaken by the powerful ideologies of race mixture?

Today's social analysts have arrived at surprisingly distinct conclusions about Brazilian race relations compared to those of an earlier generation. Current scholars emphasize exclusion; past scholars emphasized race mixture. These two generations of scholars accepted either racial exclusion or inclusion as truth while ignoring or discrediting the other. Rather than considering the possibility that both racial inclusion and exclusion may coexist, the current generation of scholars has treated that possibility as the confusion of reality with popular beliefs. Those who have argued that Brazilian society is more racially inclusive and characterized by race mixture or hybridity, have also theorized that racial inequalities and discrimination are leftover from slavery but are transitory. On the other hand, the current consensus defends the exclusivity argument and asserts that racial inclusivity, or miscegenation, is merely a popular belief that is not supported by reality.

The evidence used by the current generation is based largely on official statistics that have demonstrated high levels of racial inequality. Furthermore, these academics have marshaled plenty of evidence of discrimination to support their view. But have current scholars examined race relations widely enough and asked all the right questions? Has all the proper evidence been brought to bear? For an ideology of inclusion to be so pervasively accepted for so many years would seem to require some evidence, however limited, of its existence. What is it about the Brazilian system that supported arguments about racial inclusivity? And if there is any support for them, how can inclusiveness coexist with exclusiveness? For me, this remains the enigma of Brazilian race relations.


A common categorization in the history of thought about Brazilian race relations maintains that there have been three main stages of thinking about Brazilian race relations. Roughly speaking, these three respective currents claimed that (1) there is little or no racial discrimination but rather great fluidity among races; (2) racial discrimination is widespread but transitory; and (3) racial discrimination is persistent and structural.5 While most authors are easily categorized into one of these three schools of thought, others present a mixture of these ideas or have changed their views over the course of their careers. Also, the chronological order of important contributions is not always linear but often the outcome of multiple academic debates, partially determined by the nationality of the scholars. For the purposes of this book, I generally accept this division but collapse the latter two stages into one. Thus, I characterize scholarly perspectives on Brazilian race relations as comprising two generations. The first defended racial democracy, in which Brazil is uniquely inclusive of blacks; the second challenged racial democracy, arguing that Brazil is characterized by racial exclusion. According to the first school of thought, there is little or no racism in Brazil; for the second, racism is pervasive.

The idea of miscegenation as a positive aspect of Brazilian race relations was fully developed by Gilberto Freyre in the 1930s, and some form of this perspective was defended by North American Brazilianists, including Donald Pierson, Marvin Harris, Charles Wagley, and Carl Degler, until the 1960s, and in the case of Degler, as late as 1972. Freyre and his followers believed that any existing racial inequality was an artifact both of the enslavement of blacks and their adherence to traditional cultural values, but they predicted that it would soon disappear. For them, racial differences were fluid and conditioned by class, and racial discrimination was mild and largely irrelevant. Specifically, Harris (1952) and Wagley (1952) concluded that class, rather than racial, discrimination underlies Brazil's hierarchical social relations, even though racial prejudices and stereotypes were often voiced. In general, these scholars agreed with Freyre that "being" Brazilian implied a metaracial character, which muddled racial distinctions through extensive miscegenation.

This view would be radically challenged in the late 1950s, when Brazilian sociologists, led by Florestan Fernandes, would conclude that racial democracy was a myth. Funded by the United Nations Educational, Scientific, and Cultural Organization (UNESCO) to document, understand, and disseminate Brazil's presumed secret of racial harmony in a world then marked by the horrors of racism and genocide, Fernandes would surprise his sponsors by sharply disagreeing with Freyre and his North American counterparts on the UNESCO project. Fernandes concluded that racism was widespread in Brazilian society, although he blamed slavery and its social and psychological effects on blacks themselves for their inability to compete with whites in the newly industrializing labor market. Moreover, he believed that even though racial prejudice and discrimination were functional to slave society, they were incompatible with the competitive order established by a capitalist class society. As a result, he predicted racism would disappear with capitalist development, although whites would seek to maintain their privileged positions for as long as possible.

I attribute most of the disagreement between the two generations of race-relations scholars to their separate research emphases. The first generation focused on sociability and social relations, mostly among class equals, while the second generation emphasized inequality and discrimination. I refer to these two dimensions as vertical and horizontal social relations, respectively. The North American scholars in particular tended to follow Gilberto Freyre's emphasis on horizontal relations. Many of these scholars, including Wagley and Harris, were anthropologists, but they also included sociologists like Donald Pierson (1942). For Pierson, a student of the emerging Chicago school of sociology, segregation and intermarriage were believed to be appropriate indicators of adaptability or integration of minority groups in society. This was the dominant tradition in North American sociology. They assumed that integration would lead to eventual assimilation, where the dominant and previously subordinate racial groups would become similar and racial differences would disappear. Researchers of this school found horizontal relations to be harmonious and integrative compared to the United States, thus predicting an optimistic future for the descendants of Brazilian slaves. These scholars found any racial hierarchy, conflict, or exploitation in Brazil to be unproblematic or transitory.

By contrast, Fernandes and his Brazilian UNESCO contemporaries focused mostly on the vertical relations of racial inequality. To the limited extent that Fernandes and his followers mention horizontal relations in their work, they emphasized the distance between whites and blacks. The subjects of miscegenation and the mulatto, which were of major interest in the earlier literature, are generally ignored by Fernandes. Guimarães (1999) claims that, like other Paulistas (residents of São Paulo state), Fernandes never saw miscegenation as a value but rather was motivated by social equity and development concerns. Southern Brazilians held the notion that true Brazilians were mostly white and valued being part of a European, rather than a mestizo, nation. As the primary destination of mass European immigration, São Paulo had become an ethnic mosaic and blacks were a stigmatized minority. Moreover, a disdain for miscegenation and for mixed racial categories may have come from Fernandes's close association with the black movement. Like Abdias do Nascimento, a black activist writer, Fernandes associated miscegenation with a whitening campaign to eliminate blacks from the Brazilian population. Thus, Fernandes not only ignored any practice of miscegenation, but he rejected it as simply an ideology for legitimizing racial discrimination.

Carl Degler's explicit comparison of race relations in Brazil and the United States was especially influential in the latter. As the winner of a Pulitzer Prize, Degler's work influenced North American understandings of race in Brazil more than any other source. Writing in 1972, during the more optimistic days of the U.S. civil-rights movement, Carl Degler (1986) mixed first- and second-generation interpretations in his book. Although Freyre and his followers were already in disrepute, Degler, a self-described follower of Harris, concluded that race made little difference for mulattos but generally agreed with Fernandes that the smaller number of Brazilian blacks suffered the burdens of racism. Thus, Degler claimed that the primary difference regarding race relations between the two countries was the existence of a "mulatto escape hatch" in Brazil, which allowed mulattos to overcome racial disadvantage by avoiding classification as black but also weakened the possibility of black solidarity.

After a fifteen-year hiatus in Brazilian race research due to repression by the military government, the study of race resurfaced with the completion of Carlos Hasenbalg's 1978 Ph.D. dissertation at the University of California, Berkeley. Unlike Fernandes but like the thinking emerging within Berkeley's department of sociology, Hasenbalg concluded that racism was compatible, not incompatible, with the development of Brazilian capitalism. Hasenbalg believed that racial domination and the subordinate status of blacks would persist because racism had acquired new meanings since abolition and would continue to serve the material and symbolic interests of dominant whites through the disqualification of nonwhites as competitors. By relying extensively on government statistical data, Hasenbalg and Nelson do Valle Silva produced a series of studies about racial inequality in income, education, occupation, and infant mortality throughout the 1980s. With their studies, there remained little academic doubt about the existence of racial inequality and discrimination in Brazil. Interestingly, Silva also produced several statistical studies on intermarriage and racial classification. Some of Silva's findings seemed to support the first generation's findings, but he downplayed any such support as his body of work was neither comparative nor integrated into their general theoretical conclusions. Rather, Hasenbalg and Silva's work is remembered for greatly strengthening the dominance of the second-generation perspective.6 Unfortunately for North Americans, little of this second generation of work would become available in English, and thus Degler's 1972 book would continue to be the standard reference in the United States for nonspecialists in their understanding of race in Brazil.

In sum, the stages of Brazilian racial ideas were not discrete but overlapping, with elements of previous stages sustained in subsequent stages. Racial democracy had been seriously challenged beginning in the 1950s by São Paulo academics, while some form of the racial-democracy concept continued to be defended by U.S. academia well into the 1970s. Popular and elite support for the idea of racial democracy ended in the 1990s. Although some of its elements continue to survive in current discourse and indeed are held by some members of society. Much of the old popular discourses become grafted into the new. Surely, many of these discontinuities can be explained by differences in academic contexts, ideology, and political interests in maintaining or challenging the racial-democracy discourse. Differences between U.S. and Brazilian scholarship could also be explained by distinct conceptions of what constitutes racism and discrimination in the two countries. Additionally, a significant language barrier and limited translation prevented satisfactory mutual appreciation and awareness of a growing literature by academics in both countries.



This issue of race in Brazil has recently gained a prominent place in the work of internationally eminent sociologists Herbert Gans and Pierre Bourdieu. Interestingly, their respective interpretations of Brazil are nearly contradictory. While this may largely be due to their misunderstanding about a reasonable range of the literature, it is also unfortunately due to the literature itself, in which serious analysts of Brazil have reached opposite conclusions. Despite very limited comparative evidence, Gans (1999) confidently argues that sociological outcomes between the United States and Brazil are similar:

Brazil has not passed civil rights legislation; racial stratification, discrimination and segregation have persisted but only through the class system; a high rate of illiteracy has enabled whites to virtually monopolize the higher class; intermarriage has taken place mainly among blacks and black-white "biracials"; "biracials" gain little socioeconomic advantage from their lighter skins; and the darkest-skinned blacks are forced into slums and prisons as in the United States. (377)

On the other hand, French sociologist Pierre Bourdieu and his U.S.-based colleague Loïc Wacquant (1999) argue that, unfortunately, analysts have merely transposed U.S. conceptions of race onto Brazil, despite the empirical realities:

Carried out by Americans and Latin Americans trained in the USA, most of the recent research on racial inequality in Brazil strives to prove that, contrary to the image that Brazilians have of their own nation, the country of the "three sad races" . . . is no less racist than others. (44)

They go on to discredit an influential book by a North American scholar for misunderstanding Brazilian race relations and blame U.S. foundations for exporting the U.S. model.

Although Gans and Bourdieu reach distinct conclusions, both are guided by the question of whether Brazilian race relations are better or worse than black-white relations in the United States. While they both agree that U.S. black-white relations are deplorable, Gans believes that things are no better in Brazil, while Bourdieu and Wacquant strongly imply that Brazil is less racist. Their contrasting results do not derive from a careful or systematic understanding of the Brazilian racial system. Indeed, there seems to be little familiarity with Brazil, as far as I can tell. What seems clear is that both sets of authors, in the absence of a clear understanding, project their own alternately pessimistic or optimistic image of Brazil onto their sociological analysis. Unfortunately, the literature on race relations in Brazil allows them to have their choice.

Although we might expect more from such notable sociologists, perhaps one cannot blame Gans and Bourdieu for errors of fact because neither is a student of Brazilian society. One might wonder why they took a stab at trying to interpret race in Brazil. Apparently, they both recognized that Brazil was an important case for understanding race. For whatever reason, it is interesting that these authors based their poorly founded conclusions on particular stages in the debate on Brazilian race relations. Judging from their citations, Bourdieu and Wacquant rely on the early generation of scholarly work that largely defended racial democracy. They use this work as their gold standard from which to judge all subsequent literature. Gans, by contrast, bases his observations on more recent work that reduces racial democracy to being merely ideology or myth and claims that racism is widespread in Brazil. Thus, Bourdieu and Wacquant apparently find the first stage of research more compelling and choose to use it to discredit the second stage, misreading (or failing to read) key texts and imputing false conclusions.7 By contrast, Gans selectively uses the second stage of research while ignoring the first stage.


My goal in this book is to reexamine the arguments presented by both generations of scholars. To this end, I examine a wider range of interracial relations and behaviors in Brazil in comparative and historical context. I thus attempt to provide a more integrated and wide-ranging account of Brazilian race relations than has been presented in the past and try to flesh out the contradicting interpretations of two generations of scholars. By taking advantage of Brazil's excellent data on race and a host of socioeconomic measures, I employ a set of well-developed social science methodologies and indicators of interracial behaviors to examine a broad range of race-relations issues. Unlike my predecessors, I approach these issues by acknowledging the possibility that some form of both inclusion and exclusion, however limited, may exist.

The idea of racial exclusion reflects a consensus position in the contemporary analysis of race relations in Brazil. That consensus holds that Brazilian race relations are not much different in practice than in the United States and South Africa, even during formal segregation or apartheid. Based on the limited available evidence, I find this hard to believe. My interest in this book is thus to describe race relations at various levels by making systematic international comparisons, particularly on those dimensions that are measurable and address key sociological issues regarding the importance of race and the salience of racial boundaries. Although the terms "miscegenation" and "exclusion" may be well understood as common sense in the Brazilian context, they are often vague notions that are therefore of limited analytical value. On the other hand, they capture the central tensions in Brazilian race thinking that I seek to confront in this book.

Analytically, I find that the concept of exclusion is inadequate because it expresses a dichotomy in which persons or categories of people are either entirely in or out. Such a perspective would seem to preclude the possibility of inclusion coexisting with any exclusion. I do not believe race relations are necessarily unidimensional in this sense. Also, the reference for the term "exclusion" is unclear: included or excluded from what? Similar terms such as "marginalization" or "informalization" are problematic for similar reasons. Although they have the advantage of linking poverty and inequality to the social processes of development, I prefer the more neutral concept of "vertical relations" to capture the dimension of economic exclusion. Miscegenation suggests little or no social distance among persons of different color, although it connotes a different understanding outside of the Latin American context. Similarly, segregation is used analytically in U.S. sociology to refer to great social distance, but for Latin Americans it also connotes an abhorrent and explicit system of racial division and separation. Therefore, I use the more neutral concept of "horizontal race relations" to refer to miscegenation, or more precisely, levels of sociability, which can then be used to analyze cross-national differences.

Vertical race relations are often viewed as cause or consequence of the nature of horizontal relationships. Many U.S.-based sociological theories assume that as long as social distance remains high, particularly in intermarriage and residential segregation, prejudice and discrimination will persist.8 However, the extent and nature of discrimination at each of these levels, while not fully independent of the other indicators, may have separate logics. One cannot, for example, assume that discrimination at one level implies equal discrimination at all levels. Even if this is the case for black-white relations in the United States, it is not necessarily the case for other instances of race relations, in the United States or elsewhere. While racial injustices of any kind on any dimension are morally wrong, it is insufficient to say that Brazilian society is racist; that kind of thinking might make for good activism at some level, but it makes for poor sociology at any level. The intensity and manifestations of racism and the interrelationships between different dimensions need to be understood in order to accurately understand Brazilian reality.

I believe we can better understand the Brazilian system by separating out horizontal and vertical dimensions of race relations. This strategy permits locating and distinguishing those points in which Brazilian race relations may be more or less exclusive than previously believed or in comparison to the United States. This distinction allows us a reexamination of hypotheses about horizontal relations by the early generation of research with current data and more sophisticated methodologies. We can also use more current empirical study and theory to inform our analysis of vertical relations. To present a fair and accurate depiction of race relations requires strong, empirically based indicators that can address race relations on both the horizontal and vertical dimensions. The intensity of racism at both these levels can be further understood when we compare Brazil's indicators on both dimensions to those of a society that stands out for its troubled race relations and for being the dominant model for the study of race.


In addition to trying to make sense of the internal logic of the Brazilian race system, another main reason for writing this book is to compare Brazil with the United States (and, to a lesser extent, South Africa). As the articles by Gans (1999) and Bourdieu and Wacquant (1999) reveal, systematic comparisons of race in Brazil and the United States are seriously in need. In these countries, race has been important throughout the past five centuries or more, from the time Europeans first set foot on American (or African) soil. At the very least, the important case of Brazil needs to be understood to develop a global theory of race relations. Brazil and the United States are the two largest countries in the Western Hemisphere, both in terms of total population and African-origin population. Brazilians claiming to be black or mixed race number about 80 million, constituting nearly half of Brazil's 173 million people. A large percentage of whites in Brazil also have African ancestors, raising the number of African-origin persons to perhaps over 100 million. This compares to about 30 million blacks in the United States, or about 12 percent of a total population of approximately 270 million.9

Comparisons to the United States have often served as a backdrop for understanding Brazilian race relations, as the work of many North Americans and Brazilians alike reveals. Historians of slavery have long sought to explain differences between the United States and Brazil since Freyre (1933) and Tannenbaum (1947). Explicit and implicit comparisons to the United States are prevalent throughout the literature on Brazilian race relations, probably due to the former's economic, political, and cultural hegemony. Also, many analysts of Brazilian race relations have been North Americans (e.g., Donald Pierson, Charles Wagley, Marvin Harris, Carl Degler, George Reid Andrews, and Michael Hanchard) or Brazilians that studied in the United States (e.g., Gilberto Freyre, Nelson do Valle Silva, and Antonio Sergio Guimarães).10

The Brazil-U.S. comparison in this book relies mostly on quantitative indicators, which have been used in abundance to understand U.S. race relations. Despite strong ethnographic and historical evidence, comparisons of Brazil and the United States using quantitative indicators are rare. Also, the substantive reach of both qualitative and quantitative research has been confined to mere parts of the entire racial system. Anecdotal evidence has often been used to fill in gaps where strong evidence is lacking, leading to many misconceptions and myths in the comparisons. Carl Degler's Neither Black Nor White is a good example. Although it may provide the best comparative account even thirty years after its publication, it was unfortunately sustained on weak and often anecdotal data, not to mention that it is now greatly outdated, as race relations have changed markedly in both countries.11

I believe statistical indicators can be used for the study of race in any society, provided data are available and interpreted in the context of the particular case. They convey condensed information on various dimensions of race relations and, in this case, permit U.S. and Brazil comparisons with a greater degree of confidence than was previously possible. I expect that these indicators will help to either validate or invalidate previously held assumptions. While many of the findings based on such indicators may seem obvious, others may challenge strongly held truths or bring light to our sociological uncertainties.

On the subject of comparative indicators, a careful consideration of the issue of racial classification is fundamental. The ambiguity of Brazilian race data has led to some questioning of its reliability for capturing "real" racial differences. However, previous research that relies on large data sets has used them uncritically. In contrast, I question the reliability of race data in a context where race is thought to be ambiguous and subject to social factors. Where reliability is most questionable and where data permit, I examine race relations outcomes using two recent data sets that classify race according to both interviewer- and self-classification. Although inequality between whites and nonwhites may be so great that ambiguity is unlikely to account for the racial gap, brown-black differences may be less so and thus require more careful examination. Because racial classification cannot be taken for granted in Brazil, I dedicate an entire chapter to this issue and emphasize matters of classification where appropriate in subsequent chapters.

Ultimately, I seek to reexamine the adequacy of race-relations theories. To what extent can sociological theories account for race relations in Brazil? How can an understanding of the Brazilian case help to build better race-relations theories? What does the U.S.-Brazil comparison say about the construction, maintenance, and manifestations of racial boundaries in contemporary society? To what extent, where, and why do societies as different as Brazil and the United States set racial boundaries?

As the focus of this book is clearly the Brazilian case, my comparisons to the United States are not systematic, but instead are brought in at key moments to highlight contrasts between the two countries. Due to the fact that the U.S. literature on race is large, often hotly contested, and enters into many debates, I decided to limit the interpretation of that case to dimensions where fairly objective indicators can be found and to those areas in which there is considerable consensus. I hope that the comparisons in this study using basic sociological indicators for both the United States and Brazil will overcome misinformation and stereotypes of race relations in the United States for Brazilian readers, just as I hope it will overcome the same assumptions about Brazil for North American readers.

The dynamics of race relations in the United States are far from universal and, in many ways, they may be an exception to the more common cases of racism without racist laws. Rarely have states enforced segregation laws as strict as those in the United States (and South Africa), although many more societies--including about twenty Latin American countries, including Brazil--have had little or no formal segregation, while racializing large segments of their populations. On the other hand, for readers whose principal interest is in the U.S. case, Brazil may provide some valuable lessons about the newest phase of U.S. race relations, which has been referred to as laissez-fare racism, postracism, or discrimination with a smile for its absence of legal racism and general acceptance of antiracism.


A final reason for reexamining Brazilian race relations is to discuss them in the context of the sudden and dramatic changes in Brazilian race thinking. This new phase is reflected mostly in the new acknowledgment of racism and government attempts to redress it. The issue of race in Brazil has moved to the center of the social-policy agenda. As a result, public interest in race has skyrocketed. For the first time in Brazilian history, social policies have begun to explicitly promote social integration of blacks and mulattos. Such policies do not merely seek to eliminate or alleviate material poverty but also strive to eliminate or reduce class, racial, gender, and other discriminations that bar citizens from access to social justice. This includes both universal policies that encompass the entire population or the poor population, as well as particularistic policies that combat discrimination and promote categories of people that have been excluded on the basis of particular characteristics, including race. The designs of these policies vary widely, but together they seek to address a broad range of social exclusions that are manifested economically, psychologically, politically, and culturally. This change is a milestone in Brazilian racial thought, much like Brazil's earlier ideological transition from white supremacy to racial democracy.

Indeed, the idea of affirmative action or policies specifically designed for blacks and mulattos sounds quite odd and out of place in the Brazil-ian context. In fact, the whole idea sounded preposterous and highly unlikely just a few years ago. Brazil had been one of the first multiracial states to go beyond race, but it had become apparent that its racial democracy continued to privilege whites at the expense of nonwhites, just as it did during most of its history of white supremacy. Now that these policies are actually being implemented, Brazilian policymakers are accused of imposing U.S. policies. Why would Brazil want such policies? Opponents claim that the Brazilian context is different from the United States and such policies would be of limited effectiveness. But does Brazil have an alternative to U.S.-style raceconscious policies? As the Brazilian state begins to use race explicitly to promote blacks for the first time in its history, what consequences can be expected?

These recent changes have engendered a backlash of scholarly thinking on race in Brazil. Although largely schematic and anecdotal, it has had much influence on the policy debate, mostly because it has been advocated by several well-known senior Brazilian scholars. They argue that rather than dismiss racial democracy as mere myth, it should be used to fight against racism. Myths are not mere falsities to be discovered and discarded but rather represent a popular way of thinking, which makes Brazil unique or at least different from the United States. Prompted by the federal government's plans to implement affirmative action, they believe that the belief in racial democracy provides an ideal of racial egalitarianism, which will help Brazilians to overcome racism. Furthermore, they argue that U.S.-style affirmative action will produce negative consequences for Brazilian society, making its race relations more like those of the United States.

Finally, although recent decisions to implement affirmative action may represent the most explicit intervention ever by the Brazilian state on issues of race, I also seek to show how the Brazilian state has been very actively involved in shaping race relations throughout its history. This has included the explicit importation of European immigrants to whiten its population as well as the promotion of racial democracy through a series of actions by elites, including representatives of the Brazilian government.


In my quest to present an integrated and comparative analysis of Brazil as well as to provide a historical context and an analysis of policy, I organize the book into ten chapters. Chapters 2 and 3 provide a history of politics and ideology, which serves as background for understanding race in Brazil, followed by five chapters on contemporary race relations. Chapter 4 examines the complex system of Brazilian racial classification. Then I explore vertical relations, specifically, racial inequality in chapter 5 and discrimination in chapter 6. These chapters are followed by analyses of horizontal race relations of intermarriage in chapter 7 and residential segregation, a less intimate indicator of interracial sociability, in chapter 8. I then summarize the main points of the previous chapters and draw out the theoretical implications in chapter 9. Finally, I examine the implications of the Brazilian system for designing social policy in chapter 10.


The Brazilian censuses, annual national household surveys, and two attitudinal surveys provide a treasure trove of data on race, enabling me to map out the form and nature of race relations across large sectors of the population. These largely unexplored data are based on random sampling techniques, so that all sectors of Brazilian society are represented in their rightful proportions. The importance of such data cannot be underestimated. Unlike the majority of Latin American countries, Brazil has collected race data in a majority of its censuses and has been able to document racial inequalities. Most Latin American countries do not collect population information about race. As a result, they can more easily deny racial inequality, given the inability to prove its existence. Brazil sought to do the same in the 1970s, when it did not collect data on race. However, the proof of racial injustice in Brazil since then has come largely through such government data.

Data for chapters 4 through 8 rely mostly on the analysis of survey and census data, primarily supplied by the Instituto Brasileiro de Geografia e Estatistica (IBGE). These include the 1960, 1980, and 1991 censuses, as well as the national household surveys, Pesquisa Nacional por Amostra de Domicilios (PNAD), from 1976, 1981, 1986, 1996, and 1999. Particular chapters rely more on one or another data set, depending on the substantive appropriateness and availability of the data. All of these sources inform the time-series charts in chapter 5, which span the period of 1960-1999. Where possible, I use either the 1991 census or the 1996 or 1999 PNAD to represent a fairly current depiction of the situation. I use the 1991 census when I need a large number of cases for analysis. Unfortunately, microlevel data for the 2000 census were not yet available at the time of this writing.

Moreover, I examine residential segregation, intermarriage, and the cross-sectional effects of industrialization on inequality, using a special 1980 data set of urban areas, that was specially created by the IBGE. In the case of residential segregation, the IBGE does not make their census data available by census tracts, but they graciously agreed to calculate these indexes for the forty largest urban areas in 1980. At a later point, they calculated intermarriage, inequality, and other indexes for the seventy-three largest urban areas. I worked closely with the IBGE in producing these summary indicators, including examining computer programs to ensure that formulas for computing indicators were correctly applied. District-level maps of São Paulo and Rio de Janeiro were created using the 1991 census and the 1990 PNAD, respectively. Finally, I use two independently developed surveys, including a 1995 national survey and a 2000 survey for the state of Rio de Janeiro, in the chapters on racial classification and for occasional references to racial attitudes.

Chapters 3 and 10 focus on the new era of race consciousness and policies designed to revert racism and racial inequalities. They draw largely on information I gathered as the program officer in human rights for the Ford Foundation's Brazil office. There, I was fortunate to witness closely the dramatic changes occurring in Brazilian society, with ready access to influential academics, policy makers, and leaders of the black movement. My perspective of the sudden changes in Brazilian politics of race during the past decade and the black movement is thus largely an insider's view, which would be difficult to access from traditional sources such as meeting records, official documents, and interviews. The largely demographic analysis of the other chapters has important implications for the ways in which I understand the structural sources and the implications of those changes.


Regional differences are fundamental to understanding Brazilian society. In a country with a landmass larger than the continental United States and levels of development that vary from the highly industrialized São Paulo to the very poor Northeast, regional differences need to be considered before generalizing local findings to describe "race relations in Brazil." My own experience of having resided in Bahia, Rio de Janeiro, and São Paulo, as well as frequently visiting my wife's family in Rio Grande do Sul on many occasions, is that racial classification is distinct and race relations have a different feel in these different contexts. For one thing, the white proportion of the population in each of those places is roughly 20, 55, 75, and 85 percent, respectively. While the South and Southeast have been described as class societies marked by massive European immigration, industrialization, and early urbanization, the Northeast and North have been characterized by their especially great status differences and a castelike system, inherited from slavery but not transformed by industrialization or immigration. Throughout this book, I either directly examine regional differences or indirectly through its correlates of racial composition or industrialization. For theoretical reasons that I explain later, I tend to examine horizontal relations as they relate to racial composition (e.g., percent white), and for hierarchical relations, their correlation with industrialization.

Economist Edmar Bacha once described Brazil as "Belindia," comprising a small Belgium, reflecting a high level of development, and a large but poor India.12 Although he meant merely to describe regional differences in development, Bacha's statement could be interpreted as having racial implications as well. This is apparent in map 1.1, which shows the twenty-six Brazilian states coded by percentage of the population that is white and divided by levels of social development. Increasingly lighter shades on the map indicate states with higher proportions of whites. White majorities are found in the seven southernmost states while whites are a numerical minority in the other nineteen states. The bold line separating the large northern part of the country from the smaller southern half represents levels of social development, according to the human development index, as measured by the United Nations. The human development index measures health and educational development, encompassing levels of literacy, life expectancy and infant mortality. Social development coincides with racial composition in Brazil. All nineteen states north of the bold line have an index of human development that is less than or equal to 0.8, while the seven states below it are relatively highly developed. With the exception of Minas Gerais--in which whites are a bare majority (51.4 percent)--and Espiritu Santo--in which whites are a bare minority (47.9 percent)--all majority white states are highly developed, while the majority nonwhite states rank low on human development. Thus, whites are privileged by their location in the South and Southeast, while blacks and browns tend to reside in the less developed regions of Brazil.

The differences between the first and second generations of research were also regional. The classic studies of Brazilian race relations focused almost entirely on the northern half of Brazil, which were too often generalized to all of Brazil and may have partly accounted for differing conclusions about race in Brazil. For example, the importance given to miscegenation seems to be differentiated by region in the academic interpretation of race in Brazil. In the 1930s, Freyre (1937, 1986) reduced Brazilian society to the patriarchal family of the rural Northeast which he describes as the cradle of Brazilian civilization where miscegenation found its greatest expression. In the 1950s and 1960s, North American researchers of race in Brazil also focused on the northern regions and, like Freyre, noted high rates of racial fluidity there, especially when compared to their native United States. By contrast, their Brazilian contemporaries, such as Florestan Fernandes, focused on the white southern regions and emphasized racial discrimination and inequality, generally neglecting the issue of miscegenation.


Because race is a controversial and sensitive topic, I prefer to define the concept early on. As is the consensus in sociology, race is a social construct, with very little or no biological basis. Race exists only because of racist ideologies. In the West, which includes Brazil, nineteenth-century scientific theories established that humans could be divided into distinguishable racial types, which were hierarchically ordered according to an ideology establishing that such characteristics are correlated with a person's intellectual and behavioral traits. Even though such theories are currently discredited by the vast majority of the scientific community, beliefs in the existence of races are embedded in social practices, giving the concept a great influence on social organization. By race relations, I believe that Robert Park's (2000) definition, which he wrote in the 1930s, of "relations that exist between individuals consciousof racial differences" continues to be applicable, even though he denied that race was important in Brazil and would sometimes invoke essentialistic or biological distinctions. This definition avoids the idea of race as based on a group identity that is common in the United States but is often inappropriate for Brazil.

Race is important because people continue to classify and treat others according to societally accepted ideas. The idea of race has had enormous influence in the evolution of modern societies, including Brazil's, and it has had negative consequences for its victims. I can empathize with a concern that the use of the term "race" reifies social distinctions that have no biological value, but race continues to be immensely important in sociological interaction, and therefore sociological analysis must take it into account. As sociologists have long discovered, ideas or beliefs, like those about race, can have powerful real-world consequences. W. I. Thomas's (1922) classic sociological formulation succinctly stated "if men define situations as real, they are real in their consequences."

Concepts like race vary in their connotations in different languages, as they evolved out of distinct cultural contexts. For example, color is more commonly used in Brazil, while race is more common in the United States. Choosing race instead of color is understandable in English but clumsy in Portuguese and Spanish. Nevertheless, I find race and color in Brazil to be analytically similar and derived from similar racial ideologies. I thus decide to use the term race, which underlies both concepts. I further describe my thinking on this at the beginning of chapter 4.

The choice of English translations of racial categories is more problematic. The Brazilian system uses multiple and overlapping terms, which cannot be precisely translated into English. Since I rely greatly on data that uses the census categories white, brown, and black (branco, pardo, and preto), I will usually employ these terms in this book. Unfortunately, the common terms moreno and negro also translate as brown and black, respectively, so when I refer to these terms, I often leave them in Portuguese to avoid confusion. Because much of the literature uses the term "mulatto" to refer to mixed-race persons of black and white descent, I occasionally use it as well. Certainly, the choice of one or another term may annoy some readers but almost all the terms (except perhaps white) are problematic. However, such choices are inevitable.

A special problem is choosing a term that aggregates browns and blacks. Although it is important to sometimes make the distinction between browns and blacks as this book will show, the white-nonwhite distinction is generally the most important racial cleavage between Brazil's haves and have-nots. Although the black-movement classification system recommends that the term negro include blacks and browns, I prefer to use "nonwhite" to avoid the conceptual confusion between that use of negro and its more restricted popular use. Occasionally, I use the Portuguese negro as analogous to nonwhite, especially when I refer to government, black-movement, and journalistic uses of the term, which is the term they prefer. However, it is not always clear if negro is being used to refer to only those at the dark end of the color continuum (blacks) or if it includes intermediate color categories. Examples are the movimento negro (the black movement), social policies for negros or popular attitudes about negros. This ambiguity is discussed further in chapter 4.

In seeking to respect the black movement's attempt and right to self-identify, I could also have used the term "Afro-Brazilian" or "Afro-descendant," translations of Afro-Brasileiro(a) and afrodescendente. Although these are not commonly used in the discourse of ordinary Brazilians, they are increasingly used by college-educated persons and activists in the black movement. According to some black-movement leaders, afro-descendente has gained currency because their Spanish-speaking allies can easily use it, whereas negro is extremely demeaning in some Latin American countries. Also, black-movement leaders prefer afrodescendente because it clearly identifies the descendants of enslaved Africans, which is critical in the current reparations (for slavery) movement.13

Admittedly, the term "nonwhite" should include the small and regionally concentrated Asian and indigenous population, the other two race categories in the census, but I exclude them in my analysis. This study examines the white-to-black color continuum, which encompasses the vast majority of all Brazilians. The experiences of the small and regionally concentrated Asian and Indian populations are distinct from the regionally diverse and larger white, black, and brown populations. Other ethnic groups, including Jews and Arabs, who are largely assimilated and considered white Brazilians in the Brazilian census, as well as those of Japanese and indigenous origin, are the subjects of separate scholarly treatments, to which I defer.14

Return to Book Description

File created: 8/7/2007

Questions and comments to:
Princeton University Press

New Book E-mails
New In Print
PUP Blog
Princeton APPS
Sample Chapters
Princeton Legacy Library
Exam/Desk Copy
Recent Awards
Princeton Shorts
Freshman Reading
PUP Europe
About Us
Contact Us
PUP Home

Bookmark and Share