Available online at www.sciencedirect.

com

ScienceDirect
International Journal of Law, Crime and Justice
43 (2015) 295e309
www.elsevier.com/locate/ijlcj

How to commit a perfect murder*

Mark Cooney
Department of Sociology, University of Georgia, USA
Available online 6 June 2015

Abstract

The perfect murder is a recurring theme in many works of art, high and popular. Scientific inquiry has
generally overlooked the issue, though a considerable body of cross-disciplinary evidence documents wide
variation in the handling of homicide in human societies. At one extreme lie those intentional homicides
that result in no legal sanctions and popular praise for the killer, homicides that may be fairly termed
perfect murders. But when will killers get away with murder? The present paper draws upon pure soci-
ology to specify the conditions under which the combination of maximal legal and popular leniency for
homicide occurs. Data on the killing of civilians by police in the United States and Brazil illustrates the
continuum of murder perfection, with those in Brazil providing an especially close fit with the theoretical
model.
© 2015 Elsevier Ltd. All rights reserved.

Keywords: Police homicide; Brazil; Pure sociology

Few things are as intriguing as murder; fewer still are as intriguing as a “perfect” murder e
one for which the killer avoids all penalties. Witness our collective fascination with the murder
mystery e as film, poem, play, opera, documentary, and novel. Protean though the plots and
motifs of the murder mystery are, the form generally utilizes a conventional conception of
lethal violence: murder is wrong and ought to be punished. Secrecy is generally the murderer's
best friend: kill without being identified, even if only for as long as it takes to escape beyond the
long arm of the law. To be known or even suspected as the perpetrator but without the

*
I thank Donald Black, Scott Phillips, and James Tucker for comments on the work presented here. This paper draws
on several sections of my book, Is Killing Wrong? A Study in Pure Sociology.
E-mail address: mcooney@uga.edu

http://dx.doi.org/10.1016/j.ijlcj.2015.05.006
1756-0616/© 2015 Elsevier Ltd. All rights reserved.
296 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309

authorities being able to prove their case is more risky. For greater safety, the killer might make
it look like a suicide. Or, better yet, shift the blame by setting someone else up e an innocent
dupe or, best of all, an enemy, to be caught and punished as the culprit. But no matter how
clever the murder, the truth may seep out. Exceptional powers of detection can solve even the
deepest puzzle. Hence the artistic tension that fuels the genre: who will win the battle of wits?
Can the sleuth e Sherlock Holmes, Hercule Poirot, Phillip Marlowe, Easy Rawlins, Precious
Ramotswe or some lesser known investigator e unravel the knot of deceit, deception, and
dissembling to discover the killer's identity and see justice done? Or will evil ultimately triumph
over good, wrong over right?
Curiously, social science has ignored the problem of the perfect murder. Perhaps the issue
seems trivial (too redolent of mass culture) or morally incorrect (too sympathetic to wrong-
doers). Whatever the reason, the neglect is not justified as the topic harbors an important
scientific question: when will people get away with murder? Moreover, the data for answering
the question already exist. A large body of research conducted by criminologists, anthropol-
ogists, historians, sociologists, human rights activists and others provides a wealth of detail
about the handling of real-life homicide cases across a broad range of human societies (Cooney,
2009). Most of the work seeks to explain variation in case severity, such as why some homi-
cides result in the death penalty and others in life imprisonment (see, e.g., Gross and Mauro,
1989). Less attention has been paid to the other end of the spectrum: leniency. Yet bring the
available information together and for once the clich
e is true: fact really is stranger than fiction.
Where artistic treatments of murder tend to assume that murder is evil, empirical investigation
reveals that some murders are, in fact, good (see e.g., Wilson, 1996). When a killer truly gets
away with murder he does not merely escape punishment but receives the acclaim of his peers,
regardless of how he kills. (Since most murderers are male, I use the masculine pronoun,
though the analysis applies equally to females.) In the real world, the perfect murder is not a
whodunit.
Many examples of perfect murders can be found in human societies, including the killing of
rebellious underlings, murder for reasons of honor, and the elimination of persistently
dangerous persons. In modern societies, the most frequent instances of homicides that rarely
result in legal penalties are, paradoxically, those committed by the very individuals charged
with ensuring that people do not get away with murder: police officers. Why do police ho-
micides generate such mild outcomes? The first explanation likely to present itself e that
killings by police are, in general, legally innocuous e turns out to be unsatisfactory. The
contention of this paper is that the roots of the leniency are primarily sociological, found in
what Black (2000, 2002, 2007) calls the “social geometry” of the case e the location and
direction of police shootings in social space, particularly organizational space. The critical
importance of organizational geometry is revealed by comparing the response to police ho-
micide in Brazil and the United States. But, first, a few words about the theoretical paradigm
from which the concept of social geometry is derived: pure sociology.

1. Pure sociology

Pure sociology is a general explanatory paradigm applicable to every form of social

behavior. Black (1976) invented the paradigm, first applying it to explain variation in the
distribution of law or governmental authority. Subsequently he and others have employed pure
sociology to address a variety of topics, most notably the myriad ways social actors respond to
conflict (see Campbell, 2011).
 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309 297

Pure sociology conceives of social life in quantitative terms. Art, medicine, ideas, and
science, for instance, all vary quantitatively (Black, 1979). Law, too, is a matter of degree,
known by the severity of its sanctions (Black, 1976). Thus, an arrest is more law than no arrest,
an indictment more law than a no bill, a conviction more law than an acquittal, a longer
sentence more law than a shorter sentence; and the death penalty is the most law of all. Popular
justice e the application of non-legal sanctions e is similarly a matter of degree. A retaliatory
killing is more severe than a compensation payment, which is more severe than toleration of the
offense. The most lenient of all sanctions is praise.
Pure sociology explains the handling of deviance with its location, distance, and direction in
a multidimensional social space e its social structure or geometry. The social geometry of a
killing is known by the social characteristics of the parties to the case. Socially close offenses
(between intimates) attract less law than distant offense (between strangers). Similarly,
downward offenses (committed by higher status offenders against lower status victims) attract
less law than upward offenses (committed by lower status offenders against higher status
victims) (Black, 1976, 1993: Chap. 8). And the greater the distance in a downward direction,
the less law the offense will attract. For the same conduct, then, the greater the status superiority
of the killer over the victim, the more leniently the case ought to be handled.
As agents of the state, police officers enjoy an important kind of status over and above
civilians: organizational superiority.1 Organizational superiority is a matter of degree. Some
police officers are more superior to civilians than others, depending, in part, on how centralized
is the state of which they are agents.
Centralization here refers to how strongly states concentrate decision-making. The most
centralized states (dictatorships, autocracies) horde power in few hands. Semi-centralized states
(oligarchies, partial democracies) exhibit some measure of civilian influence (e.g., periodic
elections), but officials wield considerable unchecked power and often accrue favors or exact
bribes in return for performing their official duties. Many public institutions remain effectively
closed to public inspection and control. Police departments, for example, tend to be shielded
from effective civilian influence and inspection, operating under a militaristic model of orga-
nization. Many of the nations of the industrializing world, including Brazil, fall into this
category. Lower on the scale of centralization are modern Western democracies, which disperse
decision making across different branches of government, have regular elections, and subject
their institutions and officials to a considerable measure of civilian participation, criticism, and
scrutiny.2
Centralization elevates the organizational status of state agents, increasing their social dis-
tance from civilians. Police killing of civilians therefore have a distinctive social geometry e
characterized by greater organizational superiority e in more centralized states. And the greater
the police officer's organizational superiority, the more leniency he or she can expect (Black,
1976: 93).
Social geometry provides a radical alternative to traditional legal explanations of police
leniency. In conventional thinking, police killings are handled leniently because they are legally

1
In addition, the victims of police killings are typically low-income individuals engaged in criminal activity (Klinger,
2004). The status superiority of police officers implies that police-citizen killings ought to be handled, on average, with
greater leniency than citizen-police killings. They are: (Cooney, 2009: 67e70).
2
Centralization is a function not just of a state's political structure but of its circumstances. All polities tend to
centralize during national crises, such as wars, terrorist attacks, or natural disasters (Black, 1976: 86). Even the most
democratic democracies may behave like semi-centralized states in times of emergency.
298 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309

justified: the officer did not have an intention to kill, or did so defensively e in response to an
attack on himself or a civilian (see, e.g., Klinger, 2004). However, while the content of the law
is an empirical question its application to the facts of particular cases is a question of inter-
pretation on which lawyers can and often do disagree. Whether any given homicide is a crime
(as distinct from whether it is defined and treated as a crime) is therefore a normative issue
beyond the reach of empirical inquiry.3 Pure sociology predicts, however, that homicides
committed to right a wrong attract less law than those committed for gain or pleasure (Black,
1983). And the greater the degree of deviance to which the killing is a response, the less law it
will attract (Black, 2011b).4 Nonetheless, leniency is not confined to cases in which police kill
in response to serious attacks on themselves or on civilians. Instead, there is a continuum of
killings which police can safely commit that varies with their organizational superiority. The
handling of police homicide in the United States and Brazil illustrates this continuum.

2. Police homicide and the perfect murder

Police homicides everywhere are handled with conspicuous lack of severity. Seldom are
police officers convicted and sentenced to long prison terms or the death penalty for killing
civilians. Even so, the degree of leniency fluctuates: some settings are more conducive to
perfect murders by police than are others. Consider, first, the United States.

2.1. The United States

A study of 1500 police-civilian killings nationwide 1952e1969 reported that just three
officers (0.2%) were convicted of a crime. Moreover, of the three officers, one was sentenced to
five years to life, a second to five years, and the third to one day (Kobler, 1975). Had the parties
in the following case been reversed, for example, a similar result would have been virtually
unthinkable:
Patrolman S.S. was charged with first-degree murder, accused of deliberately shooting
C.H., a 27-year old fugitive, from a distance of four feet with a 12-gauge shotgun while
H. was crouched on his knees and elbows in a bush where he hid while trying to elude the
police. The shotgun S.S. used was an unauthorized weapon. S.S. had deliberately traveled

3
Those who analyze the criminality of police violence often disagree. Some scholars assert that many police homi-
cides are not legally justified (see, e.g., Collins, 1998). Others suggest that the great majority of such homicides are
legitimate (see, e.g., Klinger, 2004).
4
Black's (2011a) recent theory of moral time proposes that the greater the movement of social time, the greater the
conflict (including legal severity). Social time moves when actions increase or decrease stratification, intimacy, or
diversity. For example, a homicide deprives the victim of his or her livelihood and is therefore a reduction in social
status (understratification), a rape is a radical increase in physical contact (overintimacy), while the appearance of a new
street drug represents a sharp increase in cultural diversity (overdiversity). Thus, if a poor man murders a wealthy man
the understratification is greater than if the same individual murders a poor man and hence attracts more punishment, a
prediction supported by the evidence (Cooney, 2009: Chap. 3). Movements of social time help to explain why Brazilian
police homicides are generally treated with such leniency. Brazil has high crime rates e large movements of social time
e and since social control reciprocates deviance, the response to crime tends to be severe. The response is likely to be
especially severe when the crime is itself an attack on police or other state agents (a form of understratification). Still,
most police killings are not triggered by attacks on police but by crimes against citizens, suspected or real. Such crimes
are common in the U.S. as well, yet the police kill much less frequently in response to them and when they do, the
response is less lenient. In part this is because of the stronger partisanship that Brazilian officers attract, a pattern that
appears to lie outside the theory of moral time.
 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309 299

some seven miles out of his district to serve the felony warrant on C.H., and it was alleged
at the trial that S.S. was having a sexual affair with the wife of C.H. The jury was
instructed it could return one of four verdicts: first degree murder, second degree murder,
manslaughter, or acquittal. They found S.S. guilty of manslaughter. He was sentenced to
five years on that charge and three years for perjury, to be served concurrently (Kobler,
1975: 178e179).
Subsequent changes in the law and in society have resulted in greater internal scrutiny of
police killings by host police departments and by legal officials. Even so, very few officers in
contemporary America wind up being disciplined or punished for killing civilians:
Whatever protocol an agency employs, nearly all shootings are found to be ‘within
policy,’ so officers are rarely disciplined by their departments for shooting someone ….
Whatever procedures a given jurisdiction might employ to review the legality of shoot-
ings, criminal inquiries nearly always find that the police acted within the scope of the
relevant law, so officers rarely face criminal charges for the actions they take during
shooting incidents (Klinger, 2004: 205.).
Many officers are not even indicted. For example:
In the 1980s, Dallas Grand juries refused to indict officers who shot seventy-year old Etta
Collins, an unarmed black woman, through the front screen door of her house; Larry
Brice, an unarmed twenty-year old white who was shot nine times; David Horton, an
eighty-one-year-old black man who, as a crime watch coordinator for his neighborhood,
had phoned in a burglary report; and Sammy Stone, a black burglar who was shot to death
after he had been handcuffed (Swindle, 1993: 85).
More recently: “In Dallas, grand juries reviewed 81 [police] shootings between 2008 and
2012 and returned just one indictment” (Casselman, 2014). Nationwide, over a seven-year
period ending in 2011, just 41 officers were charged with murder or manslaughter while at
least 2718 killings by police officers received the official designation of “justifiable homicide”
(Elison and Palazzolo, 2014).
Conviction is even more rare. Even highlyepublicized cases, such as that of Amadou Diallo
(a West African street vendor killed after police fired 41 bullets at him), and Sean Bell (killed in
a fusillade of 50 police bullets after his bachelor party), may result in acquittals (New York
Times February 26, 2000; April 26, 2008). In New York, “despite scores of fatal shootings,” not
a single on-duty police homicide resulted in a conviction between 1977 and 1995 (Collins,
1998: 87). In Los Angeles, charges were filed in just one of the 174 fatal shootings that
occurred between 1979 and 1991 (Katz, 1991). No criminal charges were filed against any of
the 77 San Diego police officers who killed civilians from January 1, 1985 to December 20,
1990 (Petrillo, 1990). If there is a conviction, it is almost never for murder (Bulwa, 2009).5
American police killings of civilians enjoy, in effect, a presumption of legitimacy. That
presumption is not absolute, however. Officers who have a history of violence, who kill mul-
tiple civilians, or who kill with clear premeditation run a strong risk of experiencing legal
sanctions. That is not true, however, of their counterparts in more centralized states. One such
nation, for which there is good information, is Brazil, where an unusually strong tradition of

5
According to Wikipedia (as of May 15, 2015), just 15 police officers have been convicted in the United States of
murder: http://en.wikipedia.org/wiki/Category:American_police_officers_convicted_of_murder.
300 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309

investigation by human rights groups has drawn a detailed and coherent portrait of the even
greater immunity enjoyed by police officers who kill civilians.6

2.2. Brazil

In Brazil, federal authorities have some police functions, but most of the responsibility for
combating crime falls on the states. State police are organized into two separate entities: the
military police and the civil police. The military police is a force of uniformed officers charged
with patrolling public spaces, maintaining public order, and making arrests pursuant to warrants
or apprehending suspects committing crimes. The civil police investigate crimes. “Thus the
military police respond to crime while they are in progress, while the civil police respond to
crimes once they have occurred” (Jahangir, 2004, para. 24). It is the military police that commit
most of the killings of civilians.
Police homicides in Brazil and the United States differ in the aggregate in important ways.
Compared to their North American counterparts, Brazilian police:

2.2.1. Kill more civilians

Although Brazil has a “notoriously high homicide rate,” the “police are responsible for a
significant proportion of all killings” (Alston, 2009: 7e8). For every 100 non-police homicides
committed in 2008, on-duty police in the states of Rio de Janeiro and S^ao Paolo killed 20
people and 8 people respectively. Rio's rate of police killing that year was 6.8 per 100,000; S^ao
Paolo's was 0.97 (compared to 0.12 in the United States) (Delgado, 2009: 30e33). These
patterns have persisted for many years. In the early 1990s, S^ao Paolo had a rate of police
homicide over 15 times that of New York City (Chevigny, 1995: 67, 146). In 2012, across the
country as a whole, the police killed an average of five people a day (Human Rights Watch,
2014).

2.2.2. Kill more groups of civilians

Where police homicides in the United States typically involve a single victim, multiple
victims are common in Brazil. In one infamous incident in October 1992, military police
stormed a Sao Paulo prison (popularly known as the Carandiru) to quell a riot. By the time they
left, 111 prisoners were dead. Investigations revealed that “the police summarily executed
dozens of their victims, many after they had been forced to strip naked and return to their cells”
(Cavallaro, 1997: 14). In a second incident, after four police officers were killed by drug
traffickers in a favela (i.e., a squatter settlement) in Rio de Janeiro in August 1993, a group of
hooded and heavily armed men entered the favela and shot indiscriminately at local residents
for two hours, killing 21 people (Amnesty International, 1997). In March 2005, 29 people were
killed in another Rio favela when a group of heavily armed men, believed to be military police

6
Although human rights research is conducted primarily for political reasons, the social structure of the research
suggests that it yields valid empirical data. Human rights reports are typically critical of high status individuals (e.g.,
politicians) and institutions (e.g., the police) that are usually well positioned to defend themselves; the researchers are
therefore under considerable pressure to get their facts right or otherwise they find themselves publicly discredited and
perhaps banned. The reports also evince a high degree of reliability. Investigations conducted by, for example, Amnesty
International, Human Rights Watch, and the United Nations Commission on Human Rights paint a remarkably
consistent portrait of the circumstances and outcomes of Brazilian police homicides. They are also highly compatible
with Brinks (2008) scholarly research into the response to police homicide in several South American countries.
 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309 301

officers, drove through the streets and shot randomly at the inhabitants (Amnesty International,
2005).

2.2.3. Kill with more apparent premeditation

Brazilian police routinely claim that the victims fired on them first (known as “acts of
resistance”). Between January 2003 and September 2009, police in the states of Rio de Janeiro
and S^ao Paolo killed over 11,000 people in resistance killings, according to official statistics
(Delgado, 2009: 20). To back up their claim, the police frequently produce a weapon that they
claim they found on the victim (Delgado, 2009: 63e65). However, “people who are knowl-
edgeable about the military justice system say that it is common for the police to plant such
weapons, which are called in slang ‘carbitos’” (Chevigny, 1993: 9). Independent investigation
of the police version is difficult because the officers often disturb the crime scene by, for
example, taking the already deceased victim to hospital, purportedly for medical treatment (see
e.g., Penglase, 1994: 49; Delgado, 2009: 52e58). Nonetheless, when investigations have been
conducted “forensic reports and witness accounts generally show that the lethal shots had been
fired from behind and at close range, in circumstances that the person was a victim of extra-
judicial execution” (Jahangir, 2004: para. 40; see also Hinton, 2006: 111; Delgado, 2009:
52e65). Moreover, if police claims were true, the number of police victims ought to be close to
the number of civilian victims. In fact, far more civilians are killed by officers than officers are
killed by civilians: 18 times as many in S^ao Paolo State in 2008 and 44 times as many in Rio De
Janeiro State if victims of off-duty police are excluded (Delgado, 2009: 35). If, on the other
hand, police are routinely executing civilians instead of killing them in shoot-outs, many more
civilians ought to be killed rather than wounded. Again, the numbers support that interpretation:
in 1991, for example, while 78% of police shootings in metropolitan S^ao Paolo resulted in
death only 24% of police shooting in New York City and 37% in the city of Los Angeles were
fatal (Chevigny, 1995: 46, 67, 148; see also Delgado, 2009: 36). In short, “there are simply too
many killings, too large a percentage of those shot are killed, too many victims are minors with
no criminal backgrounds, for the stories about ‘shootouts’ to be true” (Chevigny, 1993: 24).7
Despite the large number, indiscriminate nature, and apparent premeditation of many Bra-
zilian police homicides, most officers are not charged with a criminal offense and hence
punishment is rare. A study of police homicides in the 1990s in two cities, S^ao Paolo and
Salvador de Bahia, reports conviction rates “well below 5%” (Brinks, 2008: 11). Punishment is
especially rare when the victims are low status civilians, such as those involved in criminal
activities (Brinks, 2003).8 Brazil has high rates of violent, property, and drug-related crime and
there is considerable public support for stern measures against criminals (see, e.g., Holston and
Caldeira, 1998; Alston, 2009: 16, note 33). The lynching of criminals is fairly common (see
e.g., Chevigny, 1995: 157). Aware of the popular feeling, police routinely claim that their
homicide victim was a criminal. If pressed, they will often devote considerable energy into
uncovering a criminal background. Should they find one, the investigation is effectively closed.
As Cavallaro (1997: 24) notes: “[I]mplicit in this procedure is the idea that the police may kill
criminals without fear of consequence.”

7
Brinks (2008: 157) estimates that the S^ao Paolo military police “tampered with the factual record” in 85% of the
cases in his sample during the 1990s.
8
Brazil's higher levels of poverty, unemployment, and homelessness means that the typical low status victim of police
homicide there is of lower status than his or her U.S. counterpart, a fact that further explains the less severe penalties for
Brazilian police.
302 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309

The poor and the homeless are other low status groups whose killing by police excites little
legal action.9 Many of the victims are children and young adults who work or live on the streets
(Penglase, 1994: 7e12). Some have committed crimes; others are merely suspected of having
done so. Regardless of their conduct, marginality itself appears to be the underlying offense.
One nine-year old boy was found strangled to death on a beach in Rio De Janeiro with a note on
his body that read, “I killed you because you didn't go to school and had no future” (quoted in
Dimenstein, 1991: 31).
Killings of street children e and other civilians e are sometimes carried out by death squads
e groups of heavily armed hooded men in civilian clothes (see, e.g., Scheper-Hughes, 2006).
The line between death squads and the police is often quite blurred as death squads are often
staffed, at least in part, by off-duty or former police officers who may be involved in criminal
activities or may simply be seeking to make extra money working as security guards for small
businesses (Jahangir, 2004: para. 43; Hinton, 2006: 112)). Whether street children are killed by
death squads or on-duty police, the legal repercussions typically range from slight to non-
existent: The child disappears never to be heard from again; end of case. If an investigation
is launched, it soon peters out (see Penglase, 1994: 40e41).

2.2.4. Third parties

A major reason why police officers fare so well in criminal cases around the world is that
they enjoy strong partisan support from other officials. Partisans enhance the social status of a
party's case (Black, 1993: 127). And the stronger their partisan support, the greater the status
boost third parties bring. Particularly valuable are high status and distant partisans, though they
are harder to recruit.
Black (1993: Chap. 7) proposes that “[p]artisanship is a joint function of the social closeness
and superiority of one side and the social remoteness and inferiority of the other” (Black, 1993:
127). Thus, wealthy, integrated, respectable litigants are likely to have more witnesses and
lawyers on their side, especially when the other side is poor, marginal, and unrespectable,
thereby augmenting their status superiority.10 Similarly, litigants with a solid corps of family
members and close friends are likely to enjoy strong partisanship, particularly when the third
parties are socially distant from the other side (see Cooney, 1994).
The high organizational status of police coupled with their social closeness to fellow offi-
cers, medical examiners, investigators, and prosecutors generally ensures that they receive
considerable support from people who can influence the course of legal events (Black, 1993:
Chap. 7).11 From start to finish, the officer is likely to receive the benefit of any doubt.
Crucially, the “facts” of the incident are likely to be construed in the officer's favor. No pun-
ishment can be imposed unless the killing is first ruled a homicide, a determination initially
made by the coroner or medical examiner, both of whom work closely with the police. A

9
Also relevant is the race of the victims. Although racial classifications are somewhat fluid in Brazil, people of color e
whether negro (black) or pardo (brown) e are, on average, of lower socio-economic status than whites and are killed, it
appears, in excess of their numbers in the population (Penglase, 1994: 2e5).
10
Consider, for example, the “dream team” of nationally known lawyers, investigators, jury consultants and forensic
experts that O.J. Simpson, the wealthy and famous ex-football player, put together to defend himself in his criminal and
civil murder trials (see, e.g., Toobin, 1996: 86e87, 128e133, 187; 212e214; Petrocelli, 1998). A typical, low status
homicide defendant could never hope to mobilize such a powerful fighting legal machine.
11
For Black (1993: 138e139; 2002: 111e112) the partisan conduct of legal officials in cases against police is an
example of a more general characteristic of law everywhere: it is a form of “slow partisanship” e far from being neutral
law typically takes sides in conflict, but does so only after a consideration of evidence and argument.
 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309 303

sociologist who observed a medical examiner's office in the United States over a three-year
period notes that “[u]nless there are extraordinary circumstances, law enforcement officials
will receive a generous benefit of the doubt” (Timmermans, 2006: 183). Consequently, some
cases that might be categorized as homicide are classified instead as self-inflicted fatalities
triggered by provoking the police to lethal violence, or “suicide by cop” (Timmermans, 2006:
183e185). A similar advantage arises when the case is investigated:
Inside every major police department, the initial investigation of any officer-involved
shooting begins as an attempt to make the incident look as clean and as professional
as possible (Simon, 1991: 377).
Police benefit too from the “code of silence,” “an unwritten rule that police will not testify
against a fellow officer and that police are expected to help in any cover-up of illegal action”
(Rudovsky, 1992: 481, note 60). So familiar is the code of silence that a neologism has been
coined to describe it: “testilying” (Koepke, 2000: 221).
Should the officer be arrested the prosecutor must decide whether to seek a conviction.
Again, the prosecutor has stronger ties to the police than to civilians. And if the case happens to
go to court, the judge too is organizationally and personally closer to the officer defendant than
to the civilian victim.
Take away some or all of the favorable third-party geometry and police homicide become less
than perfect. In civil actions for wrongful death, for example, the officer is not opposed by his
fellow professionals but by strangers e the relatives of the victims e and is more likely to lose.12
The higher organizational status of Brazilian police yields even stronger partisanship than that
enjoyed by North American officers. They commonly ignore civilians with incriminating in-
formation against their fellow officers: “witnesses critical of the police version often simply [are]
not interviewed” (Cavallaro, 1997; 24). If witnesses do find a voice, they may be intimidated or
threatened by the police, a particularly strong form of partisanship. For example, in one trial of an
officer accused of killing street children in Rio De Janeiro in 1993, several witnesses provided
investigators with information. By the time the case got to court, however, only one of the wit-
nesses was prepared to testify. He paid dearly for his pluck, suffering an attempt on his life, which
left his face partially paralyzed (Amnesty International, 1996). He is not alone. The killing of
witnesses is sufficiently common to have acquired its own nickname, “burning the archives” or
“quiema de arquivo” (Penglase, 1994: 42). Other documented examples include:

 A woman was killed under suspicious circumstance in Rio de Janeiro in 1993 after she
testified in court to police involvement in the disappearance of her son and ten friends
(Amnesty International, 1996: Appendix 3).
 A man who provided information about the killing of his brother by a death squad with
alleged ties to the police was killed in Santo Antonio de Jesus in 2003 (Jahangir, 2004:
Para 3).

12
In the United States, damage awards and out of court settlements against cities are “common” according to one
expert (Klockars, 1996: 4). Just how common is unclear in the absence of statistics (but see Pate and Fridell, 1993).
Some indication can be gleaned from the large amounts of money paid by cities to plaintiffs in damage award or settling
cases of improper police actions (including homicide): The city of Los Angeles paid just short of $80 million in civil
actions against police officers between 1991 and 1996. Between 1994 and 1996 New York paid about $70 million
(although this figure also includes traffic accidents). Detroit paid $72 million between 1987 and 1994. Many of these
awards were for lethal violence (Collins, 1998: 76e85).
304 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309

 A man who claimed he was beaten by S^ao Paolo police in 2009 and went to file a formal
complaint was told by a police officer, “I will fill you full of lead, this time you will not
escape, you shit.” Two months later he was shot 32 times by unknown assailants (Delgado,
2009: 69).

Overall, a study of S^ao Paolo police homicides from the 1990s found “clear evidence” of
witness intimidation in “nearly 25 percent” of cases (Brinks, 2008: 158).
Strong partisanship means that even when Brazilian police kill in their private capacity (e.g.,
within the family), they still enjoy a measure, albeit reduced, of immunity from punishment
(Brinks, 2008: 57). Most police homicides are investigated by the civil police who collaborate
closely with military police and have the same commander-in-chief (the state governor)
(Delgado, 2009: 87). Routinely, investigators fail to: question police officers closely, interview
all officers involved in a killing, obtain non-police eyewitnesses, conduct forensic examina-
tions, and investigate cases promptly (Delgado, 2009: 71e85). The officer under investigation
is not suspended or administratively sanctioned (unless convicted) (Jahangir, 2004: para. 22).
Prosecutors may sometimes initiate investigations proactively but they are severely hampered
by a lack of financial and human resources (Jahangir, 2004: paras 25e27).13 However, orga-
nizational inefficiency or personal incompetence cannot explain the perfunctory quality of the
investigations. When a police officer, especially a high ranking one, is the victim, police in-
vestigations display exemplary thoroughness and professionalism (Chevigny, 1993: 13e14).
Most victims of police killings, however, have little social standing and receive little or no
partisan support, a deficit which consolidates their low status and results in their cases going
nowhere. Even if their family members, friends, or neighbors do come forward, their witnesses
generally lack much credibility. Moreover, their families are usually unable to take advantage
of a provision of Brazilian law that allows the family of the victim, with the approval of the
court, to appoint an assistant to the prosecutor who can propose legal arguments, suggest
questions for witnesses, and participate in oral argument and appeals. Street people with no
effective families especially lack the requisite support and their cases are left to the vagaries of
the criminal justice system (Penglase, 1994: 46e47). As a result, the killing of social marginals
is effectively legitimate.
In combination, these patterns mean that the police effectively control the evidentiary
process. They determine the “facts” of the case. Not surprisingly, then, most cases are dismissed
for lack of evidence. Brinks's (2008: 96) reports that in his study spanning the 1990s:
Nearly 65% of all the cases that I collected in S^ao Paolo were dismissed early on by
agreement of the prosecutor and judge, ostensibly because the available evidence was
insufficient to establish that the incident involved the excessive use of force. Even the few
cases that went to trial failed more often than not to produce a conviction e 60% of the
tried cases ended in acquittals, compared to less than 25% in ordinary homicide cases.
Occasionally, however, things are different. A case that garners the attention of high status,
socially distant third parties, such as newspapers, television stations, or human rights organi-
zations may be publicized nationally and internationally. The social condition that elicit these
periodic instances of unusual partisanship are not yet clear, but when they occur, the authorities
may be shamed into taking action, and investigations, prosecutions, and even convictions may

13
Until 1996, military courts tried murders by military police (Cavallaro, 1997: 22e23).
 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309 305

result. The Carandiru prison massacre of 1992 is the prime example. In 2013 and 2014, after
two decades of campaigning by human rights groups and others, 73 police officers were
eventually sentenced to prison terms of varying lengths for the killing of the 111 prisoners
(Human Rights Watch, 2015). In another incident, widespread national and international media
coverage contributed to the conviction and 13-year sentence of a police officer who killed a 12-
year old boy he found reading a pornographic magazine (Brinks, 2003: 12e13). Brinks (2008:
164) found that adding a private assistant prosecutor to a case, something most often done by a
Catholic human rights group, increased the probability of conviction in S^ao Paolo in the 1990s
from zero to 32 percent (2008: 164). Outcomes such as these that appear incongruent with the
social geometry of the principal parties are, in general, often explained by the presence of
socially desirable partisans (Baumgartner, 1992: 148e149).

3. Popular justice

Recall that a perfect murder results not just in legal toleration but minimal popular sanctions
as well. The killer does not just get away with murder, he profits from it. Thus, the killer might
be paid, receive a promotion, or be treated as a hero: Regardless of the exact type of praise
bestowed upon him, his stock rises.
Pure sociology predicts that the same variables predict leniency in both law and popular
justice. When cases have similar legal and popular geometries they ought to occupy similar
positions on the severity scale of legal and popular sanctions.14 The downward nature of police
homicides implies, then, that any popular sanctions found are likely to be mild, except where
the victim happens to have garnered third-party support. In Brazil, where the status gap between
police and civilians is especially wide popular justice should be all the milder. That appears to
be the case.
The massive size of the state blocks out much of the light of popular justice in modern
society and restricts what is known about how civilians respond to killings by police. Even so,
some information is available on the popular reception of police killers. It suggests that Bra-
zilian police officers are often respected, admired, and honored for their killings. This takes a
variety of forms. One Sao Paolo officer who was involved in 44 homicides received a pro-
motion (Chevigny, 1993: 11). A police lieutenant in Rio de Janeiro who led or participated in
11 police operations that resulted in 18 deaths was promoted in 1996 (Cavallaro, 1997: 27e28).
Others have received pay increases for their killings (Cavallaro, 1997: 15e16). Still others have
been awarded citations and medals, at the conferring of which the number of people they have
killed is proudly announced (Amnesty International, 1996). The officer who commanded the
1992 raid on the Carandiru prison in which 111 prisoners were killed, the great majority after
already surrendering, stood for political office, choosing the number 111 as his electoral ticket.
He was elected (Amnesty International, 1999, 2001).15
The popular response to members of Brazilian death squads e many of whom are police
officers or have close ties to them e illustrates the same patterns of approbation. Not only are

14
Cases that result in, say, an acquittal will not, however, necessarily attract minimal popular sanctions since the parties
to the legal and popular case may differ significantly. O.J. Simpson, for example was acquitted by a predominately black
jury but convicted in the court of white public opinion, a conviction that led to his being shunned in a variety of ways
(Toobin, 1996; Jordan, 2001).
15
Until January 1999 he benefitted from parliamentary immunity. His 2001 conviction was overturned in 2006: http://
news.bbc.co.uk/2/hi/americas/4718592.stm. Retrieved May 15, 2015.
306 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309

death squad participants rarely prosecuted they sometimes become folk heroes. Many people
admire them and applaud their actions. The 1984 arrest in Sao Paolo of a man believed to have
killed 72 people triggered a public protest in the center of the city by residents of his neigh-
borhood. Another well-known death-squad leader was awarded an honorific title by his city for
his services and proudly stood as a candidate in local elections (Dimenstein, 1991: 45e46).

4. Conclusion

“No man can bring about the perfect murder; chance, however, can do it,” the narrator of
Vladimir Nabokov's Lolita (1958: 86) informs the reader. And perfect murders in the sense
discussed here could conceivably be statistical flukes. Whether a killer goes unpunished might
simply be the product of a host of chance factors e carelessness in the collection of evidence,
skewed prosecutorial priorities, idiosyncratic jurors and the like. Likewise, whether a person is
praised for a murder might be just a matter of good luck e the channels through which the story
flows, a demand for heroes among the public, the personal charisma of the killer and so forth.
But however large a role chance factors play in individual instances, an analysis of cases in
the aggregate reveals strong regularities at work. Look at multiple, rather than single, cases and
perfect murders exhibit patterns, occurring under some conditions more than others.
The regularities underlying perfect murders are not legal, or at least not primarily legal,
however. Perfection cannot be explained merely by the conduct and state of mind of the killer.
One of the most striking features of many perfect murders is how non-defensive and intentional
they are. It is true that the written law is more tolerant of intentional killings when they are
committed by public officials, such as police officers, in the course of their duties. But even if
most police homicides are legally justified (which is not an empirical question) the determi-
nation that they are is heavily influenced by their inherent credibility as officials of the state and
the partisanship they can evoke from fellow officers, factors that transcend legal explanations.
Not that what parties to the killing did is immaterial. A police officer who kills a civilian in
self-defense, for instance, is virtually certain to escape conviction. To explain the handling of
particular police killings it is necessary to include the actions of the parties e how their conduct
expands or contracts social space. The more deviant the conduct to which the killing is a
response, the less law it will attract (Black, 2011b; see also note 4 above). However, the
relevance of the parties' conduct is itself sociologically variable: Increase organizational su-
periority, for example, and the limitations imposed by rules on the defensiveness and inten-
tionality of killing dwindle allowing Brazilian police to get away with more deviant (aggressive
and intentional) killings than American police. But even Brazilian police often try to conceal
the true circumstances of their homicides, taking pains to make them appear to be the result of
defensive shoot-outs. Such tactics would not be necessary in still more centralized states
(including Brazil under military rule). In the most centralized states of all e Stalinist Russia,
Nazi Germany, or Maoist China, for instance e state officials have killed millions of civilians
with virtually complete impunity (e.g., Rummel, 1994).
Brazil is exceptional only in that more information is available from it than most other less
developed nations. Police homicides are common elsewhere in the industrializing nations, such as
India (see, e.g., Eckert, 2005). Death squads too operate in other countries, such as Colombia (see,
e.g., Campbell and Brenner, 2000; Massing, 2003). Yet some less wealthy nations (e.g., Uruguay)
have low rates of police homicide and high rates of prosecution for those that do occur. Perfect
police murders are not inevitable. Strengthen the perfection-diluting variables e more open police
departments, more public scrutiny of officials killings, more partisanship for victims of police
 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309 307

homicide, or, (as in Uruguay) a comparatively less disadvantaged poor population e and the
leniency with which they are treated will diminish as well (Brinks, 2008). Moreover, even where
the background geometry point toward leniency, the social structure of the particular case remains
crucial: if the officer has a poor reputation among colleagues or the victim is of higher status (high
official, wealthy, employed, married with children, no criminal record) or has the backing of
influential partisans, the likelihood that the killing will turn out to be perfect shrinks.
Police homicides are not the only incidents that fit the model of the perfect murder. The killing
of a rebellious slave by a master is perhaps the closest fit of all (see e.g., Schwarz, 1988). Other
homicides that display much of the requisite geometry include the slaying of low status blacks by
whites in the Jim Crow South, the killing of sexually deviant women in honor societies, and the
liquidation of notorious thieves or bullies among the urban poor of the world: these too are likely
to go legally unpunished and be popularly praised (see, e.g., Brundage, 1993; Peratis, 2004;
Jackall, 1997: 110). Critical to all of these instances is stratification. Perfect murders are
invariably committed downward in one or more dimensions of social space. In police killings,
organizational stratification is central, but economic (killing of slaves), familial (honor killings),
and moral (the killing of social enemies) inequalities can be critical as well. Pronounced status
disparities between the parties are the sine qua non of perfect murders.
Fact, then, deviates sharply from fiction. In the real world, the perfect murder is not
committed by an evil genius but by a moral agent acting in the name of the common good e
fighting crime, restoring honor, eliminating enemies, protecting communities. Nor is it a
mystery: the killer's identity is publicly known, and the killing is tolerated, even applauded, by
legal officials and by fellow civilians. The killer may even wear the badge of the law. In short,
while the perfect murder remains a source of aesthetic fascination, it is no longer a scientific
puzzle. To commit a perfect murder, the killer should:

 Be of as high status as possible.

 Select a low status victim who is, ideally, socially close as well.
 Have extensive strong partisanship and the victim none at all.
 Be significantly closer, socially, than the victim to those making the crucial decisions in the
case.

References

Alston, Phillip, 2009. Report of the Special Rapporteur on Extrajudicial, Summary, or Arbitrary Executions: Mission to
Brazil. http://www.unhcr.org/refworld/type,MISSION,BRA,49f6c5602,0.html (retrieved 15.05.15.).
Amnesty International, 1996. The Candelaria Trial: a Small Wedge in the Fortress of Impunity. Amnesty International
Index: AMR 19/020/1996.
Amnesty International, 1997. Brazil: Candelaria and Vigario Geral: Justice at a Snail's Pace. Amnesty International
Index AMR 19/011/1997.
Amnesty International, 1999. 7th Anniversary of Carandiru Massacre: Contempt and Neglect for 111 Lives. Amnesty
International Index AMR 19/25/99.
Amnesty International, 2005. “They Come in Shooting”: Policing Socially Excluded Communities. Amnesty Interna-
tional Index AMR 19.025/2005.
Baumgartner, M.P., 1992. The myth of discretion. In: Hawkins, Keith (Ed.), The Uses of Discretion. Clarendon Press,
Oxford, pp. 129e162.
Black, Donald, 1976. The Behavior of Law. Academic Press, New York.
Black, Donald, 1979. A strategy of pure sociology. In: McNall, Scott G. (Ed.), Theoretical Perspectives in Sociology. St.
Martin's Press, New York, pp. 149e168.
308 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309

Black, Donald, 1983. Crime as social control. Am. Sociol. Rev. 48, 34e45.
Black, Donald, 1993. The Social Structure of Right and Wrong. Academic Press, San Diego.
Black, Donald, 2000. Dreams of pure sociology. Sociol. Theory 18, 343e367.
Black, Donald, 2002. The geometry of law: an interview with Donald Black. Int. J. Sociol. Law 30, 101e129.
Black, Donald, 2007. Legal relativity. In: Clark, David S. (Ed.), Encyclopedia of Law and Society: American and
Global Perspectives, vol. 3. Sage Publications, Thousand Oaks, CA, pp. 1292e1294.
Black, Donald, 2011a. Moral Time. Oxford University Press, New York.
Black, Donald, 2011b. Electronic Posting to the Pure Sociology Network. September 2, 2011.
Brinks, Daniel M., 2003. Informal institutions and the rule of law: the judicial response to state killings in Buenos Aires
and S^ao Paolo in the 1990s. Comp. Polit. 36, 1e19.
Brinks, Daniel M., 2008. The Judicial Response to Police Killings in Latin America: Inequality and the Rule of Law.
Cambridge University Press, Cambridge.
Brundage, W. Fitzhugh, 1993. Lynching in the New South: Georgia and Virginia, 1880e1930. University of Illinois
Press, Urbana.
Bulwa, Demian, 2009. Ex BART cop is accused of murder in rare group. San Francisco Chronicle, February 15: A-1. http://
articles.sfgate.com/2009-02-15/news/17187934_1_police-officers-bart-police-dangerous-job (retrieved 15.05.15.).
Campbell, Bradley, 2011. Black's theory of law and social control. In: Rosenfeld, Richard (Ed.), Oxford Bibliographies:
Online Criminology.
Campbell, Bruce D., Brenner, Arthur D., 2000. Death Squads in Global Perspective: Murder with Deniability. St.
Martin's Press, New York.
Casselman, Ben, 2014. It's incredibly rare for a grand jury to do what Ferguson's just did. Five Thirty Eight, November
24. http://fivethirtyeight.com/datalab/ferguson-michael-brown-indictment-darren-wilson/ (retrieved 15.05.15.).
Cavallaro, James, 1997. Police Brutality in Urban Brazil. Human Rights Watch, New York.
Chevigny, Paul, 1993. Urban Police Violence in Brazil: Torture and Police Killings in Sao Paolo and Rio de Janeiro
After Five Years. Human Rights Watch, New York.
Chevigny, Paul, 1995. The Edge of the Knife: Police Violence in the Americas. The New Press, New York.
Collins, Allyson, 1998. Shielded from Justice: Police Brutality and Accountability in the United States. Human Rights
Watch, New York.
Cooney, Mark, 1994. Evidence as partisanship. Law Soc. Rev. 28, 833e858.
Cooney, Mark, 2009. Is Killing Wrong? A Study in Pure Sociology. University of Virginia Press, Charlottesville.
Delgado, Fernando Ribeiro, 2009. Lethal Force: Police Violence and Public Security in Rio de Janeiro and S^ao Paolo.
Human Rights Watch, New York.
Dimenstein, Gilberto, 1991. Brazil: War on Children. Latin American Bureau, London.
Eckert, Julia, 2005. Death and the nation: state killing in India. In: Sarat, Austin, Boulanger, Christine (Eds.), The
Cultural Lives of Capital Punishment: Comparative Perspectives. Stanford University Press, Stanford, pp. 194e218.
Elison, Zusha, Palazzolo, Joe, 2014. Police rarely criminally charged for on-duty shootings. November 24
Wall Street J. http://www.wsj.com/articles/police-rarely-criminally-charged-for-on-duty-shootings-1416874955
(retrieved 15.05.15.).
Gross, Samuel R., Mauro, Robert, 1989. Death and Discrimination: Racial Disparities in Capital Sentencing. North-
eastern University Press, Boston.
Hinton, Mercedes, 2006. The State on the Streets: Police and Politics in Argentina and Brazil. Lynne Rienner Pub-
lishers, Boulder.
Holston, James, Caldeira, Teresa R.R., 1998. Democracy, law, and violence: disjunctions of Brazilian civilianship. In:
Agüero, Felope, Stark, Jeffrey (Eds.), Fault Lines of Democracy in Post-Transition Latin America. North-South
Center Press/University of Miami, Coral Gables, FL, pp. 263e289.
Human Rights Watch, 2014. Brazil. World Report 2014. Human Rights Watch, New York.
Human Rights Watch, 2015. Brazil, World Report 2015. Human Rights Watch, New York.
Jackall, Robert, 1997. Wild Cowboys: Urban Marauders and the Forces of Order. Harvard University Press, Cambridge.
Jahangir, Asma, 2004. Report of the Special Rapporteur on Extrajudicial, Summary or Arbitrary Executions, Asma
Jahangir, on her Mission to Brazil, 16 September - 8 October, 2003. http://www.extrajudicialexecutions.org/
application/media/%28Brazil%29%20E_CN_4_2004_7_Add_3.pdf (retrieved 14.01.12.).
Jordan, Pat, 2001. The Outcast. New Yorker 42e47. July 9.
Katz, Jesse, 1991. Is training of deputies for deadly clashes adequate? Despite extensive instructions on when to shoot,
critics say officers are too quick on the trigger. Los Angeles Times. September 8: A1, A22, A23.
Klinger, David, 2004. Into the Kill Zone: a Cop's Eye View of Deadly Force. Jossey-Bass, San Francisco.
 M. Cooney / International Journal of Law, Crime and Justice 43 (2015) 295e309 309

Klockars, Carl B., 1996. A theory of excessive force and its control. In: Police Violence: Understanding and Controlling
Police Abuse of Force. Yale University Press, New Haven, pp. 1e22.
Kobler, Arthur L., 1975. Police homicide in a democracy. J. Soc. Issues 31, 163e184.
Koepke, Jennifer E., 2000. The failure to breach the blue wall of silence: the circling of the wagons to police perjury.
Washburn Law Rev. 39, 211e242.
Massing, Michael, 2003. Law in a Lawless Land: Diary of a Limpieza in Colombia. The New Press, New York.
Nabokov, Vladimir, 1958. Lolita. Putnam, New York.
Pate, Antony M., Fridell, Lorie A., 1993. Police Use of Force: Official Reports, Civilian Complaints, and Legal
Consequences. The Police Foundation, Washington D.C.
Penglase, Ben, 1994. Final Justice: Police and Death Squad Homicides of Adolescents in Brazil. Human Rights Watch,
New York.
Peratis, Kathleen, 2004. Honoring the Killers: Justice Denied for “Honor” Crimes in Jordan. Human Rights Watch,
New York.
Petrillo, Lisa, 1990. When a Cop Shoots, Who Takes a Closer Look. Here, Unlike Most Cities, an Outside Probe Rarely
Results When Officers Fire. San Diego Union-Tribune (December 21): A1, A 10.
Petrocelli, Daniel, (with Peter Knobler), 1998. Triumph of Justice: The Final Judgment on the Simpson Saga. Crown
Publishers, New York.
Rudovsky, David, 1992. Police abuse: can the violence be contained? Harv. Civ. Rights-Civil Libert. Law Rev. 27,
465e501.
Rummel, R.J., 1994. Death by Government. Transaction Publishers, New Brunswick.
Scheper-Hughes, Nancy, 2006. Death squads and democracy in Northeast Brazil. In: Comaroff, Jean, Comaroff, John L.
(Eds.), Law and Disorder in the Postcolony. University of Chicago Press, Chicago, pp. 150e187.
Schwarz, Philip J., 1988. Twice Condemned: Slaves and the Criminal Laws of Virginia, 1705e1865. Louisiana State
University Press, Baton Rouge.
Simon, David, 1991. Homicide: a Year on the Killing Streets. Fawcett Columbine, New York.
Swindle, Howard, 1993. Deliberate Indifference: a Story of Murder and Racial Injustice. Viking, New York.
Timmermans, Stefan, 2006. Postmortem: How Medical Examiners Explain Suspicious Deaths. University of Chicago
Press, Chicago.
Toobin, Jeffrey, 1996. The Run of His Life: the People vs. O.J. Simpson. Random House, New York.
Wilson, Wayne, 1996. Good Murders and Bad Murders: a Consumer's Guide in the Age of Information. University Press
of America, Lanham (revised edition; first ed., 1991).