Society for Comparative Studies in Society and History

Democracy and Violence in Brazil

Author(s): Teresa P. R. Caldeira and James Holston
Reviewed work(s):
Source: Comparative Studies in Society and History, Vol. 41, No. 4 (Oct., 1999), pp. 691-729
Published by: Cambridge University Press
Stable URL: http://www.jstor.org/stable/179426 .
Accessed: 12/12/2011 12:09
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

Cambridge University Press and Society for Comparative Studies in Society and History are collaborating with
JSTOR to digitize, preserve and extend access to Comparative Studies in Society and History.

http://www.jstor.org

Democracy and Violence in Brazil

TERESA P. R. CALDEIRA
Universityof Californiaat Irvine

JAMES HOLSTON
Universityof Californiaat San Diego
Democracy has expanded remarkablythroughoutthe world during the last
quarterof the twentiethcentury.In 1972, therewere fifty-two electoraldemocracies, constituting33 percent of the world's 160 sovereign nation-states.By
1996, the numberhad risen to 118 electoral democraciesout of 191 states, or
62 percentof the total, for a net gain of 66 democraticstates.Among the larger countrieswith a populationof one million people or more, the numberof political democraciesnearly tripledduringthe same period. If it took two hundredyears of political change from the Age of Revolution to 1970 to generate
about fifty new democraticstates, it has taken only ten years since the mid1980s to yield the same numberagain. This movementof political democratization has swept over every region of the globe, taking root in societies with
very different cultures and histories, from Papua New Guinea to Botswana,
Brazil, and Bulgaria.In the one region where it has not transformedthe nature
of nationalrule, the Middle East, it has neverthelessgenerateda plethoraof local democraticprojectsand debates.At the end of the millennium,democracy
has indisputablybecome a global value adoptedby the most diverse societies.
Although the new democratizationis overwhelmingly non-Western,the
dominant theories and evaluations of democracy remain predicated on its
Westernexperience. They remainespecially focused on the transformationof
political systems on regime change, electoralcompetition,and theirpreconditions-that is a hallmarkof Westerndemocratization.Such political considerationsare certainlyfundamental.They establishthatmost countriesin Latin
America, for example, have indeed become democraticin the sense that they
arepolitical democracies.However, the problematicand at times perversedemocratizationsof LatinAmericademonstratejust as surely thatthe consolidation of democracyrequiressocial andculturalchanges which escape the analy'We derive these data on electoral democraciesfrom the annualworld surveys that Freedom
House has compiled systematicallysince 1972. Althoughcrucialin its own right,we are criticalof
the electoral approachin evaluatingdemocracy,as will become clear below.
0010-4175/99/4423-0446$7.50 + .10 ? 1999 Society for ComparativeStudyof Society and History

69I

692

TERESA

P. R. CALDEIRA

AND

JAMES

HOLSTON

sis of this narrowpolitical perspective.Their difficulties stronglysuggest that,

althoughnecessary,political democracyis not enough to securethe civil rights
of citizenshipor producea democraticrule of law. Withoutboth, the realization
of democraticcitizenshipremainsdisabled.It is increasinglyevidentin the new
democraciesthat without this realization,political democracyloses its legitimacy andefficacy. It suffersnot only as a meansto framesocial interactionbut
also as a mode of governance.
The fundamentalproblemis that the form, substance,and developmentof
the new non-Westerndemocraciesusually differ significantlyfrom the canonized Euro-Americanexperience. It is not only that their differenthistories require a differentapproachto understand,on the one hand, the global reach of
democracyand, on the other,its specificity in diverse culturalcircumstances.
It is also thatthese otherhistories suggest the need to revise many standardassumptionsaboutdemocratization.They demonstrate,for example, both the insufficiencyof democraticelections for realizingdemocraticcitizenship,andthe
limitations of a democratictheory based on elections for understandingthis
problem.Furthermore,these other democraciesshow that the spread,timing,
and substanceof citizenship vary significantlyin differenthistoricalcontexts.
Finally, they also reveal that the developmentof citizenship is never cumulative, linear, or evenly distributedfor all citizens, but is always a mix of progressive and regressiveelements, uneven, unbalanced,andheterogeneous-in
short,what we call disjunctive.
Above all, it is the widespreadconcurrenceof democraticpolitics and systematic violence against citizens in emerging democraciesthat reveals these
limitationsof method and theory,and that requiresa differentconceptualization. This concurrencemeans that many such new democraciesexperience a
similarand defining disjunction:althoughtheirpolitical institutionsdemocratize with considerablesuccess, and althoughthey promulgateconstitutionsand
legal codes based on the rule of law and democraticvalues, the civil component of citizenshipremainsseriously impairedas citizens suffer systematicviolations of their rights. In such uncivil political democracies,violence, injustice, andimpunityareoften the norms.As a result,uncivil democraciesundergo
the delegitimationof many institutionsof law andjustice, an escalationof both
violent crime andpolice abuse,the criminalizationof the poor, a significantincrease in supportfor illegal measuresof control, the pervasive obstructionof
the principle of legality, and an unequal and uneven distributionof citizen
rights.Narrowlypolitical definitionsof democracy-those thatignore the civil componentof citizenship and its constituentelements of justice and law in
the real lives of citizens and states-overlook these dilemmas.
By the term"civil,"we refernot to the classic liberalseparationbetweenstate
and non-state,political society and civil society, public and private,or to any
such dichotomiesthattypically derivefrom the state/non-statedivide.2Rather,
2 These dichotomiesare usefully analyzedin Bobbio 1989:1-43.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

693

we use civil to refer to an aspect of citizenship, and citizenship to refer to the

prerogativesand encumbrancesof membershipin the modernpolitical community.Developing T. H. Marshall's(1977) typology (but discardinghis progressive and cumulativehistory), we distinguishcivil componentsof citizenship fromtheirpolitical, socioeconomic, andculturalcounterparts.We use civil
to specify the sphereof rights,practices,and values thatconcernsliberty,both
negative and positive, andjustice as the means to all otherrights. With regard
to liberty,the civil componentof citizenshipnot only securesits negativemeanings in the sense of guaranteeingthe autonomyof privateindividualsagainst
the abuses of the state. It also secures liberty in the positive sense of rights to
associate,assemble,andcommunicateamongprivateindividuals,who thusbecome associated individuals and thereby create the public sphere of society.
With regard to justice, the rights, practices, and values of civil citizenship
groundthe democraticrule of law.
Throughthese principles of liberty andjustice, the civil componentrelates
and regulatesboth society and the state-the latteras one of society's legally
constitutedagents.This regulationcreatesa productivemediationbetween society and state that is ambiguous, not dichotomous: the civil sphere differentiates society from the political system by defending the former from the
abuses of the latter;however, it also integratesthe two by utilizing statepower
to confrontrelationsof inequalityand dominationwithin society itself, and to
shapepeople into certainkinds of citizen-subjects.In the lattersense, therefore,
civil democracy depends on the state's capacity to impose sanctions. Highlighting this ambiguousmediation,our use of the civil embracesa paradoxof
moderndemocracy:althoughsociety needs protectionfrom the state, it is only
within the frameworkof a statethatthis is possible. Thus, citizenshipis a complex regulatoryregime by which the statemolds people into particularkinds of
subjects,andby which citizens also hold the stateaccountableto theirinterests.
We use the notionof "civil"to emphasizethis complex imbricationof stateand
society throughcitizenship.3
One of our aims in this essay is to stress the importanceof the civil component of citizenshipin democraticprocesses and, therefore,in democratictheory. We do so by looking at what happenswhen this componentis systematically violated, not underdictatorshipbut underpolitical democracy.We focus on
the case of Brazil to emphasize both the lived consequences of such violence
against citizens under democracy,and its theoretical significance for understandingdemocraticchange. We consider a numberof consequencesin Brazil
thatderive from the fundamentaldisjunctionof political democracyandthe violation of civil citizenship, and thatconcernthe individualbody, public space,
collective rights, and the rule of law. We want to emphasize at the outset that
these consequences are found in other political democracies, including those
3 For furtherdiscussion of the civil
componentof citizenship,see Holston, forthcoming,particularlyChapter5.

694

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

thatareestablishedandnot generallyuncivil, for it is also our argumentthatall

democraciesare disjunctivein the sense suggested above, at all times and in
variousways.
Whatmakes Brazil exemplaryis thatit presentswith particularand unfortunate claritythe disjunctionof the civil componentof citizenshipthatis characteristic of many emerging democracies.WhereasBrazil's political democratization in the 1980s-after twenty years of military dictatorship-promised
individualliberty,autonomy,and security,the actualitiesof everydayviolence
against Braziliansproducecontinuingor even increasingvulnerabilityof the
citizen's body. Many Braziliansfeel less individualsecurityunderdemocracy.
WhereasBrazil's democraticconstitutionof 1988 is predicatedon the notion
of a public thatis open, transparent,and accessible, a cultureof fear and suspicion has takenhold underpolitical democracythatproducesabandonmentand
lawlessness of public spaces-their conversion into no-man's land-or their
enclosure, fortification,and privatization.WhereasBraziliandemocracypresumes the rule of law, the institutionsand practicesof law andjustice are discredited and undemocratic.Their delegitimation demonstratesthat political
democracy does not necessarily or automaticallygenerate a democraticrule
of law.
In the next section of the essay, we analyze these various violations of the
civil componentof citizenship in the context of Brazil's currentpolitical democracy.In the final section, we develop the concept of disjunctivedemocracy as a means of betterunderstandingthese contemporaryforms of democratic development. The empirical basis of this investigation derives from
ethnographicresearchin the metropolitanregions of Sao Paulo and Brasilia,
which we have, at varioustimes, conductedseparatelyand together.It derives
especially from an anthropologicalconcernwith the performativedimensionof
social and institutionalrelations-that is, with the representativepracticesand
exemplaryparticularsthroughwhich these relationsareenacted,as well as with
the scripts, like democracy,that are supposed to provide a calculus for many
sets of relations, and that people must performto gain the prescribedeffects.
Ourintention,in this sense, is notjust to criticize the strictlypolitical definition
of democracy,but also to suggest an anthropologicalperspectivein its study.
We do not insist on this intentionby calling attentionto it throughoutthe discussion. Rather,we try to demonstrateits force by focusing on the civil componentof citizenshipand the lived consequencesof its violation, andby letting
these social practiceslead to a theoreticalargumentabout the disjunctivenatureof democratization.4
4 This
essay began to take shape in threepresentationswe deliveredtogetherin 1994 and 1995.
The first was given at CEDLA (Centerfor LatinAmericanResearchand Documentation)in Amsterdam.The second was at the conference"FaultLines of DemocraticGovernancein the Americas," sponsoredby the North-SouthCenterat the Universityof Miami.The thirdoccurredat a meet-

DEMOCRACY

DEMOCRACY,

VIOLENCE,

AND

AND

VIOLENCE

IN BRAZIL

695

INJUSTICE

In recent years, Braziliansociety has producednumerousevents indicative of

a disjunctivedemocratization.Some point to an expansion of democraticcitizenship, and othersto its erosion and degradation.Some indicatethe strengthening of democraticstate institutions,and others the dismantlingor "illegalization" of the state. Most frequently,events of contradictorymeaning are
coeval. The expansion of democracyis suggested by the organizationof new
social movementsthroughoutBrazil duringthe late 1970s and 1980s, the proliferationof NGOs in the 1990s, the revitalizationof tradeunions, and the organizationof variousinvestigativecommissions by the federallegislature,one
of which provoked the 1992 impeachmentof PresidentFernandoCollor de
Mello for corruption.There have also been regular,lawful, and generally unproblematicelections at all levels of government,andthe creationof numerous
new political parties, including the PT, the Workers'Party.Events indicating
the degradationof democracyinclude the rise in violent crime, police violence,
and humanrights abuses, all of which increaseddramaticallyafter the institutionalizationof democraticrule. This dismal recordof violence againstmostly
innocent and unarmedcivilians includes the massacre of indigenous populations, peasants,ruralleaders (like Chico Mendes), streetchildren,adolescents
in poor urbanneighborhoods,and prisoners (as in Sao Paulo's Casa de Deten;ao, where militarypolice killed 111 unarmedprisonersin a 1992 prisonrebellion). Police violence has reached unprecedentedlevels, and the forces of
law andorderarethemselves one of the main agents of violence in many cities.
Various police forces are plagued by corruption,entangled with organized
crime, and accustomedto violent and illegal methodsof action.
Withthe increasein criminalandpolice violence, public space in manycities
has become characterizedby muggings, assaults of various kinds, shootings,
drug traffickingand addiction,violence in traffic, and a general scofflaw attitude. Violence of one sort or anotheris a common experience of daily life. As
a result, a cultureof fear and suspicion has takenroot, giving supportto extraordinaryand often extra-legal measures for dealing with violence and crime.
For example, cities such as Rio de Janeirohave enactedpolicies that are questionablefrom the point of view of democraticconsolidation.One of these was
OperagaoRio in 1994, duringwhich the armywas sent into the city of Rio de
Janeiroin an effort to control violent criminal activity. People applaudthese
militaryoperations.They also supportillegal andprivate"actsof justice," such
ing on "Direitos,Identidadese OrdemPublica"sponsoredby theAssociacao Nacionalde P6s Graduagco e Pesquisa em Ciencias Sociais (Anpocs) of Brazil. We would like to thankthe sponsorsof
these events for their supportand encouragement.Ourpresentationat the Miami meeting in 1995
was published in a volume of conference papers, Fault Lines of Democracy in Post-Transition Latin

America, ed. Felipe Agiiero and Jeffrey Stark (Miami: University of Miami North-SouthCenter
Press) 1998:263-96. We utilize some data and passages from these papershere.

696

TERESA

P. R. CALDEIRA

AND

JAMES

HOLSTON

as the extractionof confessions throughpolice torture,and vigilante efforts to

catch suspects. In all cases, the assertionof episodic ordersupercedesconcern
for the institutionalorderof democraticlegal normsand procedures.Although
people are ready to defend democraticproceduresin the political system by
consistently supportingtrials of corruptpoliticians at all levels and defending
free elections and political organization,they also overwhelmingly welcome
actions like OperacaoRio. In the context of crime, fear, and the failure of the
institutionsof law, people considerdiscussionsaboutthe legitimacyof the military occupationin Rio or the prison assaultin Sao Paulo-and the threatthey
pose to the consolidationof democracy-largely irrelevant.They also consider that suspectedcriminalshave no humanrights to safeguard,and expect the
police to respond to violence with violence. In what follows, we look more
closely at severalkey elements thatdefine this perversityof violence underpolitical democracy.
ViolentCrime
In contemporaryBrazil, violence against civilians is the domain in which the
disrespect of civil rights and the failure of democratizationstrongly shape
everyday social interactions.Since the mid-1980s, Brazilianshave perceived
violent criminalityas the mainproblemaffectingtheircities. Not only has crime
increasedin this period,but the type of criminalityhas also changed.Crimehas
become moreorganizedandviolent, as the exampleof Sao Paulodemonstrates.
Sao Paulo has one of the highest rates of violent crime in Brazil. These rates
are also high when compared to many cities aroundthe world. In the early
1980s, violent crime representedaround20 percentof the total crime reported
to the civil police in the metropolitanregion of Sao Paulo.5 Since the mid1980s, this percentagehas been higher than 30 percent,reaching36.3 percent
in 1996. One of the crimesthatincreasedmost in the period 1981-1996 is murder (average annualvariationof 10 percent). In 1996, the rate of murderper
hundredthousandpopulationreached47.3, a value significantlyhigherthanthe
1981 rate of 14.62.6
In the last fifteen years, the proportionof violent deaths (accidents, homicides, and suicide) in totaldeathshas almostdoubledin the metropolitanregion
of Sao Paulo, accountingfor 8.95 percentof deaths in 1978, 15.82 percentin
1991, and 14.11percentin 1993.7Since 1989, violent deathshave been the sec5
Although thereis no official definitionof violent crime, for purposesof statisticalevaluation
we considerit to include murder,attemptedmurder,rape,attemptedrape,assaultand battery,robbery,and felony murder(latrocinio).
6 Rates of murderare based on the civil police recordsof reportedcrimes (boletins de ocorrencia).
7 Data on violent deaths are from the death registry.The two main sources of murderrates are
those of the police (reportedcrime) andthe healthauthorities(compulsorydeathregistration,compiled accordingto the InternationalClassificationof Disease categories).Althoughthe two sources
registersimilarpatternsof growth,the differencesbetween them are high. Rates by the deathregistry are on averagearoundthirtypercenthigherthanpolice reports.For a complete discussion of
violent crime statistics,see Caldeira(in press).

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

697

ond highest cause of death in Brazil (afterrespiratorydiseases), while in 1980

they were only the fourth(Souza and Minayo 1995:90). Murderis responsible
for the significantincreasein violent deaths, since the proportionof other"external causes" in the total numberof deaths has remainedrelatively constant.
While in 1978 murdercaused 1.44 percentof the deathsin the city of Sao Paulo,
in 1994 it caused 6.57 percent,an increaseof 356 percent.In 1994, murdersaccounted for 19.15 percentof the deaths of people between 20 and 49 years of
age in the municipalityof Sao Paulo, becoming the main cause of deathin this
age group.This rate is dramaticallydifferentfrom that of 1976, when murder
accountedfor only 4.9 percentof deathsin the same age group.In 1994, 44.40
percentof the deaths of people aged 15 to 24 were caused by murder.During
the 1980s, murdersincreased 80 percent among 10 to 14-year-olds (Souza
1994:49). In additionto murder'sincreasingeffect on the young (males more
thanfemales), thereareindicationsthatmurdervictims arepredominantlypoor.
Accordingto Pro-aimdatafor 1995, most of the districtsof the city of Sao Paulo
with the highest ratesof murder(ratesbetween 75 and 97 murdersper hundred
thousandpopulation)were very poor. In contrastto this, the lowest rates (between threeandfifteen murdersper hundredthousandpopulation)were in middle or upper-classdistricts.8
The increasein violent deathsis not a patternof Sao Paulo alone. Homicide
ratesincreasedin most Brazilianmetropolitanregions duringthe 1980s (Souza
1994:53-55). As a consequence,the homicide rates for Brazil (aroundten per
hundredthousand),which were similarto those of the United States in the early 1980s, more thandoubledAmericanratesby the late 1980s. The U.S. homicide rateis historicallyquitehigh comparedto WesternEuropeanandJapanese
rates.Duringthe periodfrom 1970 to 1990, Americanrateshave oscillatedbetween eight and ten homicides per hundredthousandpopulation,while European rates have fluctuatedbetween 0.3 and 3.5, and Japaneserates have remained at aroundone homicide per hundredthousandpopulation (Chesnais
1981:471). In other words, the contemporaryBrazilianhomicide rates above
twentyarevery high indeedif comparedto theAmerican,European,andJapanese ratesin the last few decades. However,nationalrateshide local disparities,
andmanyurbanareashave homicideratesconsiderablyhigherthanthe national
average.In the case of Brazil duringthe late 1980s and 1990s, Rio de Janeiro,
Recife, and Sao Paulo are the three most violent metropolitanregions, with
homicide rateshigher thanforty per hundredthousandpeople.9
8 Pro-aim(Programade Aprimoramentode Informacoesde Mortalidadeno Municfpiode Sao
Paulo) is responsiblefor the death statisticsof the city of Sao Paulo.
9 It shouldbe noted thatin the U.S., in 1993, some cities had muchhigherratesthanthese, while
otherAmericancities had rates comparableor lower to those of Sao Paulo.Accordingto the FBI's
UniformCrimeReportsfor the UnitedStates for 1993, some of the highest rates were in New Orleans (80.3), Washington,D.C. (78.5), Detroit (56.7), Atlanta (50.4), Miami (34.1), Los Angeles
(30.5), New YorkCity (26.5). However, the Brazilianhomicide rates have oscillated much more
than the Americanrates, and in many largeAmericancities these numbershave decreasedsignificantly since the early 1990s. It is hardto obtaincomparableinformationwith regardto otherthird
world countries.Nationaldatacompiled by the United Nations on causes of deathare not available

698

TERESA P. R. CALDEIRA AND JAMES HOLSTON

Anotherway of evaluatingthe increasein violence is to look at the registration of guns and reportsof illegal possession of weapons. The annualnumber
of registeredguns purchasedin the metropolitanregionjumped from 9,832 in
1983 to 66,870 in 1994, an increaseof 580 percent.These numbers,however,
are far from portrayingthe increase of weapons among the population,since
the apprehensionof non-registeredguns has also increasedconsiderably.Police reportsof illegal possession of guns in Sao Paulo grew an average of 9.5
percent a year between 1981 and 1996. In 1996, the police registered5,563
cases of illegal possession of guns in the metropolitanregion. As reportedin
the media, many of the apprehendedguns are smuggled into the country and
some (especially those used by drugdealers)aremorepowerfulthanthose used
by the police. The increaseof gun possession correlateswith the fact thata higher proportionof homicides arecommittedwith them.Accordingto dataof death
registration,in 1980, homicides by firearmsconstituted14.8 percentof the total of homicidesin Sao Paulo;in 1989, they were 31.2 percent(Souza 1994:55),
and in 1992, 29.26 percent.The increasein the possession of guns indicatesnot
only an increase in crime and violence, but also shows how Sao Paulo's residents are increasinglytaking the task of defense into their own hands, a practice we discuss later.
In the daily life of cities such as Sao Paulo, an importantaspect of the increase in violent crime is what Caldeiracalls "the talk of crime,"a proliferation of everyday narratives,commentaries,and even jokes that have crime as
their subject (see Caldeira,in press: PartI). This talk producesand circulates
stereotypes,bothcounteractingandprovokingfear.The narrativesof crimethat
emerge in the course of the most diverse and common conversationsoperate
with clear-cutoppositionsand essentializedcategoriesderivedfrom the polarity of good versusevil. They help to symbolicallyreordera world disruptedby
experiences of crime. But they do this in a complex and particularway. Their
reorderingbothcountersdisruptionscausedby violence, andmediatesandproliferatesviolence. More thanmaintaininga system of distinctions,narrativesof
crime createstereotypesandprejudices.They separatecategoriesof people and
reinforceinequalities.In addition,the categoricalorderarticulatedin the talk
of crime is the dominantorderof an extremelyunequalsociety.As such, it does
not incorporatethe experiences of dominatedBrazilians-the poor, migrants
from the Northeast,women, and others.Rather,it usuallydiscriminatesagainst
and criminalizesthem. The talk of crime is a productivediscoursein the sense
that it helps to producesegregation(social and spatial), abuses by the institutions of order,the negationof citizenshiprights,and,especially,violence itself.
for most African and Asian countries.LatinAmericancountrieshave had relatively high rates in
the 1990s. Colombiahas one of the highest ratesin the world:74.4 in 1990. Brazil (20.2 in 1989),
Mexico (17.2 in 1991), and Venezuela(12.1 in 1989) have the next highest. Data for LatinAmerica are from United Nations (1995:484-505) and refer to deathrates compiled by health authorities. Local situationsmay differ considerablyfrom nationalaverages.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

699

If the talk of crime generatesorder,it is not a democratic,tolerant,egalitarian

order,but the opposite.
It is not our intentionin this essay to determinethe underlyingcauses of the
increasein violent crime and the breakdownof the institutionsof law over the
past fifteen years. No doubt,the developmentof organizeddrugtraffickingand
relatedpolice corruptionin Rio de Janeiroand Sao Paulo since the mid-1980s
is an importantfactor.So, too, is the contradictionbetweenBrazil'stremendous
wealth(its industrialeconomy currentlyrankseleventh amongall nations),and
its distributionof income, which is one of the worstin the world.Madeexplicit
to Braziliansby modernmass media, this gross inequalityis, surely,at the root
of much violence. Nevertheless,our argumentis not aboutcauses, but correlations. Whateverthe origins of the increasein violence, political democracyhas
not been able to deter it or even use its occurrenceto remakerelevantinstitutions. Rather,the violence has shown Brazilian democracy-and, perforce,
Braziliansociety-to be especially weak in the areaof civil citizenshipandthe
protectionof civil rights. In turn,this weakness generatesmore opportunities
for violence to proliferate.Moreover,political democracyhas not been able to
dispel the causal link that many Brazilians-especially those on the rightmake between democracyand violence: that is, that the institutionof democracy itself is responsible for the propagationof the new violence. This is not
only because democracydefends humanrights for criminalsuspects and prisoners, therebysupposedlycreatingdifficultiesfor the police and incentives for
criminals;it is also, so the claim goes, because democracy has destructured
Braziliansociety by giving the masses a sense of political power thatconfuses
their sense of social hierarchy-so much so that "peopleno longer know their
place." Our argumentis that the response of a significantpartof the Brazilian
elite to this perceiveddestabilizationhas been to criminalizethe poor-by campaigning againsthumanrights, by flooding the media with narrativizedcrime
stories, by investing in private security and private "justice,"by retreatingto
fortified enclaves-and that this criminalizationalso contributessignificantly
to the propagationof violence.
Although the returnto democraticelections at all levels between 1985 and
1989, and the promulgationof a new Citizens' Constitutionin 1988 heralded
the creationof a new nationalpublic sphere with new freedoms and forms of
participationfor citizens, violence and the fear of violence have eviscerated
public confidence. Although political democracypromised a new horizon of
equality and fairness-a new sense of the futurefor Brazilians-its incapacity to deal with violence has eroded that horizon for many. It is interestingto
note that this promise was bolsteredby the Cardosoadministration'ssuccessful reductionof hyperinflationin 1994 to practicallyzero. In theory,the end of
rampantinflation should have expanded the sense of a productivefuture for
working Brazilians. It should have created a new time-line against which to
measurepersonaland family progress.But apparentlyit has not. In interview

700

TERESA

P. R. CALDEIRA

AND

JAMES

HOLSTON

after interview with young adults who live in the peripheryof Sao Paulo, we
found thatthey despairedof reachingthe normalizedgoals of theirparents,one
symbolized by a family, house, and steadyjob. This combinationof reduced
horizons is the context in which both violence and political democracydevelop in Brazil.
Two other elements contributeto the experience of violent crime and fear
thatframesthe everydaylife of Brazil's citizens: police violence and the inadequate response of the justice system. The failure of the institutionsof law to
combat increasingviolence and the fact that those same institutionsoften add
to the violence put the populationunderconsiderablestress. They contribute
not only to increasingthe fear of powerlessness but also to justifying private
measuresof protectionand vigilantism.
ViolentPolice
A dramaticindicationof the role of the police in the reproductionof violence
in Sao Paulois the dataon the relationshipbetweenthe numberof people killed
by the police and the total numberof murders.From 1986 to 1990, the police
committed10 percentof the total numberof killings in the metropolitanregion
of Sao Paulo;in 1991, the percentagejumped to 15.9 percent,and in 1992 rose
to 27.4 percent.A comparisonindicatesthe absurddimensionof these rates.In
New YorkCity in the 1990s, the averagepercentageof police killings has been
1.2 percent,and in Los Angeles, 2.1 percent.Althoughpolice violence has diminished in Sao Paulo since 1993, with the rates back to the level of the late
1980s, no othercity in the Americasoutside of Brazilhas a comparablerecord
of police abuse in the use of deadly force (Chevigny 1995). In Brazil, the police constitutepartof the problemof violence.
Fromits creationin the early nineteenthcentury,the Brazilianpolice's practices of violence, arbitrariness,discrimination,and disrespect of rights have
been well known.Althoughthe degree of police abusehas variedunderdifferent political regimes, duringthis entireperiodthe police have never abandoned
the practices of unsubstantiatedarrest,torture,and battering.These practices
have not always been illegal, and they have often been exercised with the support of the citizenry,even of membersof social groupswho have been the police's preferredvictims. Throughoutthis period, differentgovernmentsissued
"lawsof exception"to accommodateexisting delinquentpolice practices,or to
cover them up. Although these laws were usually issued under dictatorships,
they frequently survived under democratic rule. Thus, the legal parameters
framingpolice workhave often shifted,makingthe boundariesbetween the legal and the illegal unstable, and creatingconditions for the continuationof a
routineof abuses. The repressionof crime has targetedthe workingclasses in
particular,andhas frequentlymergedwith politicalrepression.The formulathat
the elites of the Old Republicmade famous has remainedin place: "Thesocial
questionis a matterof the police." Consequently,the poorersectorsof the pop-

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

701

ulation in particularhave unremittinglysufferedvarious forms of police violence and legal injustice.As a result, the poor have learnedto fear the police
and distrustthe justice system.
It is not necessaryto review the entirehistoryof the Brazilianpolice to make
these points. However, discussing a few contemporaryaspects of this history
will demonstratethe complicatedrelationshipbetween the police and the legal
order.The militaryregime thattook power in 1964 reorganizedthe police. Decree 667 of 1969 unified all preexistingstate uniformedpolice into a statemilitarypolice force, subordinatedto the armyand chargedwith uniformedstreet
patrolling.The objective was to train and organize this new police force accordingto a militarymodel. At the same time, the civil police continuedto exist, comprisingthe administrativeand the judiciarypolice. The 1988 constitution preservedthe dual structureof the police forces afterthe end of the military
regime. Both the civil and militarypolice forces are organizedat the state level and are underthe jurisdictionof the secretaryof public security.However,
they have differenthierarchies,training,and recruitmentprocedures.In spite
of theirunified authority,this dual organizationgeneratesconstantrivalriesand
conflicts between the two police forces. The two also seem to specialize in different types of abuses. As many human rights organizationshave shown, the
civil police, who are in charge of investigations,tend to torturepeople under
arrest,while the militarypolice are more likely to kill suspects.'?
Rules governing the currentmilitarypolice include some laws of exception
thatput them above the civil justice system. Decree-Law 1,001 of 1969-still
in force-establishes that all crimes committedby militarybodies should be
consideredmilitarycrimes andjudged by a specialmilitaryjustice, even if such
offenses were committedin peacetime and in pursuitof civilian functions. In
otherwords, since 1969 therehas been a specialjustice for the militarypolice.
This exception became the norm with the constitutionof 1988. Writtenunder
a democraticruleby a freely elected congress,the 1988 constitutionmaintained
the militarypolice as the institutionin chargeof "theostensive policing and the
preservationof the public order"(art. 144, par.5) andthe militaryjustice as the
jurisdictionfor dealing with crimes committedby militarypolicemen. In May
1996, aftera massacreby the militarypolice (in Para,in northernBrazil), President FernandoHenrique Cardoso supporteda project in Congress that proposed that militarypolicemen be tried by civil courts. Nevertheless,this project did not win congressionalapprovaluntilAugust 1997 (Law 9299), andthen
in a milder form. Its approvalin this form indicates the supportthatthe police
enjoy, despite their violent nature.The new law shifts jurisdictionto ordinary
courtsin murdersinvolving militarypolicemen and soldiers. However,all other crimes, including manslaughterand physical assault,remainin the military
10 See, for example,AmericasWatchCommittee 1987. Also see Pinheiro(1991) for one of the
first analyses that demonstratesa patternof abuse by the militarypolice.

702

TERESA P. R. CALDEIRA AND JAMES HOLSTON

system. The problemis that the right to characterizea killing as murderor as

manslaughterremainswith militarypolice investigators.Obviously,this limits
the impactof the law. Nevertheless,it is an indicationof the Cardosoadministration'sconcernto curbhumanrights violations, as we discuss below.
Varioushumanrights groups have amassed considerableevidence demonstratingthatthe militaryjustice overwhelminglyacquits(or dismisses charges
against) militarypolice officers accused of crimes againstcivilians. This evidence confirmsthatthe militaryjustice is rigorousas far as internaldiscipline
is concerned,but lax if the questionis the murderof civilians. ' The conclusion
is clear:this recordof acquittalor dismissal stimulatesan explicit sense of impunity among the police, and thereforeperpetuatesthe continuationof abuses
associated especially with excessive use of force. As Paul Chevigny (1995)
demonstratesin his analysis of police abuse in six cities in the Americas,a decrease of abuse is directly related to the enforcementof systems of accountability.When the police are not made accountablefor theirextralegalor illegal
behavior, violence and abuse escalate. The legal exception that removes the
Brazilianmilitarypolice from the civilian justice system of accountabilityincreases their impunityand their use of violence in dealing with civilians, and
indirectlyassuresthem of a wide marginfor arbitrarybehavior.
These consequences can also be demonstrateda contrario, by analyzing
cases in which accountabilityprovokeda decline in police abuse.Therearetwo
such examples from recent times. During the 1970s, membersof Sao Paulo's
civil police organizeda famousdeathsquadcalled the Esquadraoda Morte.Because they were under the jurisdiction of the civil police, judges and public
prosecutorswere able to bring them to trial-even under military dictatorship-and ultimatelyto dismantlethe squad.In recentyears,judges and prosecutors have also been able to enforce the articleof the 1988 constitutionthat
considers torturea crime not subjectto bail or executive clemency.They have
broughtcivil police officers to trial, and there are indicationsthat torturehas
diminished somewhat in Sao Paulo's civil police precincts (Americas Watch
Committee 1993:21).
In sum, althoughunderdemocraticrule, the currentorganizationof police
institutionslargely maintainsthat establishedby the militaryregime. This institutionalframeworkin large measureassures the impunityof extralegal actions by the police-especially the militarypolice, the principalrepressivepoI Centro Santos Dias, a humanrights defense group associated with the Archdiocese of Sao
Paulo,analyzed380 trialsat the courtsof the militaryjustice from 1977 to 1983 (unpublisheddata).
For these years, it found that of eighty-two police officers accused of murder,only fourteenwere
found guilty (15.9 percent).Among forty-fourpolice officers accused of crimes against property,
fourteenwere found guilty (31.8 percent).Finally, among fifty-threeofficers who faced trials for
mattersof discipline, twenty-eightwere found guilty (52.8 percent).More recently,anotherstudy
by the Nucleo de Estudos da Violencia from the Universityof Sao Paulo found that in 1995, from
a total of 344 cases that reachedthe thirdAuditoriada JusticaMilitarof Sao Paulo, 58 resultedin
convictions (16.9 percent),and 190 in acquittal(55.2 percent),while 96 (27.9 percent)were dismissed before going to trial (unpublisheddata).

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

703

lice force of the state. In this sense, insteadof helping to curbarbitrarinessand

violence, the currentstructureallows space for these practices to proliferate.
This situationreveals an especially significantpoint: usually,people associate
abuses such as unsubstantiatedarrest,tortureof prisoners,and killing of suspects with authoritarianregimes. In Brazil, however,not only do police abuses
continueto exist, but all dataalso indicatethatthey have reachedunprecedentedly high levels underthe presentpolitical democracy.12
Thereis, by now, a mountainof dataon police torture,battering,degradation
of prisoners,and excessive use of deadly force. They show that the worsening
of police violence in Sao Paulo and other Braziliancities coincides with the
consolidationof political democracy.The numberof civilians who died in confrontationswith the militarypolice in Sao Paulo increasedconsiderablyin the
late 1980s:it surpassed500 in 1989 and 1990, reached1,171 in 1991, and 1,470
in 1992 (includingthe 111 prisonersof the Casa da Deten9ao).A special division of the militarypolice called ROTAcommits a significantnumberof these
killings. Not surprisingly,ROTAwas createdduringthe militaryregime (1969)
to fight terroristattacks,especially bank robberies,in the metropolitanregion
of Sao Paulo. A few comparisonshighlight the absurdityof the numbersof
killings. In 1991, when the militarypolice in Sao Paulo killed 1,171 civilians,
the New York City police killed 27 people in confrontationsand the Los Angeles City police killed 23. In 1992, the numberswere 24 in New York,25 in
Los Angeles (Chevigny 1995:46, 67), and 1,470 in Sao Paulo. The Sao Paulo
killings suggest somethingmore like a civil war or a regime of terrorthananything resemblinga democracy.
The civilian deathsin Sao Paulo cannotbe consideredaccidental,or a result
of the increasedviolence by criminals,as the militarypolice claim. If the latter
were the case, we would expect thatthe numberof police killed would also increase. But this has not happened.Although high when comparedto the U.S.
statistics, the numberof Brazilianmilitarypolice killed in confrontationshas
stayed more or less stable during the 1980s and 1990s, averagingthirty-nine
peryear.Moreover,the proportionof civilians killed in Sao Paulo in relationto
those woundedis completely abnormal.Usually,the police wound many more
thanthey kill. Duringthe 1980s and 1990s, for each civilian killed in New York,
an average of three were wounded;in Los Angeles, the ratio was 1 killed to 2
wounded.Yet in the metropolitanregion of Sao Paulo,4.6 civilians were killed
for each personwoundedin 1992 by the police; in 1991, the ratiowas 3.6 killed
12 One might argue that this increase is artificialbecause censorshipwas effective undermilitary dictatorshipand curtailedinformationabout police abuse. Under democraticrule, this information is readily available in the everyday media. It is true that much more of this kind of information is accessible today and statistics are better, though still far from reliable. However,
accordingto every availablecomparativeindex, it is indisputablethatnot only have the victims of
police abuse changedfrom political to civilian, but also the numbersof victims have risen dramatically in recentyears.Moreover,frequentmediaexposureof violent police actionshas usuallygeneratednot condemnation,but supportamong the populationat large.

704

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

to 1 wounded;for otheryears in the 1980s and 1990s there was an averageof

more than two deaths for each person injured.These data indicate that the
police in Sao Paulo and in other Braziliancities-such as Rio de Janeiroand
Recife, for which there are similar data-shoot to kill ratherthan to subdue.
As shown below, the practice of shooting to kill not only has broad popular
supportbut is also "accepted"by the "toughtalk"of official policy.
Probablyno single event more tragicallyexemplifies the routineassociation
of these componentsof police violence than the massacreof 111 prisonersat
Casa de Detenqcoin October 1992. Not a single police officer died; none was
woundedseriously.Yet prisonerswere randomlyshot and summarilyexecuted
after surrendering.In a countryin which humanrights violations coexist with
a free press, images of the massacreended up in the mass media, revealing a
concentrationcamp vision of piles of bullet-riddledbodies on the prisonfloor,
nakedinmatescarryingcorpses,androws of open woodencoffins arrangedside
by side along the corridorsof the Instituteof Legal Medicine.Althougha civilian criminalprosecutorpresentedchargesagainstone of the commandersof the
operationand a militaryprosecutorpresentedchargesagainst 120 officers and
soldiers for various crimes, including homicide, not a single one has been
broughtto trial almost five years afterthe massacre.Most of the accused continue to hold theirjobs in the militarypolice and to participatein public life. A
few have even run for elected office on the basis of their performancein the
assault.13
The massacreat the Casa de Detencao is an egregious example of what has
become an accepted, if not encouraged,routineof police abuse. Some of this
violence is no doubt associatedwith corruption.Possibly, some violence may
be accountedfor in terms of Brazilianjudicial procedures.Especially in the
precinctsof the investigativecivil police, the prevalenceof routinetorturemay
well be relatedto the fact thatconfession is still the centralpiece of judicial evidence. The hellish condition of prisons surely has something to do with restrictedstate budgets. Police violence has, however, even more to do with official policy and popularsupport.We can see the effect of the formerthrough
a contrastamong the experiences of several Braziliangovernors.Sao Paulo's
first governor after the end of militarydictatorship,FrancoMontoro (198387), and governorMairioCovas (1995-99, who has been reelectedfor another
3 For
example, in 1994, MP Colonel UbiratanGuimaraes,the commanderof the Casa de Detencaooperation,ranfor a seat in the stateassembly.Among the manyshockingaspectsof his campaign was his choice of the number111 to identifyhimself as a candidateon the ballot, exactly the
numberof prisonerskilled in the assault.Althoughhe received 26,156 votes, it was not enough to
be elected. More recently,in May 1996, the EighthHouse of Public Law of the Courtof Justiceof
the State of Sao Paulojudged the police action "legitimate"and absolved the state of all civil and
financial responsibilities.The superiorjudge who heard the case, Raphael Salvador-also vicepresidentof the PaulistaAssociationof Judges-justified his decision by blamingthe prisonersfor
the massacre:"They startedthe rebellion,destroyedthe prisonblock, and forced society, through
its police, to defend itself" (Estadode Sdo Paulo, 4 May 96: Al). Othercourtsare also hearingthe
case, and the decision is underappeal.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

705

term) were particularlydeterminedto cut down the abuses of the police, improve humanrights,andenforce the ruleof law. They were able to achieve their
goals partially,in spite of strongopposition, not only from the police but also
fromright-wingpoliticiansandthe populationat large,especially in Montoro's
case.
In contrast, the two governors who succeeded Montoro, Orestes Quercia
(1987-91) andLuizAntonioFleuryFilho (1991-95), adoptedthe oppositepolicy.14They believed that only a "tough"(meaningviolent) police force would
be able to curbthe risingratesof crime,andthey reversedall of Montoro'smeasures aimed at controllingpolice abuses. Undertheiradministrations,the number of deaths caused by the militarypolice rose continuously.After the massacre at the Casa de Detencao, when the annual number of police killings
approached1,500, Fleury was forced to substitutehis own secretaryof public
securityandchangehis "tough"policy. As a result,the numberof people killed
by the militarypolice droppedto 409 in 1993 and 453 in 1994. The numbers
continued to decrease during Mario Covas's administration,reaching 250 in
1996 and 1997. The new policies adoptedby Covas include strictercontrolof
the use of deadly force, the creationof an ombudsmanfor the police, and the
adoptionof a State Programfor HumanRights in 1997, which replicates the
NationalPlanfor HumanRightsadoptedby the Cardosoadministrationthe previous year.Although the new policies adoptedat both state and federal levels
have had positive effects on efforts to control the disrespectof humanrights,
they arenot easy to implement.This is evident in the numberof civilians killed
by the police, which is still extraordinarilyhigh accordingto internationalstandards.It is also demonstratedin a series of strikesandriots by the police forces
of variousstatesduringthe monthsof JuneandJulyof 1997, in responseto state
initiativesto reformtheirstructure.At thatmoment,Congresswas debatingthe
law thatwould makemilitarypolicemen accountableto the civil courts,the federal government(throughits nationalsecretaryof humanrights) was elaborating a projectof police reformto be sent to Congress,and GovernorCovas presented a proposalfor transferringall patrollingactivitiesto the civil police and
eliminatingthe division between the two police forces. The aggression,strikes,
and public demonstrationsby the police, as well as exchanges of gunfire and
aggressionbetweenthe two forces, were broadlydocumentedby the mediaand
indicatetheir deep resistanceto reform.
This resistanceis ultimatelygroundedin broadpopularsupportfor a "tough"
police force. The hard fact is that a significant part of the populationof Sao
Paulo supportsviolent police action. For example, accordingto variousnewspaper surveys conductedat the time, between twenty-nineand forty-fourpercent of the residentsof Sao Paulo supportedthe police in the massacreof prisoners at the Casa de Detenico. Every serious study of popularsentimentshows
14

IV).

For a complete analysis of Sao Paulo'spolicies of public security,see Caldeira(in press:Part

706

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

widespread approvalof violent police actions in dealing with criminals, including tortureand killing. Thus, the police have continuedto be violent, not
only becauseof an assuredimpunity,but also becauseof the supportof the population.Paradoxically,even the main victims of police violence-the working
classes-support some of its forms. Given this widespreadapproval,it is not
surprisingthatthe majorityof the populationarehostile to the notion of human
rights and to campaignslaunchedby humanrights groupsto enforce a rule of
law that respects individual rights. Elsewhere, Caldeira(1991) has analyzed
this hostility and the failure of human rights campaigns to deal with it adequately.The point we wish to suggest here is thatthe population'ssupportfor
police violence indicatesthe existence not only of an institutionaldysfunction,
but also of a pervasiveculturalpatternthatassociates orderand authoritywith
the use of violence, and that, in turn,contributesto the delegitimationof the
justice system and the rule of law.15
This culturalidentificationgeneratesa whole series of ambiguousand confusing practices because most Brazilians-particularly poor and/or black
ones-also fear the police.16 Most poor people have experiencedpolice mistreatmentandabuse,andtheirnarrativesaboutthem arefull of indignation.On
the one hand, they consider that the police routinely mistake "workers"for
"criminals"(these two being opposed local categories) and are thereforeviolent with them. On the other,they believe thatthe police aresoft with real criminals, who can bribethem, but hardwith honest andpoor workers,who cannot.
In this triangularrelation,people tend to express a confusion among all vertices: the police treatworkersas criminals,workersview the police as corrupt,
and in some cases workerseven considercriminalsas protectingthem against
the discriminationof the police. Moreover,when both rich and poor describe
the police as workers,they mean to emphasizethat they are not well prepared
for theirjobs, because they come from the lower classes and thereforelack the
requisiteeducation,leadership,and good judgment.Even when the poor themselves express it, such criticismof the police tends to be combinedwith prejudice againstthe poor.
In this context of confusion, in which police violence is praisedand feared,
15 This culturalpatternis quite complex. One of its main elements, which we cannot analyze
here, is a certaindominantconception of the body that Caldeira(in press) calls the "unbounded
body."It is a concept with two complementaryaspects. First, the unboundedbody is one around
which thereareno clearboundariesof separationand avoidance,a body thatis permeableandopen
to intervention,and that can and even should be manipulatedby others. Second, the unbounded
body is unprotectedby individualrights and indeed results historicallyfrom the lack of their enforcement.Thus, in Brazil, where the judicial system is openly discredited,the body (and the person) are generally not protectedby a set of rights thatbind it, in the sense of establishingbarriers
and setting limits to interference,intervention,or abuse by others. In Brazil, the unboundedbody
is evident not only at Carnival,but also in the extraordinarilyhigh rates of caesareanbirths,cosmetic surgery,and physical punishmentof childrenby parents.
16 This discussion about the population'sview of the police and the justice system is based on
researchconductedby TeresaCaldeiraamongresidentsof all social classes in Sao Paulo.The complete study is presentedin Caldeira(in press:PartII).

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

707

desiredand distrusted,people do not easily associatethe police with the rule of

law. The synecdoche "Herecomes the law," used to refer to the arrivalof the
police in AmericanEnglish does not exist in BrazilianPortuguese.The Brazilian identificationis instead with "authority"and then with the abusive and often violent use of it. People do not consider the police in terms of law, rights,
and citizenship-not to mentionjustice and fair treatment-but ratherin terms
of incompetence,corruption,injustice,and brutalforce.
Judicial Discredit

Caughtin this combinationof political democracyand violence, the vast majority of Brazilians are resigned to an undemocraticfate: they cannot rely on
the institutionsof stateto secure theircivil rights,eitheras positive protections
or as negative immunities.Moreover,once theirrightshave been violated, it is
equally unlikely for Braziliansto expect redressthroughthe courts.For example, in April 1998, the state of Rio de Janeirocreateda judicial ombudsmanof
the police (one internalto its organization)to judge citizen complaintsagainst
both civil and militarypolice. In the first nine months of operation,this court
received 1,586 complaints,includingchargesof torture,extortion,and abuseof
authority.It decidedon indictmentsin 20 percentof the cases andhandeddown
convictions in 7 percentof them (112). The sentences carriedvariouskinds of
punishments,includingprison,demotion,and warning.In fact, however, not a
single convicted policeman remainedin jail or was expelled from the police
force.
This example is representativeof judicial ineffectiveness: Brazilians perceive the judiciaryas an institution"withoutteeth"in most cases, an incapacity which creates the belief that crime pays, that breakingthe law goes unpunished, that citizens cannot enforce their legal rights, and that their legal
disability allows criminals to act with impunity.It is not that people have no
hope. Theirwillingness to try new inventionslike the police ombudsmanbelie
such a possibility. Nor is it that there are not adequatelaws on the books. It is
ratherthatthe courtsand the police cannotmake the law "stick."Justas the police do not representthe law-as-rightfor most people, thejudiciaryis so remote
as a reliableresourcethatmany residentsof Sao Paulo did not even mentionit
in our interviews about violence and crime. When they did respondto a specific question about the judiciary,their reply was most often some version of
"It'sa joke!" Even for educatedBraziliansthe judiciaryis a closed, conservative, enigmatic institution,protectedby practicallyimpenetrablebureaucratic
formalitiesand fiercely defended corporateprivileges. Beyond a very narrow
professionalcircle, remarkablylittle is known about its personneland organization.Nationally,aboutseventy-twopercentof all Braziliansinvolved in criminal conflicts do not use thejustice system to resolve theirproblems,according
to data gatheredin 1988 (PNAD statistics, cited in Adorno 1994:136). When
conflicts occur,people frequentlyuse the phrase"Go look for your rights"(Vai

708

TERESA P. R. CALDEIRA AND JAMES HOLSTON

procuraros seus direitos).This phraseoriginatedin the field of laborlaw, where

it means thatalthoughrightsexist, they arehardto realize withoutmuch struggle (typically againstbureaucracy).It can also be used in a more cynical tone
by someone accused of an offense. In such cases, it means that even if the accuser has rights, he or she will never realize them throughthe justice system.
The message is to forget the law and either accept what happenedor try for an
extrajudicialresolution.
In the discussion thatfollows, we want to emphasizethatour focus is on the
judicial protectionof the civil component of citizenship, especially on those
civil rights concernedwith life and liberty and effected by urbanviolence. In
some other areas of law there have been hopeful changes in recent years. For
example, the Public Ministryhas developed into an proactiveand sometimes
effective prosecutorialinstitution,chargedwith defendingthe public interest.17
In 1990, a nationalconsumerprotectioncode establishedeffective consumer
rights and continues to support their enforcement through special "proconsumer"offices thatreceive and evaluatecomplaintsin most cities.18In addition, in 1998, both federaland several statelegislaturesinstitutedparliamentary investigativecommittees(CPIs) on drug traffickingand organizedcrime,
with broadpowersto subpoenawitnesses, review bankandtelephoneaccounts,
and issue warrantsof arrest.Althoughin manycases the courtshave voided the
arrestsfor lack of sufficientevidence, the CPIs have neverthelesssucceededin
leading an offensive against organized crime, based on judicial-like public
hearingsthatgeneratemuch mediacoverage andpublic support.Moreover,the
federal legislatureopened a CPI to investigate the judiciaryitself for various
kinds of corruption,with the ultimateobjective of formulatinga comprehensive reformof the courts. Needless to say, the judiciaryrebelled. Many of the
highest judges simply refused to testify or cooperate.Nevertheless, after only
a few monthsof operation,the CPI uncoveredseveralcases of stupendouscorruption,giving it public legitimacy that the judiciarycould not deny. Finally,
we wish to recognizethattherearemanyhonest,dedicated,andeffectivejudges
and prosecutors.They often wage what amountto heroic strugglesagainstentrenchedcorporateinterestsand archaiclegal practicesand culturesthat have
little to do with a democraticprojectof justice.
Althoughpolice abuse has received considerableattentionin the evaluation
of Brazil's democratization,the judiciary'sfailure to achieve to a democratic
rule of law has received comparativelylittle consideration.However,the judiciary's prevailinginabilityto secure and communicatea sense of effective justice, fair and timely treatment,andreasonableaccess for all Braziliansnot only
17
Althoughthereareto datefew books devotedto the developmentandsignificanceof the Public Ministry,see the volume edited by Vigliar and Macedo Jdnior(1999) for a recent discussion.
18 On consumerprotection,see Lopes (1992) and Macedo Jr. (1998). Publishedby the Instituto Brasileirode Politica e Direito do Consumidor,the Revista do Consumidorprovidesconsumer
advocateswith case studes,jurisprudence,analysis of judgments,and debates.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

709

rendersit an isolated and even irrelevantinstitutionfor most people: such failures also crippleBraziliandemocracywith an undemocraticrule of law. As we
discuss later,this impairmenthas not been well addressed,because most studies of democratizationassume that political democracy automaticallyor necessarilyproducesa democraticrule of law. They presupposethatthe rule of law
is, by definition, democratic,without establishingwhat the rule of law is, and
what would make it democratic.
In additionto ineffectiveness, isolation, and formalism,Braziliansview the
judicial system as profoundlyclass-biased.Although most, if not all, judicial
systems in the world betterserve the rich, the bias in Brazilis particularlyegregious becausethe overwhelmingmajorityof Braziliansarepoor.Forthe Brazilian rich,the courts,when needed,have been a reliablemeansof manipulationespecially because theirstaggeringbureaucraticcomplexitiestie up conflicts in
formal complication to the point of exhaustion and/or extra-judicialsolution
(see Holston 1991). For the poor, however, the judiciaryhas historicallybeen
little more thana source of humiliation.19This abusive rule of law embodies a
double discriminationthatis a "ruleof thumb"in Brazil:the poor suffer criminal sanctionsfrom which the rich are generallyimmune, while the rich enjoy
access to privatelaw (civil and commercial)from which the poor are systematically excluded. This double bias pollutes the entirefield of law, discrediting
the judiciary and the law generally as a means to justice. Thus, the courts do
not providea genuineforumwithinwhich contemporarysocial conflicts can be
engaged with a sense of fairnessandequalitybefittinga democracy.The courts
remainespecially ineffective in arbitratingsocial relationsin ways that would
impose sanctions on the offenses of the powerful and protect citizens from
abuse by the state and its agents. These incapacitiesproducegeneralized expectationsof eitherimpunityor abuse from the justice system.
Confirmationsof these expectations aboundin every area. Consider a few
examples. Between 1965 and 1990, the AmericasWatchCommittee(1991) has
registeredthe murderof 1,681 ruralworkers.Of these cases, there have been
only twenty-six trials and fifteen convictions. It is, moreover, uncertain,
whether any of those convicted have remainedin jail for any length of time.
The convictedmurderersof rurallaborleaderChico Mendesarea case in point.
The conclusion is certain:hired guns murderwith impunityin ruralland and
laborconflicts. In the case of child labor,the constitutionoutlaws the employment of those younger than fourteenyears of age. Yet, the 1991 nationalcensus shows that there are more than three million childrenbelow fourteenemployed in the formaland informaleconomy. Conclusion:the legislaturemakes
laws thatthe courtscannotor will not enforce,employersoperatewith impunity
in blatantviolation of the law, and workers are abused. With regardto urban
crime, of all incidents reportedby the civil police for the municipalityof Sao
19 The one
exception for the workingpoor is labortribunalsfor peculiarhistoricalreasons (see
Santos 1979).

710

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

Paulo in 1993 (389,178 boletins de ocorrencia), only 20.4 percentresultedin

the type of police fact-findingproceedings(inqueritosinstaurados)necessary
for judicial action of any sort. For the last decade, thatratehas variedbetween
seventeen and twenty-onepercent.In 1993, for crimes of murder,it was a low
73.8 percent,though for drug dealing, it reached94.4 percent (Seade, unpublished data).Althoughwe do not have data on the numberof conflicts that actually go to trial,it is widely thoughtto be low. Moreover,as we have seen, conviction does not necessarily mean punishment.20In the area of white-collar
economic crime, the federal governmentclosed several large private banks,
such as the Banco Economico and the Banco Nacional,in the last few yearsfor
executive fraud,illicit enrichment,and othercrimes. However,not one executive of these bankshas ever been broughtto trial,muchless put in jail. The governmentbailed out the banks and repaid all depositorsso that even investors
who made fortunesthroughrisky ventures(such as those who later defaulted
on loans at exorbitantrates)did nothave to assumeresponsibility.In these cases
and so many others, de facto impunity is the result of the rule of law under
Brazil's political democracy.
In evaluatingthejudiciary,it is also importantto considerthatthe courtshave
a special responsibilityin every democracyto protectcitizens from the abuses
of arbitraryexecutive action. Such protectionis surely among the foundational and legitimatingvirtuesof democracy.One of the most importantbarriersto
this abuse is the constitutionalrequirementthat the state may deprive an individual of basic liberties(such as life, assembly,and property)only on the basis
of law and its due process. Some form of this principleof legality appearsas a
fundamentalguaranteein every democraticconstitution.However, the application of this principleis the greaterhallmarkof democracy.Historically,consolidated democracies have depended on the judiciary-especially the high
court-to interpretthe normsof legality so thatthey apply to real social problems in ways congruentwith their intent. Most important,this applicationhas
dependedon the judicial interpretationof liberty.In the developmentof Western democracy,this interpretationhas entailed an expansionof the categoryof
libertyto includenot only freedomof contractandotherclassically liberaleconomic liberties,but also, and more fundamentally,the civil libertiesof speech,
assembly,personalsecurity,and so forth, consistentwith due process. To give
constitutionalnorms utility and relevance, the courts must grapple with the
question of what sort of proceduredue process requires,and what liberty entails.2
20
In addition,S6rgioAdornofoundthatbothconvictionand sentencingpatternsareaffectedby
racialbiases. In his study of 297 criminaljury trialsin Sao Paulo, he observeda convictionrateof
threeblacks to one white (Adorno 1994:140).
21 As David P. Currie
(1988) shows in his study of the Americanconstitution,for example, the
U.S. SupremeCourthas tended to equate due process with fairness, going beyond common law
procedureto meet that standard.It has interpretedlibertybroadlyto give civil rights not only the
protectionof due process but also to give "FirstAmendmentliberties a 'preferredposition' enti-

DEMOCRACY AND VIOLENCE IN BRAZIL

711

In Brazil, every democraticconstitutionsince the first RepublicanCharterof

1891 has contained adequateprovisions for due process and the fundamental
rightsof life, liberty,andproperty-provisions directlyinspired,in fact, by the
U.S. Constitution.In reality,however, Braziliancourts have consistently protected only property,and only certainkinds at that.22When tested, they have
not given life and libertyrigorousjudicial protectionagainstthe infringements
of the state. It is not just that citizens have not used the high court to protect
their non-economic constitutionalrights. It is also that the Braziliancourts do
not invite such use because they do not have a traditionof defending them.
Ratherthan forbiddingthe state to deprive persons of their rights, the courts
tend to acquiesce to thatdeprivationwhen they considerit at all-as the failed
challengesto governmentcensorship,illegal detention,or coercedconfessions,
for example, illustrate.Civil rightscases often languishfor years with no resolution untilthey become moot. Whathas been almostcompletely missing from
Brazil'sjudicial traditionis the sense that courts protectthe rights of citizenship and the principleof legality, even though these norms have been written
into every democraticconstitution.
What we want to emphasize in conclusion are the consequences of judicial
discreditfor the democraticrule of law: not only are civil rights generally unenforcedwithin thejustice system, but they are also skewed by the double bias
of impunityand exclusion discussed earlier.From the perspectiveof most citizens, therefore,the right to justice as a key civil right and matterof law lacks
both institutionalconsolidationand personalpractice.For example, due to this
discredit,the new social movementsof the 1970s and 1980s, which did so much
to generatea new conceptionof Braziliancitizenship,ignoredthe courtsas an
arenaof redress.These movementswere unprecedentedin theircreationof new
kinds of rights outside the normativeand institutionaldefinitions of the state
and its legal codes. In particular,these rightsaddressednew collective andpersonal spaces of daily life in the city, especially in the residentialneighborhoods
of the peripheries.As these "rightsto the city" expandedcitizenshipto new social bases, they also creatednew sources of citizenship rights. Yet, until very
recently,these social movementsbypassed the judiciaryin their struggles.Instead,they have mostly workedwith models of rightsthattend to limit the concept of citizenshipto political participationand thus political rightson the one
hand, and to insertioninto the system of governmentservices and thus sociotling them to greaterjudicial protectionthanordinaryeconomic liberties"(Currie1988:49).Thus,
to ensure a fair trial, the U.S. SupremeCourtdecided in 1932 that "due process entitles indigent
defendant[s]to counsel at state expense thoughcommon law did not"(Currie1988:124).
22 For
example, in a case of land disputeinvolving millions of poor families in the peripheryof
Sao Paulo, the SupremeFederalTribunalwas petitionedin 1957 (Acao Cfvel Orginaria164-A) to
sort out the propertyinterestsof the federalgovernment,the stateof Sao Paulo, and variousprivate
parties.By law, the Courthas to hear all cases broughtbefore it. Although this particularcase involves constitutionalissues that directly affect the well-being of countless citizens, it has simply
languishedin the court system withoutdecision for more thanforty years (see Holston 1991).

712

TERESA P. R. CALDEIRA AND JAMES HOLSTON

economic rightson the other.Withsome exceptions (such as minorityandfeminist movements),they have largely disregardedthe courtsas means of change
and focused instead on securingentitlementsdirectlyfrom the executive and,
secondarily,from the legislative institutionsof government.
Fromthe perspectiveof the courts,the neglect of the rightof all Braziliancitizens to justice and civil rights means that the judiciaryhas been very slow to
confrontthe transformationsof contemporarysociety. Of the branchesof government,it has remainedthe most resistantto democratictransformation.As a
result, the developmentof Braziliancitizenshipremainsstrikinglydisjunctive,
almost a decade afterthe successful institutionof politicaldemocracy.
The Privatization of Justice and Security

One of the most importantconsequences of judicial discreditis the privatization of justice and security.23The combinationof fear of the police and distrust
of thejustice system leaves people feeling vulnerable.Some resign themselves
to this feeling; others seek alternatives.These alternativesare usually outside
the boundariesof legality and are of two types. In one alternative,people considerreactingprivatelyand takingthe law into theirown hands.It is important
to add that such vigilantism is usually an alternativemore at the level of discourse than of practice,althoughlynchings have, in fact, increasedconsiderably in the 1990s. People express their discontent by defending personal
vengeance, which does not mean that they act vengefully, at least not as frequentlyandvehementlyas they defend such responses.In the otheralternative,
people supportthe use of deadlyforce againstalleged criminals.Both areparadoxical reactionsto a delegitimatedjustice system, for people usually want the
police-whom they fear and accuse of being violent-to be violent "toward
the side that deserves it," even though they know that the police routinelyaggrieve innocentpeople. Their intent is neverthelessclear:they want criminals
killed. Given theirdistrustof thejustice system, summaryexecutionby the police is the only guaranteedway to remove the threatof crime. But the paradox
remains:by supportingvigilantism and violent police methods, people both
propagateviolence and greatlyincreasetheirown chancesof becoming its victims.
Whatwe call the privatizationof justice and securitydoes not include the revenge killings that markorganizedcrime in Brazil, including those relatedto
drugtraffickingin thefavelas (squattershantytowns). Although such killings
increasethe annualmurderrate considerably,and althoughthey belong to the
same field of violence as privatesecurityandjustice, they are distinctfrom the
protectiveandreactivemeasuresconsideredby citizens who arevictims of fear
andcrime. Some writershave called these organizedcrime killings a system of
"alternativejustice."They do so to emphasizethatthese killings have become
23

See Caldeira(in press) for a fuller discussion of this problem.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

713

dominantin some favelas, organizingthe residents'everyday lives on the basis of terrorand filling the void left by the absence of state institutions,especially the justice system. We thinkthatto considerthem acts of justice confuses the importantanalytic distinctions between revenge, self-help, crime, and
law. It also implies a misapplicationof the notion of plurallegal spheresto acts
thatin our view properlybelong to the field of crime and thatarerecognizedas
such even by thefavelados who live wherethey occur.To describethem as acts
of alternativejustice both exoticizes and depoliticizes them by removingthem
from the sphereof a failed nationaljustice system.
OurresearchindicatesthatBraziliansof all classes generallythink that it is
too risky to takejustice into theirown handsbecause doing so may lead to a lot
of trouble,especially from a moralpoint of view. However,they are more likely to arguethatif this kind of justice were carriedout by the "right"institution,
such as a police force that defends innocent people, it would constitutean effective solution to crime. This type of reasoningleads people from all classes
to supportsummaryexecutions by the police, and to evaluate police violence
positively. It is in this context thatROTAandthe Esquadraoda Mortearewidely admired.Poor people see these two organizationsas "tough"with criminals
andnot corrupt.Moreover,they tendto believe thatthese organizationskill "the
rightpeople"-even againstmuchevidence to the contrary-and thereforecarry out justice. Poor people perceive these organizationsas more efficient than
a justice system in which the death penalty does not exist and the judicial
process takes forever.This same reasoningleads them to admireand occasionally use vigilantes, calledjusticeiros (literally,justice makers).In an interview
conductedby Caldeira,a young man who lives in a working-classneighborhood said the following abouta famous police death squad:
I wishtheEsquadrao
daMortestillexisted.TheEsquadrao
daMorteis thepolicethat
daMorteis justicedoneby one'sownhands.I thinkit should
onlykills;theEsquadrao
stillexist.It'snecessaryto takejusticeintoone'sownhands,butthepeoplewhoshould
do thisshouldbe thepolice,theauthorities
themselves,notus.Whyshouldwe geta guy
and kill him? What do we pay taxes for? For this, to be protected.... It's not worth it

forusto lynch.Theyshouldhavetheright,theyhavetheduty,becausewe paytaxesfor

this.... Thelawmustbe thisone:if youkill,youdie.
People from the upperclasses also defend summaryexecutions, andmay use
exactly the same argumentsaboutthe failureof thejustice system and the need
to "kill the right person,"to "solve the problem definitively."However, in a
country with huge social inequalities such as Brazil, the way in which justice
does not work for the upperclasses is not the same as for the workingclasses.
In fact, for the upperclasses, the non-functioningof the justice system may be
just anotherprivilege.24In contrastto the workingclasses-who are frequent
victims of police violence, constantlyrunthe risk of being mistakenfor crimi24

See Holston (1991) for an analysis of this perverse"privilege-at-law."

714

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

nals, and suffer accordingly-the upper classes are rarely victims of police
abuse or of the justice system. Rather,they have the luxuryof choosing to disrespect the law. They can rely on their perceptionthat the law does not work,
or works only for them, and they have the privilege of bypassing or manipulating it. As an upper-classwoman told Caldeirawith considerableirony during an interview:
Normally,lawsareenforcedagainstthelowerclasses,theclassesof smallpurchasing
power.Forthem,thelawsarewell "respected."
Theymakethemfollowthelaw,obey
the law.Wefromthe middleclass,fromthe upperclass,we don'tneedto respectthe
lawbecausewe payforit withmoney.I don'tthinkthisis just.
Thus, the everydayexperienceof violence andof the institutionsof law leads
to a pervasive and comprehensivedelegitimationof the rule of law among all
social classes. Poorerpeople are victims of arbitrariness,violence, and injustices committedby agents of the law. As a result, they feel that they are left
without alternativesinside the law. In contrast,the rich find it in their best interestto take advantageof the failuresof legal institutions.They have the privilege of being able to choose to ignore the law and do what they think is more
appropriate.What is similar for both groups, however, is that their reactions
tend to be framedin privateand frequentlyillegal terms.In both cases, the rule
of law is discredited.
If we considerthe performanceof the police and thejustice system in a context of growing ratesof violent crime, it is not difficultto understandwhy Sao
Paulo'sresidentsincreasinglyadoptprivatemeasuresto protectthemselvesand
deal with violence. Because these measuresare privateand often violent, they
can only contributefurtherto the delegitimationof the rule of law and the reproductionof violence. Moreover,given the structureof inequalitythatcharacterizesBraziliansociety,suchprivatemeasuresemphasizesocial discrimination.
Private measuresto deal with crime and to carryout justice are of various
sorts. The most visible measuresare the ubiquitouswalls and bars that people
put up in frontof theirhouses and apartmentbuildings.These barriersare dramatically changing the landscapeof Brazilianbig cities, as well as the social
interactionsof their public spaces. In these cities of walls, residents are very
suspicious and change their habits to avoid interactionsin public, especially
with people perceivedas being different.The walls not only separateresidences
but also createsemi-publicenclaves, such as shoppingcentersand office complexes, where entrancescan be controlledand social homogeneityguaranteed.
In this sense, fear of crime legitimates practicesof segregationand considerably changes the characterof public space. In a society wherepeople from differentsocial groupstend not to interactor even encountereach otherin public,
the chances for propagatingdemocraticpracticesare surely diminished.25
The walls do not stand alone. They are part of a complex of measures,in25
See Caldeira(1999) for a discussion of how the proliferationof fortifiedenclaves transforms
the characterof public life, both in Sao Paulo and in Los Angeles.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

715

cluding various technologies of security,from video camerasto identification

of visitors, from electronic fences to all kinds of alarms.These measuresalso
includethe hiringof doormenandarmedprivateguards.As in manyothercountries, such practiceshelp multiplythe profitsof the rapidlygrowing privatesecurity industry.This industryhas various faces in Brazil, as it adaptsto serve
differentsocial classes. At some levels it mixes with the illegal actions of the
police, as a significant numberof privateguardsare off-duty policemen, frequentlyworkingwith police guns. For the workingclasses, however, these organizedprivateservices andtechnologiesaremostly unaffordable.At times, the
poor may benefit from the vigilance of justiceiros hired by a local merchant,
but they are also just as likely to suffer the adverseconsequences.
Collectively, these measurescause deep transformationsin the way people
carryon theirdaily affairs,interactwith others,and move aroundthe city. It is
possible to observe new gestures, new body postures,and new instinctive reactionsof suspicionanddistancing.We mightcall this a new cultureof fear,using an expressionthat was used for many years to refer to everyday life under
authoritarianregimes. In the presentBraziliancase, however,thereis no political repression,and police violence is routineand uncensorednews. Thus, this
new cultureof fear takes shape in the context of a democraticpolitical system
with a free political organizationand press. Its developmentis contemporary
with the transitionto democracy.
A common argumentin discussions aboutviolence and democraticconsolidation in Brazil is that the concurrenceof elements we have described including the abuses of the police, the delegitimationof the justice system, the
privatemeasuresof justice and security,the generalizeddisrespectfor law, the
cultureof fear, and the relatedtransformationsof public space-constitutes an
authoritarianenclave or a survivalof authoritarian
rule in democracy.Although
this concurrencecertainlyrepresentsan obstacle to the consolidationof a democratic society in Brazil, we want to call attentionto its novelty as a postdemocratizationphenomenon,and cautionagainstblamingthe (military)past.
The contemporaryviolence is new not only because of its unprecedentedlevel, but also because of its publicness, disregardfor ideological rhetoric,negation of the notions of a common andjust futurefor all Brazilians,and its perverse associationwith the expansionof political citizenshipandrelatednotions
of agency. The contemporaryviolence is not an invention of authoritarian
rulers,but the perversedevelopmentof a deeply unequalsociety in which the
expansion of some rights occurredsimultaneouslywith the abandonmentof
modernideas of developmentand progressby many Braziliansof all classes,
underminingthe sense of a common project for the future (see Caldeira,in
press, Chapter2).
DISJUNCTIVE

DEMOCRACY

The coexistence in Brazil of political democracyand violence againstcitizens

exemplifies a particularbut common kind of democraticdisjunctionin which

716

TERESA

P. R. CALDEIRA

AND

JAMES

HOLSTON

the civil component of citizenship is systematicallyunderminedand significantly weaker than other components.Thus, Braziliansvote in generally free
and fairelections, have theirbasic rightsembodiedas constitutionalprinciples,
and benefit from a minimumof socio-economic rights, well-groundedin public demands. However, the vast majority cannot rely on the institutions of
state-particularly on the courts and the police-to respect or guaranteetheir
individual rights, arbitratetheir conflicts justly, or stem escalating violence
legally. In this sense, the killing of "marginals"describedearlieris an extreme
expression in Braziliansociety of the everyday marginalizationof the anonymous individualas citizen and bearerof civil rights. This kind of disjunctive
democratizationinevitably brings new forms of violence and injustice.These
are forms specific to a democracywith a discreditedcivil sphere.As we have
shown, when the civil componentis discredited,social groupsat all levels come
to supportthe privatizationof bothjustice and security,and illegal or extralegal measuresof controlby state institutions,particularlythe police.
In effect, the developmentof Braziliancitizenship underpolitical democracy has been very uneven, in a numberof significant ways. On the one hand,
its civil component-including, civil rights,access to justice, due process, and
the applicationof law-is unevenly and irregularlydistributedamong Brazilian citizens. On the other,althoughsystematicallyviolated, the civil sphereof
citizenship has not been a prominentconcern for many of the principalforces
of democratizationin Brazil. These forces include the new social movements,
labor unions, political parties,and universities,which remainfocused on other aspects of citizenship and democracy, especially the political and socioeconomic. Humanrights organizations,the women's movement, and various
minority-rightsgroups(such as those for gays andblacks) do defend civil concerns. But they do not defend these as common rights of all citizens, of every
man andwoman,rich andpoor,regardlessof raceor sexual orientation.Rather,
Braziliansvery often perceive these groups as defending the rights of minorities, "marginals,"and special interests-to such an extent, for example, that
many oppose humanrights as "privilegesfor bandits,"as we discussed earlier. Thus, althoughpeople complain bitterly about everyday violence and injustice, Braziliandemocracylargely ignores the civil aspect of citizenship as
a fundamentalconcern for all Brazilians.To use Marshall's(1977) typology,
this disjunctionmeans that in comparisonwith social and political rights, civil rights have not been effectively woven into the fabric of Brazilian citizenship.
It is particularlydifficultto analyzewhy civil rightsare so impracticableand
disregardedin Brazil withoutbeing historicallyreductiveor culturallysimplistic. It requiresunderstandingan intersectionof culturalformulationsaboutlaw,
citizenship,andindividualautonomythatis too complex to explorehere.As we
statedearlier,our concernin this essay is not to establishunderlyinghistorical
causalities. Rather,our focus is on both the lived and theoreticalsignificance

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

717

of the disjunctionin Brazilianpolitical democracywherecivil rightsarenot experienced,perceived, or appreciatedas common rights of citizenship.
How can we accounttheoreticallyfor a democracyin which the civil componentof citizenshipis systematicallyviolated?Whatsense does it maketo call this
Brazil a democracy?It only does so if we recognizethatthese combinationsof
contradictorydevelopmentsreveal a fundamentalcharacteristicof democratization itself-namely, that it is normallydisjunctive.By calling democracydisjunctive, we want to emphasizethatit comprisesprocesses in the institutionalization, practice, and meaning of citizenship that are never uniform or
homogeneous.Rather,they are normallyuneven, unbalanced,irregular,heterogeneous, arrhythmic,and indeed contradictory.The concept of disjunctive
democracystresses,therefore,thatat any one momentcitizenshipmay expandin
one areaof rightsas it contractsin another.The conceptalso meansthatdemocracy's distributionand depthamong a populationof citizens in a given political
space are uneven.It is in this lack of balanceand unevennessthatcontemporary
Brazilexemplifiesa disjunctiontypicalof manyemergingdemocracies.26
The notion of disjunctionwe suggest is differentfrom, but complementary
to severalotherconsiderationsof temporaland spatialissues in the studyof democratization.For example, in the work of BarringtonMoore (1966), Eric
Nordlinger(1971), and LeonardBinder et al. (1971), argumentsabout timing
focus on the sequences of various crises of nationalidentity, state formation,
modernization,and social structureas historical prerequisitesof democracy.
Centralto otherstudies of political developmentis the problemof the penetration of the stateas a centralauthoritythroughouta society and territory(for example,JosephLaPalombara'scontributionto Binderet al.). GuillermoO'Donnell and Philippe Schmitter's(1986) seminal work considersregime transition
as an unfolding historicalprocess of analyticallydistinct sequences, patterns,
and stages. In addition,dependencytheoryanalyzesthe importanceof both the
timingandthe spatialreferentsof a region'sinsertioninto the internationalmarket as factorsthatrelatecapitalistdevelopmentanddemocratizationon a world
scale. All of these studiesconsidertime and space as dimensionsof change between differentkinds of political regimes, systems, or stages, which they treat
as more or less comprehensivewholes.
By contrast,our notion of disjunctionis specifically internalto democracy.
It emphasizesthatdemocracyentails a complexityof processes andinstitutions
of citizenship, always becoming and unbecoming, often confusing and unstable, ratherthan a set stage of institutions,actors, social structures,and culturalvalues. The notionof disjunctionsuggests thatalthoughtherearerulesthat
might define an ideal democracy,such a regime has neverexisted. Rather,there
are always various and often contradictorysets of rules and games in play in
any democracy,at the same time and in the same space, some of which may be
26

The concept of disjunctivedemocracyis developed furtherin Holston (forthcoming).

718

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

considerablyless democraticthanothers.This theoryof disjunctionhas a number of conceptualrequisites, some of which we now consider in terms of the
extension of democracyand citizenship beyond the political, the rule of law,
and the questionof alternativedemocraticformations.
The Extension of Democracy beyond the Political

The theory of disjunctivedemocracydepends on one's thinkingof both modern democracyand citizenshipas extendingbeyond the political to encompass
the social, economic, and cultural spheres of life. Given the coincidence of
democracyandviolence in new democraciessuch as Brazil,it seems evident to
us thatit is only as a fraudor as a severely impairedregime thatdemocracycan
standalone-either conceptuallyor actually-as a kind of political methodin
the context of social, economic, and culturalconditions hostile to democratic
citizenship. If thatis so, then democracyhas to be consideredas much a qualification of society as of politics. This is not to adopt a maximal definition of
democracy,or to make a particularpolitical cultureits precondition.Rather,it
is to insist thatthe extension of democracyto the social sphereis as centralto
the concept as its extension to the political sphere.It is to contendthatboth of
these colonizations-of society by the state and the state by society-constitute the contemporaryand enabling form of democraticdevelopment.Our argumentderivesfromBobbio's (1989) observationsaboutthe modernextension
of democracyfrom the political to the social. With the criticisms alreadysuggested, it also develops Marshall's(1977) analysis of the intersectionof class
and citizenshipandthe relationsbetween whathe defined as the threeelements
of citizenship:namely,the political,civil, andsocioeconomic.Most modernliteratureon citizenship alreadyincorporatesthis broaderconception of citizenship.27 But, curiously, the literature on democratizationmostly considers
democracy in terms of political rights and related electoral institutions and
processes, as if the other arenasof citizenship were disconnectedfrom them.
In our view, democracyand citizenship are necessarily and inherentlyconnected in a muchfuller sense-for the democraticstateexists only throughthe
political participationof citizens in both its organizationand its articulation
with civil society, and citizenshipexists only throughthe state's applicationof
the principleof legality and its defense of the equality,liberty,and dignity of
citizens in both their public and privatelives. This inherentconnection means
that democracyneeds to be evaluatedin termsof the realizationof citizenship
in the full sense of the term. By that, we mean not just political citizenshipi.e., electoral performanceand enabling institutionsof government-but also
citizenship's substantivesocial, cultural,and economic conditions.
27 Some recentwork on
citizenshipconsidersits extension to the culturalas well. For example,
see Taylor(1992) and Kymlicka (1997). In Canadaand throughoutLatinAmerica-especially in
the Andean countries and Mexico-many questions of indigenous rights, sovereignty,and legal
pluralismturnon issues of culturalcitizenship.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

719

To some, the argumentthatpolitical democracyis not enough both in theory andin fact may seem obvious. Indeed,one of the most astuteanalystsof contemporarydemocracy,GuillermoO'Donnell (1992:49), writes thatdemocratic
consolidationrequires"theextension of similarlydemocratic... relationsinto
other[notjust political] spheresof social life." However,we would suggest that
the means and modes of this extension (and frequentretraction),andtheirconsequences for the theory of democracy,remain inadequatelyinvestigated or
conceptualized. This is especially the case because most contemporaryobserversuse a political definitionof democracythatneglects to considerthatthe
social conditions of citizenship are constitutive of its political possibilities.
Such a definition excludes, therefore, democracy's extension to the social
sphereout of hand.Furthermore,we arguethatthe democratizationof stateand
society are mutuallydefining in a consolidateddemocracy;thatis, democratic
extensionsbetween stateand society arereciprocal.In an importantsense, consolidationmeansthatthe statedoes not monopolize,in fact or claim, the sources
of democracy-or of citizenshipor law, for thatmatter.
The dissociationof democracyfrom society and its conditionsof citizenship
derives directly from the continuedreign-at least in Americanpolitical science-of Joseph Schumpeter'sdefinitionof democracyover the study of both
new and establisheddemocracies:"Thedemocraticmethodis that institutional arrangementfor arrivingat political decisions in which individualsacquire
the power to decide by means of a competitive struggle for the people's
vote"(1947:269).28Our intentionhere is not to become embroiledin debates
about competing definitions of "democratic-ness"(see Karl 1990 for an enlightening discussion). As Bobbio has wryly remarked,"Everyregime is democraticaccordingto the meaning of democracypresumedby its defendants
and undemocraticin the sense upheldby its detractors"(1989:158). Rather,our
intentionis to suggest why the studyof the civil componentof citizenship,with
its attributesof law andjustice, does not seem to fit the predominantconceptualizationof democracy;or, put in anotherway, to suggest that currentapproachesgenerallyfail to relatetheir conceptionsof democracyto the real extent and exercise of citizenship and that this dissociationis a serious problem.
The Schumpeteriandefinitionprivileges political democracyand the proceduralminimums necessary to achieve it. When interpretednarrowly,it holds
that democracyis fundamentallya means of governance,a method or technical instrumentof politics, and democratization,therefore,is primarilya question of establishingadequategovernmentalinstitutions.In this minimalistversion, the modernstate is the locus of the democraticproject.Thus, the narrow
28 As Samuel
Huntington(1991:6-7) observes, "By the 1970s the debatewas over, and Schumpeterhad won. Theorists... concludedthatonly [Schumpeter'sproceduraldefinitionof democracy] providedthe analyticalprecision and empiricalreferentsthat make the concept a useful one.
Sweepingdiscussions of democracyin termsof normativetheorysharplydeclined,at least inAmerican scholarlydiscussions."

720

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

political definition of democracyfocuses on elections and therebyeliminates

all concernfor the exercise of anythingotherthansome aspectsof political citizenship. For the same reason,even thoughthe political definitionemphasizes
institutions,it does not consider the effectiveness of the justice system or the
qualityof the rule of law, even in the latter'smorerestrictiveform as state law.
Karl(1986) has convincingly criticizedthis view as "electoralism"and argued
that Schumpeterhimself would not have supportedit, because he considered
civil rights necessaryfor the operationof democracy.
Many currentstudies agree with Karland go beyond mere electoralcompetition to include in their definitionsof democracyguaranteesfor civil liberties
throughthe rule of law, as well as two forms of accountability:namely,that of
the governorsto the governedand of the militaryto the civilian. For example,
O'Donnell and Schmitter(1986:8) defined a broaderdemocraticminimumas
"secretballoting, universaladult suffrage,regularelections, partisancompetition, associationalrecognitionand access, and executive accountability."Similarly,for her "middle-rangespecification"(thatis, neitherminimalnor maximal), Karl(1990:2) identifies contestation,participation,and accountabilityof
rulers,and adds civilian controlover the military.Curiously,neitherdefinition
specifies the rule of law or civil rights.Although accountability,participation,
and elections presume fundamentalcivil rights, there are many other rights,
such as the rightto justice, thatare not so obviously included.Moreover,it is a
mistake in our view to presumea democraticrule of law when the conditions
of political democracy have been met. Furthermore,although this broader
Schumpeterianperspectiveclaims to includethe reachof the stateinto civil society, the extension has to remainof secondaryimportance.It is so limited because it is groundedin the classic dichotomy between state and civil society
(andassociateddivides, such as public andprivate)thatlocates the realmof the
political in the formerand distinguishesthatrealm from the social relationsof
the latter.Thus, if democracyis narrowlyconceived as a political method,this
dichotomizationmeans thatit cannotalso be a conditionof society.
The deeperdissociationof democracyfrom society producedby this political definitionmeans thatmany studies-including those by governmentaland
funding agencies-examine only the political and formalcomponentsof citizenship. Whenjustified in the literature,the principalreason given for avoiding the other, more substantiveaspects is commonly the claim that to study
them-and thereforethe real textureof social life-would be to open the door
to ideological and evidentiaryconfusion.Thus, the social complexity in which
every democraticgovernmentmust survive and with which it must come to
termsto act democraticallyfalls outsidethe definitionalscope. To make, as we
suggest, the legal, ethical, and performativedimensions of citizenship fundamentalto the conceptionof democracyseems to make it difficult for many observers"to find any actualdemocraciesto study"(Karl 1990:2) or "to assign a
reasonable closure to the second transitionprocess" (Valenzuela 1992:60),

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

721

because in these terms no democracyis really consolidated.But this apparent

difficulty is an artifactof a classificatoryscheme that insists on homogeneous
categories and terminalprocesses. If we accept thateven consolidateddemocracies are disjunctivethen the difficulty evaporates,and we are compelled to
study the full experience of democraticcitizenship to understandthe development of democracy.
This full experience is so unbalancedand heterogeneousin so many contemporarydemocraciesthatits traditionalpolitical definitionin termsof membershipin the nation-stateis as unconvincingtheoreticallyas it is unfaithfulto
the new empiricalconditions.Divided from social considerations,the political
definitiongenerallytreatscitizenshipin termsof abstractand uniformrightsof
membershipin the nation-state.This treatmentassumesan even distributionof
these rights across nationalspace and society. However, we know that actual
democraciesbehave very differently.We know that there are vast, substantive
differencesof citizenshipbetween social groupsandregions at subnationaland
transnationallevels, even when participatorypolitical rights are nationallyeffective. If formalcitizenshiprefersto membershipin the territorialnation-state,
and substantivecitizenshipto the arrayof political, civil, socio-economic, and
culturalrights people possess and exercise, then much of the turmoilin contemporarydemocracies(both emerging and established)derives from the disjunctive relationbetween the formaland the substantive.29As many observers
arebeginningto realize,the conditionof formalmembershipwithoutmuchsubstantivecitizenship is characteristicof many of the societies that have experienced recent transitionsto democracyin LatinAmerica,Asia, EasternEurope,
andAfrica.
Focusing on such characteristics,O'Donnell (1993) makes a similar argument in his recent discussions of "low-intensitycitizenship"in the "brownareas of new democracies."Moreover,he stresses the importanceof an effective
rule of law andthe state's legal responsibilityto extend and enforce citizenship
rightsuniversally.The failureof the stateto do so curtailscitizenshipand compromises democracy.Althoughhe maintainsa political definitionof democracy, O'Donnell's analysis is close to our own understandingof democracy'sdisjunctive natureand the importanceof studyingit from this perspective.As he
argues"Evena political definitionof democracy(such as thatrecommendedby
most contemporaryauthors,and to which I adhere)should not neglect posing
the question of the extent to which citizenship is really exercised in a given
country"(1993:1361). We would only disagreewith his limitingthe concept of
low-intensity citizenship "specifically to the political sphere, to the political
theory of political democracy"-a limitationthat leads him to the dubiousargumentthat"thedenial of liberalrightsto (mostly but not exclusively) the poor
29 For further
analysis of the disjunctionbetween the formaland substantiveaspects of citizenship, see Holston andAppadurai(1999).

722

TERESA P. R. CALDEIRA AND JAMES HOLSTON

or otherwisedeprivedsectors is analyticallydistinctfrom, and bearsno necessary relation to various degrees of social and economic democratization"
(1993:1361). Yet,as O'Donnell admitsin the same discussion,the two "gohand
in hand"empirically.Hence, the argumentof "no necessary relation"seems
drivenmoreby the need to maintainthe political definitionthanthe need to account betterfor the natureof real democracies.
If we agree that contemporarydemocracies are significantly disjunctivein
the political, civil, social, andculturalaspects of citizenship,and that such disjunctionscan delegitimatepolitical democracy,then it seems rightto insist that
the extension of democracyto the social sphereis as centralto the concept of
democracyas its extension to the political. It seems rightto arguethatthese extensions are reciprocaland mutuallydefining. It then becomes unnecessaryto
arguethatdemocracy'sdisjunctionsare incidentalor extraneousto the theory.
The Democratic Rule of Law

One of the consequencesof maintainingthe political definitionof democracy

is that the meaning of the rule of law is seldom investigatedin studies of democratization,and the crucial question of the performanceof the justice system seldom posed. Rather,it is usually assumedthatbecause the rule of law is
a fundamentof democracy,the institutionalizationof political democracywill
produce a rule of law that is inherentlydemocratic.30However, the study of
electoraldemocracieslike Brazil thatare also uncivil indicatesthatthe relation
betweendemocracyandthe rule of law is far more uncertain.It shows the need
to assess, ratherthan assume, this relation.In Brazil, there is no doubt that violence againstcitizens and disrespectfor law have increasedafter the formal
transitionto democracy.
When we analyze the civil sphereof citizenshipin termsof its legal, moral,
and performativeattributes,it becomes evident thatthe rights,institutions,and
practiceswhich give it substancerequiremore thanformallegislation or an independentlegislatureto become effective. Justas the institutionalizationof democraticpolitical rights does not by itself secure the integrityor development
of a democracy'scivil sphere,our analysisof violence in Brazil shows thatcivil rights do not depend only on legislaturesto draftlaws. More than executive
and legislative initiatives, civil rights depend on the justice system to secure
their realization.Although Marshallidentifiedthe importanceof the courts in
this regard,the justice system entails many more elements thanhe recognized,
all of which arecrucial to the realizationof the civil sphereof citizenship.The
justice system includes the courts with their bureaucracyand administration,
30 Thus, it is
very generallypresumedthatpolitical democraciesguaranteea greaterrespectfor
law andhumandignitythantheirauthoritarian
predecessors.Indeed,in manyLatinAmericancountries,this suppositiongroundedpoliticalargumentsfor the replacementof dictatorshipwith democracy.When political democracyfinally came, it was thus burdenedwith an expectationin this regard that it could not meet in many cases. As a result, many inauguraldemocraticgovernmentsat
the nationaland local levels suffereddisappointmentsand loss of prestige.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

723

the bar,law schools, and the police (even though some of its divisions may be
partof the executive branchof government),all organizedin termsof the rule
of law. Basic to these elements is the rightto justice, the rightto all otherrights.
Each in its own way, these various agents of the justice system establish a reciprocitybetween the power of law at their disposal and the capacity of people
and institutionsto act accordingto expectationsaboutthe rule of law and their
rightto justice.
In emphasizingthis reciprocity,however,we do not mean to say thatthe rule
of law as a scheme of rules, procedures,principles,and institutionsnecessarily legitimatesor securesdemocracy.To the contrary,uncivil democracydemonstratesthat the relationbetween rule of law and democracyis not a given. It
mustbe assessed in specific cases, not assumedin general.This conclusionsuggests three correlates.One is that, as we have seen, political democratization
does not necessarily producea rule of law centeredon democraticconsiderations. The second is that the rule of law is not necessarily democratic.It may
be necessary for full democracy,but it is not exclusive to it. We have only to
think of England's constitutionalmonarchy,France'sphysiocraticdespotism,
Nazi Germany'slegally constitutedgovernment,and South African apartheid
to realize that the rule of law can coexist with non-democraticpolitical forms,
is not necessarilyjust, and can be focused on non-democraticconcerns. Both
of these correlationscontradictcommonidealizationsaboutdemocracyandthe
rule of law. The thirdcorrelateis thatthe kind of rule of law that is a prerequisite for democracyneeds to be gauged, as we suggest below, to the concernsof
democraticcitizenship.
There are a numberof ways to understandthe rule of law that are useful in
sortingout its relationto democracy.One can view it as an amalgamationover
time of essential characteristics,as do many historiansof law. If we examine
the nine hundred-yearGermantraditionof Rechtsstaat,it becomes evident that
the ancientconcept of a "governmentof law and not of men"embodyingthese
characteristicsdoes not have a democraticorigin. It also becomes clear thatthe
rule of law correlatesin specific ways with various political forms, though it
develops most often with democracy(see Skinner 1989 for this correlation).
However,one can also considersuch an essentializedspecificationof the rule
of law impossible, and even undesirablein both a theoretical and practical
sense. One can try insteadto give it an abstracttheoreticalaccount,as do legal
theoristsRonald Dworkin and Cass R. Sunstein. From opposed perspectives,
they both analyze the natureof the legal, and the reasoningpeculiarto its rule.
Thus Dworkin writes:
Thelawof a community
on thisaccountis theschemeof rightsandresponsibilities
that
meet[thefollowing]complexstandard:
theylicensecoercionbecausetheyflow from
past[political]decisionsof therightsort.Theyaretherefore"legal"rightsandresponsibilities.Thischaracterization
of theconceptof lawsetsout,in suitablyairyform,what
is sometimescalledthe"rule"of law(1986:93).

724

TERESA

P. R. CALDEIRA

AND

JAMES

HOLSTON

With very differenttheoreticalintentions,Sunstein (1996:12) understandsthe

rule of law as a combinationof rules that limit "untrammeleddiscretion"and
forms of discretionthat limit rule-boundjustice so as to "makespace for particularisticforms of argument."
Despite otherdifferences,both DworkinandSunsteindrawan importantdistinctionbetween law andjustice in characterizingthe rule of law: a rule of law
can be legal but not particularlyjust. Dworkin(1986:97-98) proposesthatlaw
establishes which rightsjustify using or withholdingthe coercive force of the
state, while justice establishesthe best theory or standardof rights.Thus, justice is a matterof a moralor political theoryaboutthe rightfulmeasuresof human conduct.AlthoughDworkinviews theoriesof justice as "imposedby personal convictions," we suggest that democracy offers a more compelling
considerationof justice, if not a theory.Sunsteinarguesthat
it cannotbe saidthata systemcomplyingwiththeruleof lawmustbejust... Totake
in SouthAfricawas
oneexample,thefundamental
problemwiththesystemof apartheid
thebasicfeaturesof
notthatit violatedtheruleof law.Onthecontrary,
emphatically
couldbe madeentirelyconsistentwiththeruleof law.Rulesdo notguarantee
apartheid
justice(1996:193).
If we cannotassumethata rule of law is democratic,or thatpolitical democracies have a democraticrule of law, then we have to investigatethe extent to
which a particularrule of law engages and realizes a project for democracy.
Such a rule of law is necessaryfor full democraticcitizenship,on which the legitimacy of democracyas both a political and a social project ultimately depends. This necessity is easy to show. We can imagine fair trialsoccurringunder non-democraticregimes and unfairtrailsunderdemocraticones. However,
we cannot imagine anythingother than a sham democracywithout fair trials.
Therefore,a democracymust secure the legitimacy of law on its own termsof
citizenship.If not, it becomes discredited.
What,then, does a democraticrule of law entail and how does a democracy
secure it? These are immensely importantand complex questionsthat we cannot adequatelyaddresshere. Basically, we agree with Sunstein'sproposalthat
the best way to securejustice befitting a democracyis not to suppose that the
rule of law is intrinsicallyjust or democratic,but to referenceit to criteriaof
justice forged in the democraticarenasof society. In askinghow a legal system
supportsa well-functioningdemocracyand vice-versa we want to know how
the system responds to considerationsthat favor democratic outcomes and
processes. One way to learnthis is by focusing on how a particularrule of law
realizesthe values of justice thataregermaneto democraticcitizenship.We can
referencewhat goes on in the legal system to such criteriaand evaluateits performance.We want to propose three kinds of considerationfor furtherstudy.
First,we can determinewhetherthe exercise of judicial and police authorityis
justifiablein termsof the powers and limits establishedin a democraticconstitution.Second, we can analyze the performanceof a legal system in relationto

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

725

those aspectsof a particulardemocraticdevelopmentthatareweakest and least

reliable. For example, we can evaluate how the judiciary reacts when fundamental civil rights are at stake, or when a disadvantagedminorityor a politically vulnerablegroupis at risk. Finally,we can evaluatea rule of law in terms
of the degree to which citizens participatein thejustice system, and to which it
is accountableto their oversight.The latterhas been analyzedmostly in terms
of civilian oversight of the police (see Chevigny, 1995). Accountabilityof the
courts to citizens has not received comparablestudy,perhapsbecause it is often thoughtthat adjudicationought to be protectedfrom democraticinfluence.
But the problemsof uncivil democracyindicate that a measureof citizen participation in the judicial process and in the oversight of judges would both
broadenaccess to justice and encouragethe judiciaryto respondmore directly
to democraticchange. The means of such citizen participationinclude the jury
system, alternativedisputeresolution,thejudicial appointmentprocess,andthe
externalreview, election, and recall of judges.
Citizens of an electoraldemocracywith a democraticrule of law participate
not only in free elections and variousforms of political association,but also in
a justice system in which they are confidentof fair and equitabletreatment,to
which they have reasonableaccess, to which all are liable, which is accountable to theiroversightin significantways, and which regulatesaccordingto the
due process of law not only theirpracticesbut also those of the stateas a legally constitutedand accountableagent of society. These five sorts of considerations-fairness, access, universality,accountability,andlegality-characterize
a democraticprojectfor a rule of law. That a rule of law has never existed in
such perfectiondoes not lessen the importanceof an approximation.That importancebecomes apparentas soon as people perceive thattheirrightto justice
lacks institutionalconsolidation,above all in the courts.As we have seen, citizens of an electoral democracy without a democraticrule of law find themselves in a disastrouschain reaction:the justice system and those who defend
it become discredited,impunityand violence prevail, and largely as a result, a
cultureof vigilantism,exceptionalism,andprivatizedpowerpredominates.Uncivil democracyresults.
Other Histories, Other Democracies

If democracywith mass citizenship has been one of the hallmarksof Western

modernity,its conceptualizationhas rarelyconsideredviolence amongcitizens
as a characteristic,ratherthan episodic condition of its development.Instead,
democratictheoryin the West has generallyproceededas if the problemof internalviolence had alreadybeen solved. Thatis, it assumes the pacificationof
societal violence throughthe development of a civilizing process in Norbert
Elias's sense, the moderndisciplines in Foucault'ssense, or the modernnationstate that combines rule of law with the monopolizationof the means of coercion, in Max Weber'ssense, and uses these developmentsto control violence.

726

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

Democratic theory has thus assumed that the problem of internal violence
againstcitizens has been resolved-or, better,dissolved-by assigningit to the
Westerndevelopmentof cultureand the nation-state.
However, studying the new democracies like Brazil reveals the extent to
which this supposedresolutionof violence is an expression-even a charadeof the particularhistories of a few nation-states,ratherthan a general or universal traitof democratizationitself. In the new non-Westerndemocratization,
violence remainsentangledwith democracy.This entanglementis distinctive
because people both engage and oppose democratizationin the idioms of violent social relationsthatremainconstitutiveof societies with differentcultures
and histories. Many are like Brazil in that they have political democraciesin
which citizens suffersystematicviolence from forces of public andprivate,organized and unorganizedcoercion, which act with the confidence of impunity.
In these democracies,the citizen's body is not boundedand protectedby civil
rights. Rather,it is vulnerableto this violence and trappedin the reproduction
of it that privatizedjustice promotes.Culturesof fear, indifference,illegality,
andabuseof powercharacteristicallyproliferatewithinthe publicbody of these
political democracies,coexisting with democraticvalues of public life. In this
political democratization,urbanpublic space becomes both newly politicized
and newly dangerous,is alternativelyabandonedand fortified.In these political democracies,the rule of law becomes both brutallyviolent and ineffective
to combatviolence.
At this point,thereare few studiesof the new democraciesin these more anthropologicalterms.FernandoCoronilandJulie Skurski(1991) show how political violence in Venezuelais regularlyreenactedin democraticcontexts.They
arguethat violence is "wielded and resisted"(289) in the terms of a society's
distinctivehistory,in relationto which, therefore,it has to be analyzed. Contemporaryviolence in Venezuela,they suggest, continuesto be framed"inConquest terms,"mobilizing notions of a barbaricpeople and a civilizing government(of elites). Theirstudyis an exampleof anothertype of cultureandhistory
in which modernizationand democraticpolitics have always been entangled
with violence. Taussig (1987) demonstratesa similarprocess for Colombia in
his study of the use of violence in the rubberboom and the creationof what he
calls a "cultureof terrorand space of death"(1987:3). His analysis also shows
how the internalpacificationof Europeanstatesandthe creationof a cultureof
fear in LatinAmericacoincided.
These studiessuggestthatas morecountriesdemocratize,anddemocracybecomes more diverse, the specificities of its Europeanand NorthAmericanexperience-including its relationto internalviolence-become more apparent.
In this context, it seems unlikely thatpolitical theoriesof democracyanchored
in these specificities of Westernhistory and culture are adequatefor understandingdemocracy'sglobal reach and its non-Westernexperience.But if that
is the case, how do we assess the quality of democracyin such diverse situa-

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

727

tions?Are there,fundamentally,culturallydifferentways of being democratic?

Are there, in other words, alternativeconfigurationsof democracyand different ways of reachingit? Or areclaims of real differencemerely excuses for undemocraticpractices?To what extent can democracybe culturallyrelativized
before it becomes unrecognizable?
How do we establish the limits of what constitutesan alternativeformation
of democracyandnot a differentorderof being?This is, admittedly,a complex
anddifficultproject.But thatdifficulty shouldnot blind us to the differencesof
history with regardto democracy.It should not make the minimalpolitical definition an attractiveresponse to the increasingdiversity of democracy.If that
definitionhas the advantageof being parsimonious-and to that extent seemingly universal-we hope we have shown thatit is superficialbecause it misses many of the principaldilemmas of contemporarydemocratization.At this
point, however, there are too few studies of the social and culturalconditions
of citizenshipin the new political democraciesto answerthe kinds of questions
just posed. We have suggested thatthinkingaboutdemocracyas disjunctiveis
useful in studyingthese conditionsbecause the premiseof this approachis that
although all contemporarydemocracies are markedby disjunctions,they are
not necessarilyor even likely to be disjunctivein the same ways. This approach
dependson the idea thatdifferentsocieties and culturesmust, by force of their
differenthistories, producedifferencesin democracy.Therefore,there is not a
single model, recipe, history, or cultureof democracy.Although we have not
discussed the disjunctionsof Americandemocracy,for example, they are not
difficult to specify in terms of the weakness of socio-economic rights for all
Americans,the violent disrespectof civil rightsfor some, andthe excessive professionalizationof politics.
Specifying the disjunctive qualities of a democracy-describing Brazilian
democracy as uncivil, for example-refers to and criticizes only specific aspects of what is normallya complex democraticproject.It does not condemn
the entireproject.Moreover,it does not supposethatbecomingcivil meansnecessarily becoming just like some other democracy.Uncivil democracies are
democracies nevertheless, with complex disjunctionsthat need to be understood as inherentto democraticdevelopment. The problem is to account for
their disjunctionsfrom within the process of democratizationwithout disrespecting their democratic intentions, or predeterminingthe antidote to their
problemson the basis of convergence to ideal types that are modeled on particular,and usuallyWesternexamples. If, as we think,such convergenceis unlikely, then democratictheory must adaptto the developmentof new species.
REFERENCES

Adorno,Sergio.1994.Crime,justicapenale desigualdade
juridica:as mortesque se
contamno tribunaldojdri.RevistaUSP21:132-51.
AmericasWatchCommittee. 1993. UrbanPolice Violencein Brazil: Tortureand Police

728

TERESA

P. R. CALDEIRA

AND JAMES

HOLSTON

Killings in Sdo Paulo and Rio de JaneiroafterFive Years.New York:AmericasWatch

Committee.
Americas Watch Committee. 1991. Rural Violence in Brazil. New York: Americas
WatchCommittee.
AmericasWatchCommittee. 1987. Police Abuse in Brazil-Summary Executionsand
Torturein Sdo Paulo and Rio de Janeiro. New York:AmericasWatchCommittee.
Archdiocheseof Sao Paulo. 1986. Torturein Brazil. New York:VintageBooks.
Binder,Leonard,JamesS. Coleman,JosephLaPalombara,LucianW. Pye, Sidney Verba, andMyronWeiner.1971. Crises and Sequencesin Political Development.Princeton: PrincetonUniversityPress.
Bobbio, Norberto.1989. Democracy and Dictatorship:TheNatureand Limitsof State
Power Minneapolis:Universityof MinnesotaPress.
Caldeira,C6sar. 1995. OperacaoRio e cidadania:as tens6es entre o combate a criminalidadee a ordemjuridica.Paperpresentedat Anpocs XIX, Caxambu,Brazil.
Caldeira,TeresaP.R. In press. Cityof Walls:Crime,Segregation,and Citizenshipin Sdo
Paulo. Berkeley:Universityof CaliforniaPress.
Caldeira,TeresaP.R. 1999. "FortifiedEnclaves:The New UrbanSegregation."In Cities
and Citizenship,114-138. Ed. JamesHolston. Durham:Duke UniversityPress.
Caldeira,TeresaPires do Rio. 1991. Direitos humanos ou "privilegiosde bandidos"?
Novos Estudos Cebrap30: 162-74.
Chesnais,Jean-Claude.1981. Histoirede la Violenceen Occidentde 1800 a nos Jours.
Paris:Pluriel.
Chevigny,Paul. 1995. TheEdge of the Knife:Police Violenceand Accountabilityin Six
Cities of theAmericas.New York:New Press.
Coronil,Fernando,andJulie Skurski.1991. "DismemberingandRememberingthe Nation: The Semanticsof Political Violence in Venezuela."ComparativeStudies in Society and History 33(2):288-337.
Currie,David P. 1988. The Constitutionof the UnitedStates:A Primerfor the People.
Chicago:Universityof Chicago Press.
Dworkin,Ronald. 1986. Law'sEmpire.Cambridge,MA: HarvardUniversityPress.
Holston, James. Forthcoming.UnsettlingCitizenship:Disjunctionsof Democracy and
Modernity.
Holston, James. 1991. "TheMisruleof Law: Landand Usurpationin Brazil."Comparative Studies in Society and History 33(4):695-725.
Holston,James,andArjunAppadurai.1999. "CitiesandCitizenship."In Cities and Citizenship, 1-18. Ed. JamesHolston. Durham:Duke UniversityPress.
Huntington,Samuel P. 1991. The ThirdWave:Democratizationin the Late Twentieth
Century.Norman,OK: Universityof OklahomaPress.
Karl,TerryLynn. 1990. "Dilemmasof Democratizationin LatinAmerica."Comparative Politics 23(1):1-21.
Karl,TerryLynn. 1986. "ImposingConsent?ElectoralismversusDemocratizationin El
Salvador."In Elections and Democratizationin LatinAmerica, 1980-1985, 9-36.
Ed. PaulW. Drake,andEduardoSilva. San Diego: Centerfor IberianandLatinAmerican Studies, Universityof California.
Kymlicka,Will. 1995. MulticulturalCitizenship.Oxford:OxfordUniversityPress.
LaPalombara,Joseph. 1971. "Penetration:A Crisis of GovernmentalCapacity."In
Crises and Sequences in Political Development.LeonardBinder, et al. Princeton:
PrincetonUniversityPress.
Lopes, Jos6 Reinaldode Lima. 1992. ResponsabilidadeCivil do Fabricantee a Defesa
do Consumidor.Sao Paulo:EditoraRevista dos Tribunais.
Macedo Jr.,RonaldoPorto. 1998. ContratosRelacionais e Defesa do ConsumidorSao
Paulo:EditoraMax Limonad.

DEMOCRACY

AND

VIOLENCE

IN BRAZIL

729

Mainwaring,Scott, GuillermoO'Donnell, andJ. SamuelValenzuela,ed. 1992. Issues in

Democratic Consolidation:The New SouthAmericanDemocracies in Comparative
Perspective. Notre Dame: Universityof Notre Dame.
Marshall,T. H. 1977[1950]. "Citizenshipand Social Class." In Class, Citizenship,and
Social Development,71-134. Chicago: Universityof Chicago Press.
Moore Jr., Barrington.1966. Social Origins of Dictatorship and Democracy. Boston:
Beacon Press.
Nordlinger,Eric. 1971. "PoliticalDevelopment,Time SequencesandRates of Change."
In Political Developmentand Social Change,2nd ed. Ed. JasonL. Finkle and Robert
W. Gable. New York:JohnWiley.
O'Donnell, Guillermo. 1993. "On the State, Democratizationand Some Conceptual
Problems:A Latin American View with Glances at Some Post-CommunistCountries." WorldDevelopment21(8): 1355-69.
O'Donnell, Guillermo. 1992. "Transitions,Continuities,and Paradoxes."In Issues in
Democratic Consolidation:TheNew SouthAmericanDemocracies in Comparative
Perspective, 17-56. Ed. Scott Mainwaring,Guillermo O'Donnell, and J. Samuel
Valenzuela.Notre Dame: University of Notre Dame Press.
O'Donnell, Guillermo,and PhilippeC. Schmitter.1986. TransitionsfromAuthoritarian
Rule: TentativeConclusionsAbout UncertainDemocracies. Baltimore:Johns Hopkins UniversityPress.
Revista do Consumidor.Sao Paulo: InstitutoBrasileiro de Politica e Direito do Consumidor(Brasilcon).
Pinheiro,Paulo Sdrgio. 1991. Police and Political Crisis:The Case of the MilitaryPolice. In Vigilantismand the State in ModernLatinAmerica, 167-188. Ed. MarthaK.
Huggins. New York:Praeger.
Santos,WanderleyGuilhermedos. 1979. Cidadaniae Justica. Rio de Janeiro:Campus.
Schumpeter,JosephA. 1947. Capitalism,Socialism,and Democracy 2d ed. New York:
Harperand brothers.
Skinner,Quentin. 1989. "The state."In Political Innovationand ConceptualChange,
90-131. Ed. Terence Ball, James Farr,and Russell L. Hanson. Cambridge:CambridgeUniversityPress.
Souza, Edinilsa R. de. 1994. Homicidios no Brasil:o grandevilao da sa6de piblica na
decadade 80. Cadernosde SauidePublica 10(1):45-60.
Souza, Edinilsa, and MariaCecilia de Souza Minayo.1995. O impactoda violencia social na saudep6bicado Brasil:d6cadade 80. In Os MuitosBrasis: Saudee Populadio
na Decada de 80, 87-116. Ed. MariaCecilia de Souza Minayo. Sao Paulo: HucitecAbrasco.
Sunstein,Cass R. 1996.Legal Reasoningand Political Conflict.New York:OxfordUniversity Press.
Taussig,Michael. 1987. Shamanism,Colonialism,and the WildMan:A Studyof Terror
and Healing. Chicago:University of Chicago Press.
Taylor, Charles. 1992. Multiculturalismand the Politics of Recognition. Princeton:
PrincetonUniversityPress.
United Nations. Demographic Yearbook,1993. New York:United Nations.
Valenzuela,J. Samuel. 1992. "DemocraticConsolidationin Post-transitionalSettings:
Notion, Process, and FacilitatingConditions."In Issues in Democratic Consolidation: The New SouthAmericanDemocracies in ComparativePerspective, 57-104.
Ed. Scott Mainwaring,GuillermoO'Donnell, andJ. SamuelValenzuela.Notre Dame:
Universityof Notre Dame Press.
Vigliara, Jose Marcelo Menezes, and Ronaldo Porto Macedo Junior,ed. 1999. Ministerio PutblicoII: Democracia. Sao Paulo:Atlas.