Hostname: page-component-cd9895bd7-jn8rn Total loading time: 0 Render date: 2024-12-25T07:14:02.235Z Has data issue: false hasContentIssue false

Quase Pretos de Tão Pobres? Race and Social Discrimination in Rio de Janeiro's Twentieth-Century Criminal Courts

Published online by Cambridge University Press:  05 October 2022

Brodwyn Fischer*
Affiliation:
Northwestern University
Rights & Permissions [Opens in a new window]

Abstract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the ‘Save PDF’ action button.

Conceived as a contribution to debates about the role of state institutions in perpetuating racial inequality in modern Brazil, this article explores the relative importance of social and racial characteristics in determining defendants' treatment in Rio de Janeiro's criminal courts between 1930 and 1964. Focusing on rarely noted aspects of defendants' class and citizenship status, and emphasizing the importance of judicial procedure, it argues that social discrimination was open in Rio de Janeiro's courts, but that race alone was a relatively poor predictor of defendants' fates. At the same time, it suggests that racial and social characteristics ought not to be seen as separate and competing categories, both because “social' language had important racial meanings and because ”social“ discrimination had significant racial implications. Institutionalized social prejudice may thus go far in explaining the stubborn persistence of racial inequity in an age when ”racial democracy“ became a national hope and mantra.

Type
Research Article
Copyright
Copyright © 2004 by the University of Texas Press

Footnotes

*

I wish to thank members of the University of Chicago Latin American History workshop (especially Dain Borges and Claudio Lomnitz), participants in the Five-College Social History workshop, and the anonymous LARR referees for their valuable critiques of earlier versions of this article. I am grateful also to the Fulbright Comission, Harvard University, and the Social Science Research Council for support provided over various years of research. Thanks, too, to Jonathan Brown, Peter Ward, and all of the members of the LARR editorial team for their kindness throughout. Finally, my gratitude to Emilio Kourí, for his insightful readings, and for his loving support.

References

1. Gilberto Freyre, Casa grande e senzala (Rio de Janeiro: Mia and Schmidt, 1933); Sobrados e mucambos (Rio de Janeiro: Cia. Editora Nacional, 1936). On Brazilian intellectuals, scientific racism, and influences on Freyre's thought, see Thomas Skidmore, Black into White (Durham, N.C.: Duke University Press, 1993); Dain Borges, “Puffy, Ugly, Slothful and Inert,” Journal of Latin American Studies 25, no. 2 (May 1993): 235–37; Lilia Schwartz, O espetáculo das raças (São Paulo: Cia. das Letras, 1993); Jeffrey Needell, “History, Race and the State in the Thought of Oliveira Viana,” Hispanic American Historical Review 75, no. 1 (February 1995):1–30 and “Identity, Race, Gender and Modernity in the Origins of Gilberto Freyre's Oevre,” American Historical Review, 100, no. 1 (February 1995): 51–77.

2. Charles Wagley, Race and Class in Rural Brazil (Paris: UNESCO, 1952); L.A. Costa Pinto, O negro no Rio de Janeiro (São Paulo: Cia. Editora Nacional, 1953); Thales de Azevedo, Les elites de couleur dans une ville bresilienne, (Paris: UNESCO, 1953); Roger Bastide and Florestan Fernandes, Relações raciais entre negros e brancos em São Paulo (São Paulo: Anhembi, 1955). For insightful analysis of the UNESCO studies, see Marcos Chor Maio, “O Projeto UNESCO e a agenda das ciências sociais no Brasil dos anos 40 e 50,” in Revista Brasileira de Ciências Sociais 14, no. 41, (October 1999): 141–58; “Uma polemica esquecida: Costa Pinto, Guerreiro Ramos e o tema das relações raciais,” in Dados—Revista de Ciências Sociais 40, no. 1, (1997); “A questão racial no pensamento de Guerreiro Ramos,” in Marcos Chor Maio and Ricardo Ventura Santos, eds., Raça, ciência, e sociedade (Rio de Janeiro: Centro Cultural do Banco do Brasil, Fiocruz, 1996), 179–94.

3. George Reid Andrews, Blacks and Whites in São Paulo (Madison, Wis.: University of Wisconsin Press, 1991); Carlos Hasenbalg, Discriminação e desigualidades raciais no Brasil (Rio de Janeiro: Graal, 1979); Lilia Moritz Schwarcz and Renato da Silva Quieroz, comp., Raça e diversidade (São Paulo: Editora da Universidade de São Paulo [EdUSP], 1997); Pierre-Michel Fontaine, Race Class and Power in Brazil (Los Angeles: University of California at Los Angeles [UCLA], Center for Afro-American Studies, 1985); Peggy Lovell, org., Desigualidade racial no Brasil contemporâneo (Belo Horizonte: CEDEPLAN, FACE, Universidade Federal de Minas Gerais [UFMG], 1991); Carlos Hasenbalg and Nelson do Vale e Silva, Estrutura social, mobilidade e raça (São Paulo: Vertice; and Rio de Janeiro: Instituto Universitário de Pesquisas do Rio de Janeiro [IUPERJ], 1988) and Relações raciais no Brasil contemporâneo (Rio de Janeiro: Rio Fundo Editora, 1992); Rebecca Reichmann, ed., Race in Contemporary Brazil (University Park, Penn.: Pennsylvania State University Press, 1999); Michael Hanchard, Orpheus and Power (Princeton, N.J.: Princeton University Press, 1994); Carlos Antônio Costa Ribeiro, Côr e criminalidade (Rio de Janeiro: Ed UFRJ, 1995); Yvonne Maggie and Claudia Barcellos Rezende, Raça como retórica (Rio de Janeiro: Civilização Brasileira, 2002); Robin Sheriff, Dreaming Equality (New Brunswick, N.J.: Rutgers U. Press, 2001).

4. See, especially, Valle e Silva, Hasenbalg, Andrews, and Hanchard, op. cit.; Robin Sheriff, “Exposing Silence as Cultural Censorship: A Brazilian Case,” in American Anthropologist 102, no. 1 (March 2000): 114–32; Peter Fry, “Color and the Rule of Law in Brazil,” in Juan E. Méndez, Guillermo O'Donnell, and Paulo Sérgio Pinheiro, eds., The (Un)Rule of Law and the Underprivileged in Latin America (Notre Dame, Ind.: University of Notre Dame Press, 1999); and Chor Maio and Ventura Santos, op. cit.

5. Caetano Veloso, Haití, from Caetano Veloso and Gilberto Gil, Tropicália 2, copyright 1995.

6. This observation is especially relevant for the work of Hasenbalg and Valle e Silva, whose statistical studies often show the strong influence on socio-economic inequality of apparently non-racial factors such as regional origin and education, but whose conclusions focus mostly on race. See, for example, Valle Silva, “Cor e pobreza no centenário da abolição,” in Valle Silva and Hasenbalg, op. cit., 1992, 119–38, and Hasenbalg, “Desigualdades raciais no Brasil,” in Hasenbalg e Valle Silva, op. cit., 1988, 115–43. George Reid Andrews' work on São Paulo stands out as an important exception to this trend.

7. Records from the archives of the Primeira Vara Criminal, Maço 407, no. 4184 (formerly held in the State Judicial archive of Rio de Janeiro) and from the Arquivo Judiciário of the Rio de Janeiro Fórum, Vara de Execuções Penais, Processo no. 2242/1959.

8. Steady, legal factory employment was, in fact, quite hard to come by in mid-twentieth-century Rio and is now widely acknowledged to have been a relatively privileged attainment throughout Brazil. For rates of employment by sector in Brazil, see Brodwyn Fischer, “The Poverty of Law: Rio de Janeiro, 1930–1964,” Ph.D. diss., Department of History, Harvard University, 1999; for the relationship between factory work and other forms of employment with regards to race, see Hasenbalg, “O negro na indústria: proletarização tardia e desigual,” in Hasenbalg and Valle Silva, op. cit., 1992, 101–18; Lovell, op. cit.; and Andrews, op. cit.

9. Citation unavailable; found within the processo.

10. Jagunço, often translated as “bandit,” implies a particular kind of outlaw, one tainted and brutalized by the environmental and racial handicaps that elite Brazilians viewed as shaping Northeastern sertanejo (rural inland) populations. Euclides Da Cunha's Rebellion in the Backlands famously explores the term.

11. Servente implies a worker of humble status, not exactly a “servant,” but also not “working class.”

12. Brasil, Serviço de Estatística Demográfica, Moral e Política, Crimes e contravenções (Rio de Janeiro: Serviço Gráfico do IBGE, 1942–47; Departamento de Imprensa Nacional, 1948–64).

13. Between 1942 (when city officials began to compile criminal statistics) and 1963, on average 90 percent of all defendants in carioca criminal trials were male. Women were frequently victims of crime—especially domestic violence, violence among neighbors, sexual crimes, and thefts and robberies—but they were not especially frequent victims of physical violence and murders in public spaces and in the workplace. Information on defendants, ibid.; on victims, based on my own sample.

14. This pattern held true long before the twentieth century and far beyond Rio's borders; Lyman Johnson and Sonya Lipsett-Rivera, The Faces of Honor (Albuquerque, N.M.: University of New Mexico Press, 1998) and Sueann Caulfield, Lara Putnam, and Sarah Chambers, Honor, Status and Law in Modern Latin America, under review, Duke University Press.

15. The exceptions to this trend in my own sample are thefts and robberies, for obvious reasons.

16. On differences between positivist theories of criminology and the so-called “classical” school that had dominated the 1890 code, see Peter Fry, “Direito positivo versus direito classico: a psicologização do crime no Brasil no pensamento de Heitor Carrilho,” in Sérvulo A. Figueira, org., Cultura e psicanálise (São Paulo: Brasiliense, 1985); Fry and Sérgio Carrera, “As vicissitudes do liberalismo no direito penal brasileiro,” Revista Brasileira de Ciências Sociais 1, no. 2 (1986): 48–54; Carlos Antônio Costa Ribeiro Filho, “Clássicos e positivistas no moderno direito penal brasileiro: uma interpretação sociológica,” in Micael M. Herschmann and Carlos Alberto Messeder Pereira, A invenção do Brasil moderno: medicina, educação e engenharia nos anos 20–30 (Rio de Janeiro: Rocco, 1994). For a longer analysis of the transition from older to newer criminal and procedural codes, see Fischer, op. cit.

17. These percentages refer only to cases that actually went far enough for rights violations to be determined, 249 criminal trials including 62 murder and attempted murder trials.

18. This sort of language was the lingua franca of many sociologists, anthropologists, social workers, and policymakers throughout the 1950s, 1960s, and even 1970s. For a concise explanation, see Janice Perlman, The Myth of Marginality (Berkeley and Los Angeles: University of California Press, 1976).

19. Although some researchers argue that racial categories used by the police were highly mutable, I did not generally find this to be the case. Although an individual's racial category would occasionally shift over the course of a trial, these shifts were very rare and were almost never across the white/nonwhite divide. These words obviously reflect the social perception of an individual's color rather than an exact analysis of biological descent or “race,” but seemed to be extraordinarily consistent in the sample I considered.

20. Giorgio Mortara, “Atividade e posições na ocupação nos diversos grupos de côr da população do Distrito Federal” and “A composição da população do Distrito Federal segundo a côr,” in IBGE, Pesquisa sobre os diversos grupos de côr nas populações do Estado de São Paulo e do D.F. (Rio de Janeiro: n.p., 1951).

21. See note 2, above. On Bahia, for an interpretation that diverges somewhat from the standard view, see Antonio Sérgio Alfredo Guimarães, “Côr, classes e status nos estudos de Pierson, Azevedo e Harris na Bahia, 1940–1960,” in Chor Maio and Ventura Santos, op. cit., 143–58. For the standard interpretation of these studies' importance in the history of race relations in Brazil, see Carlos Hasenbalg, “Racial Inequalities in Brazil and Throughout Latin America: Timid Responses to Disguised Racism,” in Elizabeth Jelin and Eric Hershberg, Constructing Democracy: Human Rights, Citizenship, and Society in Latin America, (Boulder, Colo.: Westview Press, 1996).

22. The classic here is of course Florestan Fernandes' The Negro in Brazilian Society, translated by Jacqueline D. Skiles, A. Brunei, and Arthur Rothwell (New York and London: Columbia University Press, 1969).

23. See, especially, Lovell, ed., op. cit., 1991, and Valle Silva and Hasenbalg, op. cit., 1992. There is considerable disagreement as to whether these disparities are significantly less visible for pardos than they are for pretos.

24. Hasenbalg and Valle Silva, op. cit.

25. Ibid; Andrews, op. cit.; Lovell, op. cit.

26. Valle Silva, “Distância social e casamento inter-racial no Brasil,” in Valle Silva and Hasenbalg, op. cit., 1992, 17–52; Sueann Caulfield, In Defense of Honor (Chapel Hill, N.C.: Duke University Press, 2000).

27. Edward E. Telles, “Race, Class and Space in Brazilian Cities,” International Journal of Urban and Regional Research 19, no. 3 (September 1995): 394–406; Fischer, op. cit.

28. Michael Hanchard, op. cit.; France Winddance Twine, Racism in a Racial Democracy: The Maintenance of White Supremacy in Brazil (New Brunswick, N.J.: Rutgers University Press, 1997).

29. Robin Sheriff, “Exposing Silence,” op. cit.; Dreaming Equality, op. cit.

30. Martha Abreu and Sueann Caulfield have observed extensive adherence to racialized notions of female sexuality and the dangers of urban modernity in carioca legal texts, police stations, and courtrooms; Olivia Gomes da Cunha has revealed the pseudo-scientific racism that persisted at the highest levels of the carioca police and legal-medical establishments well into the 1930s; and Sérgio Carrara has indicated indirectly the extent to which race tinged the practice of criminal psychology. Martha Abreu, Meninas perdidas, Rio de Janeiro, 1989; Sueann Caulfield, op. cit.; Olivia Gomes da Cunha, Intenção e gesto: pessoa, côr e a produção cotidiana da (in)diferença no Rio de Janerio (Rio de Janeiro: Arquivo Nacional, 2002); Sérgio Carrara, Crime e loucura (Rio de Janeiro: EdUERJ; São Paulo: UdESP, 1998).

31. Racial difference had a remarkably weak overt presence in Brazilian law from its origins. Before abolition, the key legal distinction was based on slave or free status, not upon race, and in the post-emancipation period explicit legal distinctions between the races were virtually non-existent.

32. Nelson Hungria, “A Criminalidade dos homens de côr no Brasil,” Revista Forense, ano XLVIII, vol. CXXXIV (Março 1951): 21–30.

33. Hungria, p. 26, quoting Donald R. Taft, Criminology (New York: MacMillan, 1948), 91. It is clear throughout the article that Hungria is using the word culture in the same broad sense as the American criminologists he cites, and not in the narrow Portuguese sense of educational and social level.

34. For two early works disproving these assumptions in relation to favela populations at large, see Vitor Tavares de Moura, Relatório sobre o problema das favelas (Rio de Janeiro: n.p., 1940); and Maria Hortência do Nascimento Silva, Impressões de uma assistente sobre o trabalho na favela (Rio de Janeiro: Prefeitura do Distrito Federal, Gráfica Sauer, 1942). For sociological evidence against these early stereotypes, see Perlman, op. cit.; Elizabeth Leeds and Anthony Leeds, A sociologia do Brasil urbano (Rio de Janeiro: Zahar Editores, 1977); Licia do Prado Valladares, Passa-se uma casa (Rio de Janeiro: Zahar Editores, 1978)

35. Prefeitura do Distrito Federal, Secretaria Geral do Interior e Segurança, Departamento de Segurança e Estatística, Censo das favelas: aspectos gerais (Rio de Janeiro: Prefeitura do Distrito Federal, 1949).

36. IPEME, A vida mental dos favelados (Rio de Janeiro: IPEME, 1958), 31–34.

37. For a thoughtful reflection on this merger of the racial and the social in Brazil, see Guimarães, op. cit.

38. See note 6.

39. Most work on the institutional mechanisms that perpetuate Brazilian racial difference has focused on employment and earnings patterns rather than governmental practices. Our knowledge of the Brazilian state's role in this process is limited, a glaring gap in light of its enormous twentieth-century role in distributing both concrete benefits such as social security, public housing, legal land titles, welfare, vocational education, and intangible goods such as workers' and citizens' rights. Emerging exceptions include studies of immigration (see, especially, Jeffrey Lesser, Negotiating National Identity (Durham, N.C.: Duke University Press, 1999); Giralda Seyferth, “Construindo a nação: hierarquias raciais e o papel do racismo na política de imigração e colonização,” in Chor Maio and Ventura Santos, op. cit., 41–58; Skidmore, op. cit.; and Schwartz, op. cit‥ For education, see Jerry Dávila, Diploma of Whiteness (Durham, N.C.: Duke University Press, 2003); public health, Gilberto Hochman, A era do saneamento (São Paulo: HUCITEC-ANPOCS, 1998) and Nísia de Trindade Lima and Gilberto Hochman, “Condenado pela raça, absolvido pela medicina,” in Chor Maio and Ventura Santos, op. cit., 23–40.

40. See Borris Fausto, Crime e cotidiano (São Paulo: Brasiliense, 1984); Alba Zaluar, A maquina e a revolta (São Paulo: Brasiliense, 1985); Caulfield, op. cit., 2000; Abreu, op. cit.; Gomes da Cunha, op. cit.; Costa Ribeiro, op. cit.; and Sérgio Adorno, “Racial Discrimination and Criminal Justice in São Paulo,” in Reichmann, op. cit., 123–38.

41. For well-conceived exceptions, see Caulfield, op. cit., 2000, and Gomes da Cunha, op. cit., though only Caulfield couples quantitative and qualitative analysis. Roberto Kant de Lima's dissertation is an illuminating study of police procedure and practice, but he does not privilege race in his analysis (Roberto Kant de Lima, “Legal Theory and Judicial Practice: Paradoxes of Police Work in Rio de Janeiro City,” Ph.D. diss., Department of Anthropology, Harvard University, August 1986).

42. A partial exception is Joana Domingues Vargas' analysis of sexual crimes in Campinas (“Indivíduos sob suspeita: a côr dos acusados de estupro no fluxo do sistema de justiça criminal,” in Dados 42, no. 4 [1999]: 63–82). Her sample, though, was too small, and her statistical techniques too simple, to reach definite conclusions, and she did not consider procedural violations.

43. Citizenship documents included civil identification cards, worker identification cards known as carteiras de trabalho and papers certifying military service and voter registration. Carlos Antônio Costa Riberio has probably gone further than anyone in considering a number of diverse factors in analyzing judicial outcomes (“As práticas judiciais e o significado do processo de julgamento,” Dados 42, no. 4 [1999]: 691–727), but his sample is too small—and his time frame too narrow—to yield definitive results

44. The cases were chosen from records held in the Brazilian National Archive, Rio's now privatized state judicial archives, and the archives of Rio's principal sentencing court.

45. I chose to analyze diverse crimes because patterns of police and judicial behavior tend to vary widely, as do socio-economic profiles of defendants and victims, and I aimed to capture as broad spectrum as possible. The percentage of white defendants did not vary much from one sample crime to another, reducing the chances that drawing the sample from a variety of crimes distorted my findings. When there are variations, they involve larger percentages of white defendants in crimes that were less vigorously prosecuted and prone to fewer rule violations than average, meaning that if anything my cross tabulations tend to exaggerate whiteness' significance in producing lenient treatment. In the multivariate analysis, crime type was considered as an independent variable.

46. Because there are no citywide figures available for the 1930 period, we cannot know whether or not this group of defendants was representative of all. Compared to all cariocas, these defendants were less white (53 percent as opposed to 70 percent), slightly more illiterate (26 percent as opposed 22 percent), and more likely to be single (60 percent as opposed to 51 percent). Statistics are drawn from the 1940 census. In the 1950s and 1960s, citywide crime statistics are available (though not, disaggregated), and they indicate that defendants in my sample were slightly less white, somewhat more literate, and more frequently single than their counterparts in all of the sample crimes committed in Rio. Compared to the citywide population in 1960, defendants in my sample were more often Afro-Brazilian (60 percent as opposed to 30 percent), slightly more literate (82 percent as opposed to 72 percent), more frequently single or involved in a consensual unions (66 percent as opposed 46 percent), and more likely to live in favelas or other forms of informal housing (34 percent as opposed to 10 percent).

47. There are no thorough historical studies of criminality rates by class and race in Brazil, though there is evidence that poor, black Brazilians tend to be more frequent victims and perpetrators of crimes in modern Brazil. Many researchers have concluded that exaggerated police vigilance and abuse are the main reasons that blacks are over-represented in populations of defendants and prison inmates, citing as evidence U.S. studies to the effect that there are no racial differences in criminal propensity. While this is most certainly true in abstract, and while the very definition of which acts are crimes is generally infused with racism and classism, theory is no substitute for further study. See Adorno, op. cit., and Costa Ribeiro, 1995, op. cit.

48. This smaller percentage of white reus in the latter sample may reflect a citywide trend; between 1942 and 1963, the proportion of white defendants fell from 54 percent to 45 percent. See Crimes e Contravenções, op. cit.

49. The personal characteristics considered differed according to the type of analysis. For the cross tabulations of the 1930s sample, I began with eleven commonly noted characteristics: region of birth, civil status, race, educational level, occupational level, employment status, housing type (shack, rooming house, apartment, etc.), neighborhood type (formal, informal, or mixed), access to legal counsel (considered only for cases that went to trial), prior arrest record, and possession of a state identification document (already, by the 1930s, an important marker of citizenship status). For the later sample, I added two categories; possession of the carteira de trabalho, and each defendant's characterization in the police vida pregressa. For the logistic regressions, I necessarily reduced the number of factors considered, eliminating categories in order to maximize the predictive power of each model.

50. Small by statistical standards, my samples are nonetheless among the largest utilized so far in historical studies of Brazilian criminal justice.

51. Due to space constraints, the discussion of statistical results that follows is condensed. For tables and a more extensive discussion of methodology, please see my Northwestern University website.

52. See, especially, Adorno, op. cit.

53. The low social status and mixed racial composition of the Rio's police dated back at least to the late nineteenth century; See Thomas Holloway, Policing Rio de Janeiro (Stanford: Stanford University Press, 1993), and Marcos Luiz Bretas, Ordem na cidade: o exercício cotidiano da autoridade policial no Rio de Janeiro, 1907–1930 (Rio de Janeiro: Rocco, 1997).

54. In the 1930s, cases against white defendants were dismissed 10 percent more often than the mean, cases against pardos were average, and those against pretos were dismissed 6.6 percent less than the mean. After 1950, pardos and brancos were in the average range, while pretos‘ cases were dismissed 8.2 percent less than the mean. The only really surprising result here is the consistent difference between pardos and pretos, given modern researchers’ doubts about the validity of the distinction. For a discussion of this debate as related to the criminal justice system, see Costa Ribeiro, op. cit.; Adorno, op. cit.

55. Characteristics with average dismissal rates more than 10 percent different than average in the early sample were: professional occupation (26.3), formal housing type (10.1), white skin (10), previous criminal inquiries (-10.2), occupation as an unskilled laborer (-12), previous conviction (-12.5), illiteracy (-13.7), shack housing (-14.6), residence in an informal neighborhood (-19.2), and unemployment (-21.7). In the latter sample, they included occupation as a housewife (36.4), occupation in the police or military forces (11.8), formal housing type (10), occupation as a white-collar employee (-10.8), residence in an informal neighborhood (-11), weak literacy (-11.9), lack of employment (-13.8), previous conviction (-22.8), and a negative VP (-23.8).

56. Models are standard logistical regression equations, where the natural log of the odds that any individual will go to trial, be found guilty, spend time in prison, or experience civil rights violations are equal to a constant added to the “B” corresponding to each characteristic possessed by an individual. Models were constructed to maximize their overall predictive power, and predict correct outcomes between 71 and 86 percent of the time, considerably higher rates than those produced from random guesses. Results are expressed as odds; odds lower than 1 are low, and those over 1 are high. In 1930, whites' odds of going to trial were .422, residents of formal housing faced odds of .574, residents of shacks and tenements faced odds of 2.233, professionals faced odds of .162, factory workers and skilled laborers faced odds of 2.336, unskilled laborers faced odds of 2.207, and those with primary literacy faced odds of .569. All findings here are significant to .05. See my Northwestern web page for full discussion.

57. Residents of informal housing had high odds of going to trial (2.389, at .05 significance), as did white-collar workers (2.938, at .05), those with previous convictions (2.698, at .05–.15) and those with a negative VP (2.876, at .05). Unskilled laborers, those with a clean criminal record, and those with a positive VP all faced low odds of trial (.522, .517, and .136, all at .05-.15 significance). The negative results for white-collar employees may be due to a tendency among defendants to claim “employment in commerce” in order to disguise unemployment or criminal occupations. The positive results for unskilled laborers can only be explained through further research, although they—like positive results for Afro Brazilians detailed below—may be due for the system's tendency to dismiss as unimportant crimes committed among low-status individuals (see Caulfield, op. cit.). Crime type influenced odds in predictable ways in all regressions.

58. Before 1942, deviations of more than 10 percent in the percentage of innocent rulings appeared for defendants who had access to private counsel (17.7), lived in formal housing (12.9), worked as professionals (11.9), had high levels of literacy (10.2), lived in shack housing (-11.1), had no lawyer (-12.8), had previous convictions (-12.8), had a public defender (-16.5), or were unemployed (-22.1). After 1950, deviations of more than 10 percent appeared for the highly literate (46.8), professionals (44.2), housewives (37.5), policemen and soldiers (32), those with a carteira de trabalho (26.3), foreigners (25.6), those with private counsel (23.2), those with civil identification card (15.9), those with positive VPs (11.6), those with previous convictions (-10.5), Northeastern migrants (-11), those with public defenders (-11.5), those with rudimentary literacy (-16.2 percent), residents of collective housing (-16.6 percent), those with negative VPs (-16.9), unskilled laborers (-18.3), and the unemployed (-23.6).

59. Brancos‘ odds were 2.3, and pretos‘ odds were .097, though only the latter is significant to .05–.15. This is quite possibly an anomalous result, given its magnitude, but it is not unique in pointing to more lenient treatment of pardos and pretos, a phenomenon that may be related to the status of their victims.

60. Before 1942, white defendants' odds of a guilty verdict were 2.3, while pretos' were .097 (to .05–.15 sig.); those with a private lawyer faced .397 odds (.05 significance) and those with a public defender 1.749 (.05–.15). Less significant results also indicate low conviction odds for residents of formal housing (.661 at .15–.25 sig.) and with a work card (.275 at .15–.25 sig.). After 1950, the most significant results (.05 significance) indicate very low conviction odds for holders of the carteira de trabalho (.071) and very negative ones for unskilled laborers (odds of 4.55). Less significant results (.05–.15) indicate high odds for white-collar workers (3.302), and less significant numbers still (.15–.25) indicate low conviction odds for professionals (.194) and those with no criminal record, and high odds of a guilty verdict for residents of informal housing (1.719).

61. Before 1942, variations of more than 10 percent from the average percentage of defendants who spent no time in jail were found for housewives (24.9), the highly literate (17.7), police and military personnel (15.6), professionals (13.6), residents of formal housing (12.7), skilled laborers (-12.5), those with previous criminal inquiries (-19.5), those with previous convictions (-29.2) and the unemployed (-40.9). After 1950, variations occurred for housewives (30.2), professionals (23.1), police and military personnel (18.5), those with the carteira de trabalho (16), those with positive VPs (13.3), those with civil identification cards (12.1), those who were married or widowed (10.1), those with previous criminal inquiries (-10.4), white-collar workers (-10.7), skilled laborers (-11.2), the barely literate (-11.5), the unemployed (-20.8), those with a criminal conviction (-34.9) and those with a negative VP (-38.1).

62. Before 1942, results significant to .05 indicate low odds of prison time for pretos (.391) and for holders of civil identification cards (.384), but high ones for those with previous convictions (3.693). Results significant from .05–.15 indicate low odds for foreigners (.479), residents of formal housing (.651), and residents of the Zona Sul (.433) and high ones for brancos (1.909), and factory workers or skilled laborers (2.359). Results at .15–.25 significance indicate high odds for rural-urban migrants (1.528). After 1950, the only results significant to .05 give high odds of prison time to those with negative VPs (3.147) and previous convictions (2.576); those with .05–. 15 significance indicate low odds of prison time for pardos, pretos, and those with a positive VP (.4, .44, and .523), and very high odds for white-collar workers (2.204). Results at .15–.25 indicate reduced odds for the literate (.643).

63. Although police and judges rarely explicitly acknowledged procedural violations in trial documents, such violations emerged in a variety of ways. Rarely, defendants had highly competent lawyers who pointed out procedural violations and demanded that something be done about them. More frequently, trial delays could be identified simply by looking at the dates in trial records and comparing them with the regulations set out in the procedural codes, and illegally long pre-trial imprisonments could be detected in the same way. Violations involving absent or incompetent lawyers were similarly easy to discern, and defendants and lawyers alike often brought complaints about police abuse directly to judges, whom they seemed to think capable of rectifying such injustices.

64. Variations of more than 10 percent from the average percentage of cases without violations were found before 1942 for police and military personnel (21.5), professionals (19.1), the highly literate (17.3), foreigners (12.3), residents of formal housing (12.1), those with civil identification cards (11.6), white-collar workers (11.3), the solidly literate (11.1), those who lived in shacks (-11.4), those who lived in informal neighborhoods (-13.2), the illiterate (-14.7), unskilled laborers (-16.7), those with previous criminal inquiries (-17), those with previous convictions (-27.8), and the unemployed. After 1950, significant variations occurred for housewives (30.4), police and military personnel (29.8), professionals (23.2), those with the carteira de trabalho (22.7), the highly literate (21.2), those with civil identification cards (17.8), those with positive VPs (15.2), residents of formal housing (13.7), those with no criminal record (10.4), those with previous criminal investigations (–10.1), those with no civil identification card (–11.1), unskilled laborers (–12.9), those with rudimentary literacy (–13.3), residents of informal neighborhoods (–13.7), the unemployed (–33.9), those with previous convictions (–35.3), and those with negative VPs (–37.4).

65. Before 1942, results significant at .05 indicate low odds of civil rights violations for foreigners and holders of civil identification cards (.296 and .365) and high ones for unskilled laborers (3.001). Results at .05–.15 significance indicate low odds for pretos (.459) and residents of formal housing (.589) and high ones for shack and tenement residents (1.738), residents outside of the Zona Sul (2.107), and those with previous convictions (2.319). Results at .15–.25 significance indicate high odds for migrants (1.56) and skilled and factory workers (1.785), and low ones for Zona Sul residents (.414) and professionals (.161). After 1950, results significant at .05 indicate high violations odds for residents of informal housing (2.731), those with criminal records (3.072), and defendants with a negative VP (3.351); results significant from .15-.25 indicate low odds for formal housing residents (.161) and defendants with the carteira de trabalho (.379).