The Criminal Community of Victims and Perpetrators:
Cognitive Foundations of Citizen Detachment From Organized
Violence in Mexico
Human Rights Quarterly, Volume 38, Number 4, November 2016, pp. 1038-1069
Published by Johns Hopkins University Press
For additional information about this article
Access provided at 9 Sep 2019 22:05 GMT from University of California, San Diego
HUMAN RIGHTS QUARTERLY
Human Rights Quarterly 38 (2016) 1038â€“1069 Â© 2016 by Johns Hopkins University Press
The Criminal Community of Victims and
Perpetrators: Cognitive Foundations of
Citizen Detachment From Organized
Violence in Mexico
After its successful transition to democracy, Mexico has experienced an
epidemic of organized societal violence known as the drug war that, to
date, has caused well over 100,000 casualties. Most of this violence has
been consigned to oblivion, without proper investigation or prosecution.
Victims have been organizing and protesting, yet ordinary citizens have
remained quiet, except for two short lived waves of nationwide protest.
As I hypothesize, a primary reason for their acquiescence is cognitive. The
framing of organized violence as a self-contained war among criminals
(â€œbounded violenceâ€) erodes the attitudinal foundation of citizen solidarity
and sympathy with the victims of injustice. I explore the cognitive foundations of citizen attitudes towards victims on the basis of original data from
the Mexican 2013 National Survey on Organized Violence. Logistic regression analysis confirms the expected framing effect. Even when controlling
for alternative explanations, such as personal proximity to violence and
social proximity to its victims, the notion of bounded violence within a
criminal community induces citizens to view its victims with indifference.
* Andreas Schedler is Professor of Political Science at the Center for Economic Teaching
and Research (CIDE) in Mexico City. His most recent book publications are The Politics of
Uncertainty: Sustaining and Subverting Electoral Authoritarianism (Oxford University Press,
2013) and En la niebla de la guerra: Los ciudadanos ante la violencia criminal organizada
[In the Fog of War: Citizens and Organized Criminal Violence in Mexico] (CIDE, 2015).
2016 The Criminal Community of Victims and Perpetrators 1039
This article examines the logic of political solidarity in epidemics of organized
criminal violence that reveal structural failures of both public security and
public justice. Political solidarity is the willingness to assist the victims of
severe and systematic injustice within a political community. It is a perennially scarce resource. Citizens will often escape its demands by denying
its basic moral requisite: sympathy with the victims. In this paper, I analyze
one cognitive mechanism which I expect to undercut citizen sympathies
with victims of organized criminal violence: their symbolic identification
with perpetrators. When victims are seen to belong to the same imagined
community as their assassins, when they are seen as criminals killed by
criminals, their fellow citizens, I hypothesize, may easily deny them their
sympathies. They may easily retreat into a position of detachment from the
horrors of war.
Generally speaking, the identification of victims with perpetrators may be
based on social class or ethnicity. In this paper, I analyze its working in an
extreme case of simultaneous collapse of public security and justice, which
is, the so-called drug war that has been unfolding in Mexico over the past
fifteen years. Here, the symbolic identification of victims and perpetrators
has been based on the presumption of common organizational membership. The notion of drug violence has offered an easy narrative of citizen
detachment, as it frames the war as an exclusive affair among members of
criminal organizations (bounded violence).
I explore the cognitive foundations of citizensâ€™ attitudes towards the
victims of drug violence on the basis of original data from the Mexican
2013 National Survey on Organized Violence. Logistic regression analysis
confirms the expected framing effect. Even when controlling for alternative
explanations, such as personal proximity to violence and social proximity
to its victims, the â€œframes of warâ€ mold the perception of its victims.1
the extent that citizens hold criminal violence to be a self-contained affair
among criminals, they tend to sympathize less with the victims and their
II. SOURCES OF POLITICAL SOLIDARITY
Whenever severe violations of human rights such as torture, murder, and
disappearance spread within a modern nation state, ordinary citizens face
1. Judith Butler, Frames of War: When Is Life Grievable? (2010)
1040 HUMAN RIGHTS QUARTERLY Vol. 38
discomforting questions (even when they avoid them).2
What do they know?
What should they know? What do they want to know? How much do they
care? How do they relate to victims and perpetrators? Whom do they sympathize with? Which acts of violence do they condemn, which condone?
What can they do to stop or alleviate the suffering of victims? What do they
actually do? These questions touch the essence of political solidarity, which
is: citizensâ€™ willingness to assist the victims of severe and systematic injustice.
Whether they recognize it or not, citizens face such questions whenever
direct human intervention by either public or private actors produces suffering and death on a massive scale: under repressive dictatorship, in the face
of genocide, in civil war, and in epidemics of criminal violence.
A. Attitudinal Foundations
More often than not, citizens fail to meet the high demands of political
solidarity. They fail their responsibility to protect their co-citizens. They let
atrocities run their course, know little, care little, and do little about the
fate of victims. Even worse, citizens often act in opportunistic ways, taking
personal advantage of acts of injustice; in a complicit manner, encouraging
victimization by omission or commission; or collaboratively, participating
directly in the organization of violence and injustice. Still, the history of
violence and injustice is also a history of resistance to violence and injustice.
In contexts of harsh repression, acts of solidarity carry extreme risks. They
are no less than heroic and nevertheless do take place, even under the most
What explains empirical variations in political solidarity? When are
citizens able and willing to mobilize the moral resources of solidarity? By
contrary, when do they refuse to assist co-citizens who suffer from injustice?
Since most of contemporary political science comprehends political actors
as acting under the primacy of self interest, the discipline has paid scarce
attention to moral interventions in the face of injustice. Modern political science is not a science of solidarity, but one of utility. Reflections on political
solidarity have been largely left to normative political theorists,4
2. I use the notion of â€œhuman rights violationsâ€ in a broad way that covers public as
well as private perpetrators. See, e.g., Tristan Anne Borer, A Taxonomy of Victims and
Perpetrators: Human Rights and Reconciliation in South Africa, 25 Hum. Rts. Q. 1088
(2003). Similarly, I use the notion of â€œcitizensâ€ in a wide, minimalistic way, as members
of a modern territorial state (which may be dictatorial or failing), rather than carriers of
rights in an effective democratic polity.
3. See Arno Lustiger, Rettungswiderstand: Uber Die Judenretter in Europa WÃ¤hrend der NS-Zeit
4. See, e.g., Hauke Brunkhorst, Solidarity: From Civic Friendship to a Global Legal Community
(Jeffrey Flynn trans. 2005); Richard Rorty, Contingency, Irony, and Solidarity (1989).
2016 The Criminal Community of Victims and Perpetrators 1041
of totalitarian systems,5
and to a certain extent to students of social movements, interest groups, and ethnic mobilization.6
Much theoretical reflection and empirical work on the logic of moral
intervention on behalf of victims has taken place in sociology and social
psychology. The language is usually different. Authors rarely frame their work
under the positively connotated heading of solidarity, but rather focus on its
opposites, such as moral disengagement,7
and passive bystanding.9
Despite differences in theoretical approaches and methodological strategies, these strands of literature show broad empirical convergences. Among
other things, they all highlight one basic attitudinal prerequisite of solidarity:
sympathy with victims.
B. Sources of Sympathy
Conceptual logic as well as empirical evidence suggest that sympathetic
attitudes, â€œfeelings of pity and sorrow for someone elseâ€™s misfortune,â€10 are
contingent on two cognitive factors that form part of the conceptual core of
political solidarity: the recognition of victims as â€œone of usâ€ (identification)
and the recognition of victims as victims (perceived injustice).
Identification: the concept of solidarity is bounded. It describes reciprocal
horizontal obligations among members of a community of equals. It differs
from neighboring concepts such as benevolence which refers to relations
among individuals, charity which refers to hierarchical relations between
donors and recipients, or clemency which refers to relations between victors
and losers. The original Roman concept of solidarity was a legal concept. It
described reciprocal financial obligations among members of a community
and their joint liability for personal debts. â€œOne for all, all for one. Everyone
assumes responsibility for anyone who cannot pay his debt.â€ 11 The modern
5. See, e.g. Peter Longerich, â€œDavon Haben wir Nichts Gewusst!â€ Die Deutschen und die Judenverfolgung 1933â€“1945 (2006).
6. For a synthesis, see Sally J. Scholz, Political Solidarity 2â€“16 (2008).
7. See, e.g. Albert Bandura, Claudio Barbaranelli, Gian Vittorio Caprara, & Concetta
Pastorelli, Mechanisms of Moral Disengagement in the Exercise of Moral Agency, 71 J.
Personality & Soc. Psychology 364 (1996).
8. See. e.g. Israel W. Charny, A Classification of Denials of the Holocaust and Other
Genocides,â€ 5 J. Genocide Res. 11 (2003); Stanley Cohen, States of Denial: Knowing about
Atrocities and Suffering (2001).
9. See, e.g. Amanda B. Nickerson, Danielle Mele, & Dana Princiotta, Attachment and
Empathy as Predictors of Roles as Defenders or Outsiders in Bullying Interactions, 46 J.
School Psychâ€™gy 687 (2008); Pozzoli, Tiziana & Gianluca Gini, Why Do Bystanders of
Bullying Help or Not? A Multidimensional Model, 33 J. Early Adolescence 315 (2013).
10. â€œSympathy,â€ Oxford Dictionary of English (3d ed. 2014), available at http://www.oxfordreference.com.
11. Brunkhorst, supra note 4, at 2.
1042 HUMAN RIGHTS QUARTERLY Vol. 38
concept of political solidarity is a moral concept. It describes reciprocal
moral obligations among strangers who share membership in an â€œimagined
communityâ€ 12 defined by some abstract criterion such as class, occupation,
gender, age, ethnicity, nation, or humanity.
Group membership is often binary. Either you are in or you are out.
Either you qualify as a potential addressee of group solidarity or you do
not. To the extent that politics polarizes between opposing camps of friends
and foes, such dichotomies of belonging tend to map the scope of political
solidarity well. Yet, the relevant social psychological literature on moral
action neither deals with social groups nor with situations of political polarization. It studies individual relations in more ordinary settings in which
group boundaries are fuzzy, groups are internally diverse, and people belong to multiple groups at a time. Accordingly, this literature does not ask
whether spectators categorize victims in a binary fashion as either insiders
or outsiders of some abstract community. Rather, it asks where they place
them along a continuum of proximity versus distance. Just as â€œsocial distanceâ€ between victim and perpetrator increases the probability of criminal
violence,13 â€œpositive attitude toward the victimâ€ 14 increase the probability
of defensive intervention by third parties.
Injustice: solidarity is a response to the suffering of others. Political
solidarity more narrowly, is a response to injustices co-citizens suffer. It
involves the acceptance of â€œpositive duties in response to a perceived injustice.â€ 15 It is the perception of injustice that creates the demand for solidary action. A just world in which everyone gets what he or she deserves,
has no need for political solidarity. Citizens can uphold the â€œdelusion of a
just worldâ€16 either by ignoring acts of injustice (the logic of denial) or by
re-describing them as acts of justice (the blaming of victims). By pleading
blindness, citizens can remain deaf to the calls of solidarity. Alternatively, if
they invert responsibilities by blaming the victims for their own misfortune,
that is, if they hold them to be deserving of punishment rather than worthy
of protection, then their very sense of justice will impel them to side with
perpetrators rather than victims.17
In this article, I wish to draw attention to a cognitive frame that undercuts both sources of sympathy: the identification of victims with perpetrators. If victims appear to belong to the same social group as perpetrators,
12. Benedict Anderson, Imagined Communities: Reflections on the Origin and Spread of Nationalism
13. See, e.g., Gresham M. Sykes & David Matza, Techniques of Neutralization: A Theory
of Delinquency, 22 Am. Sociological Rev. 664 (1957); David A. Grossman, On Killing: The
Psychological Cost of Learning to Kill in War and Society (1995).
14. Pozzoli & Gini, supra note 9, at 320.
15. Scholz, supra note 6, at 6.
16. See, e.g. Melvin J. Lerner, The Belief in a Just World: A Fundamental Delusion (1980)
17. See, e.g,. id.; William Ryan, Blaming the Victim (1971).
2016 The Criminal Community of Victims and Perpetrators 1043
to the same imagined community of criminals, then citizens can easily
deny them recognition as co-citizens as well as recognition as victims. Of
course, sympathy alone is not enough for solidary intervention. Attitudes
do not translate smoothly into action. Citizens must also see effective ways
of helping victims (the expected effectiveness of intervention) and of doing so without exposing themselves to unbearable danger (the perceived
risks of intervention). Situations of organized criminal violence, like other
situations of systemic violence, place tight limits and high risks on solidary
action. Under the shadow of illegitimate violence, it is difficult to see what
individual citizens could possibly do to protect victims and anything they
might be doing is likely to entail considerable threats to their own physical
integrity. However, even if not a sufficient condition of active solidarity,
sympathy towards victims does appear to constitute a necessary condition.
Citizens who are either indifferent or hostile to victims are unlikely to assist
them in their plight.
C. The Triangle of Violence
The classic criminal triangle of perpetrators, victims, and spectators allows
for four ideal-typical alliance structures, which are illustrated in Figure 2:
1. Sympathy: the imagined community between citizens and victims. Citizens
sympathize with victims, identify with their plight, and recognize them as
victims of injustice who are worthy of protection. Typical example: ordinary
citizens in the face of ordinary crime.
2. Complicity: the imagined community of citizens and perpetrators. Citizens
sympathize with perpetrators, identify themselves with their cause, and
recognize them as agents of justice who deserve support. They are either
hostile or indifferent towards victims. Typical example: pro-regime actors who
sympathize with repressive campaigns against â€œthe enemies of the peopleâ€
3. Polarization: the confrontation between communities of perpetrators, victims,
and citizens. In violent conflict among communities, citizens sympathize with
those victims and perpetrators who belong to their own imagined community.
Drawing a sharp line between insiders and outsiders, they side with â€œourâ€
victims (martyrs) and â€œourâ€ perpetrators (heroes) against â€œtheirâ€ victims (who
are deserving) and â€œtheirâ€ perpetrators (who are evil). Typical examples: the
distribution of national sympathies in international war, the distribution of
ethnic sympathies in ethnic war, and the distribution of political sympathies
in internal political war.
4. Detachment: the community of victims and perpetrators. Citizens sympathize
or identify with no one. In their perspective, both perpetrators and victims
belong to a community separate from their own. They are not community
members, but some sort of aliens entangled in extraneous violent encounters.
1044 HUMAN RIGHTS QUARTERLY Vol. 38
Both are barbarians, neither of them merits support. The classic example
is the neutrality of moderate actors in the face of revolutionary warfare. In
such situations of ideological polarization, neutral actors draw an equation
between warring parties: all commit atrocities, none is justified.18 In situations
of nonpolitical violence, neutral actors draw an equation between victims
and perpetrators: they are all criminals, no one is innocent.
The imagined criminal community of victims and perpetrators may
be ascriptive (for example, based on ethnic membership) or voluntary (for
example, based on organizational membership). In public imagination,
ascriptive communities of criminals (such as the various ethnically defined
mafias) are composed of what we may call â€œintrinsic barbarians,â€ while
voluntary communities of criminals (such as drug cartels or terrorist organizations) are composed of what we may call â€œelective barbarians.â€ In the
case of the Mexican drug war, public discourse is about the latter, a war
among self-selected barbarians.
III. ORGANIZED CRIMINAL VIOLENCE IN MEXICO
In Mexicoâ€™s 2000 presidential elections the victory of opposition candidate
Vicente Fox of the conservative National Action Party (PAN) capped a long
process of democratization by elections and ended seven straight decades
of hegemonic rule by the Institutional Revolutionary Party (PRI). Yet as its
fledgling democracy has been struggling to find its way, the country has
suffered a pandemic escalation of violence related to organized crime.
A. The Escalation of Violence
In 2006 after a close and contentious election, PANâ€™s Felipe CalderÃ³n assumed the presidency amid a lingering security crisis. During Foxâ€™s term in
office, violent competition among drug-trafficking organizations (so-called
cartels) provoked more than a thousand homicides per year, and the number
was rising. Although it had not been an issue during the election campaign,
President CalderÃ³n decided to make the fight against drug cartels the defining
policy of his presidency, only to see that fight turn into his termâ€™s defining
failure. During his six years in office, violence escalated both qualitatively
18. See, e.g., Elisabeth Jean Wood, Insurgent Collective Action and Civil War in El Salvador
2016 The Criminal Community of Victims and Perpetrators 1045
In qualitative terms, modes of assassination moved toward demonstrative
cruelty, routinized, and ritualized, including the public display of tortured,
dismembered, and decapitated bodies. In quantitative terms, the number of
annual homicides attributed to criminal organizations shot up from around
2,200 in 2006 to more than 16,600 in 2011. In 2012, drug-related homicides
started to decline. This trend continued through 2014, even though annual
figures of â€œexecutionsâ€ have remained at a level (around 8,000) many times
higher than in the early 2000s (see Figure 1). All these numbers must be
read with great caution, though. The problems that cluster around the task
Figure 1. Number of homicides attributed to organized crime in Mexico,
Sources: For 2001â€“2006: General Attorneyâ€™s Office, cited in Marcos Pablo Moloeznik, â€œMilitarizing Mexicoâ€™s Public Securityâ€ (Washington, dc: National Defense University, Center for
Hemispheric Defense Studies), chds Regional Insights 11 (15 February 2009). For 2007â€“2010:
Presidency of the Republic, â€œDataset of Deaths by Presumptive Criminal Rivalry.â€ For Januaryâ€“September 2011: General Attorneyâ€™s Office, â€œDataset of Deaths by Presumptive Criminal
Rivalryâ€ (http://www.pgr.gob.mx). For October 2011â€“December 2015: Eduardo Guerrero, Lantia
Consultores, â€œDataset of Violence by Organized Crimeâ€ (http://www.lantiaconsultores.com/).
1046 HUMAN RIGHTS QUARTERLY Vol. 38
of compiling accurate data on the violence are massive. Besides, thousands
of people have â€œdisappearedâ€ after being abducted.19
19. The number of disappeared persons is estimated to lie around 22,000. Sandra Jessica
Ley GutiÃ©rrez, Citizens in Fear: Political Participation and Voting Behavior in the Midst
of Violence, Durham, Duke University, Department of Political Science, Ph.D. Dissertation, at Ch. 5 (2014). On disappearances and mass graves (â€œnarcofosasâ€) related
to organized crime, see also Cory Molzahn, Octavio RodrÃguez Ferreira, & David A. Shirk,
Trans-Border Institute, Joan B. Kroc School of Peace Studies, University of San Diego, Drug
Violence in Mexico: Data and Analysis through 2012, at 18â€“19 (2013).
Figure 2. Patterns of citizen identification in the triangle of injustice
a) Sympathy b) Complicity
Note: Gray shadings denote imagined communities (perceived alliance structures)
2016 The Criminal Community of Victims and Perpetrators 1047
B. The Expansiveness of Violence
When confrontations between armed groups within a state cause more than
a thousand â€œbattle-related deathsâ€ per year, conflict scholars speak of â€œcivil
war.â€ At least since 2001, democratic Mexico has experienced levels of
â€œinternal warâ€ that surpass this conventional threshold. Classic conceptions
of civil war require that the parties in conflict are â€œpolitically and militarily
organized, and . . . have publicly stated political objectives.â€ 20 The Mexican drug war is different. It is not a political civil war in which ideological
insurgencies fight to topple state power or transform the political regime.
It rather looks like a prototypical â€œnewâ€ civil war, fought for material gain
and not social justice.21
Yet if this is a war, it is not one but many. Its major lines of conflict run
between criminal enterprises. Many perhaps most, acts of private coercion
are hostile acts within a multilateral war among competing cartels. Yet while
the so-called drug war entails various interacting nonstate conflicts, it also
contains elements of one-sided violence that criminals unleash against civilians. Profit-oriented participation in illicit markets forms only a portion of
organized crimeâ€™s activity. The drug cartels are also massively engaged in
predatory crimes involving unilateral violence against civilians. Organized
homicides have only been the tip of the violent iceberg. As criminal organizations have diversified their activities, the country has seen the dramatic
expansion of kidnapping, human trafficking, and extortion (mafia-like protection rackets).22 In addition, as the cartels wage a guerrilla war against state
20. Nicholas Sambanis, What Is Civil War? Conceptual and Empirical Complexities of an
Operational Definition, 48 J. Conflict Res. 814, 829 (2004).
21. See Luis de la Calle & Andreas Schedler, Political and Economic Civil Wars, 111th Annual Meeting, American Political Science Association (APSA), San Francisco (3â€“6 Sept.
2015). For a critical discussion, see Stathis N. Kalyvas, How Civil Wars Help Explain
Organized Crimeâ€”and How They Do Not, J. Conflict Res. 1 (2015). Seminal texts on
â€œnewâ€ civil wars have been, among others, Hans Magnus Enzensberger, Aussichten auf den
BÃ¼rgerkrieg (1996), Mary Kaldor, New and Old Wars: Organized Violence in a Global Era
(2d ed. 2007); Stathis N. Kalyvas, â€œNewâ€ and â€œOldâ€ Civil Wars: A Valid Distinction?,
54 World Pol. 99 (2001).
22. See, e.g., HÃ©ctor Aguilar CamÃn, Eduardo Guerrero, Alejandro Madrazo, AndrÃ©s Lajous, Jorge
HernÃ¡ndez Tinajero, Joel ChÃ¡vez, & Dante Haro, Informe Jalisco: MÃ¡s allÃ¡ de la Guerra de las
Drogas 93â€“95 (2012), available at https://seplan.app.jalisco.gob.mx/biblioteca/archivo/
play/580; Marcelo Bergman, La Violencia en MÃ©xico: Algunas Aproximaciones AcadÃ©micas, 40 Desacatos 65, 70â€“72 (2012); Carlos Bravo, Marc Grau Vidiella , & Gerardo
Maldonado HernÃ¡ndez , Elecciones, Violencia y Estructura Social (EVES): Base de Datos Integral
de Municipios Mexicanos 33â€“34 (2014), available at http://www.ine.mx/docs/IFEâ€“v2/CDD/
CDD-estructura/DOCS/Informe_Final_TomoI-BaseEVES.pdf; Edgardo Buscaglia, MÃ©xico
Pierde la Guerra, Esquire (Mar. 2010), Panorama EstadÃstico de la Violencia en MÃ©xico (Carlos
Javier Echarri CÃ¡novas ed., 2012), available at http://cedua.colmex.mx/images/PDFs/
echarri/panorama.pdf; OAS Hemispheric Security Observatory, Report on Citizen Security in
the Americas 2012, at 70â€“75 (2012) , available at https://adamblackwell.files.wordpress.
1048 HUMAN RIGHTS QUARTERLY Vol. 38
agents, they participate in a kind of criminal insurgency. They have carried
out numerous attacks against the state, such as the kidnapping, torture,
and murder of security officials and assaults on police stations using hand
grenades and heavy weapons.
Thus the Mexican state is a warring party, too. In theory, it has a
monopoly on the wielding of legitimate violence. In practice, it commits
criminal violence on a large scale. International human rights groups agree
in ascertaining â€œwidespreadâ€ human rights violations perpetrated by security agents. In part, these violations are expressions of state abuse. They are
the unintended but inevitable consequence of acting with brute force, little
actionable intelligence, and no oversight in an â€œirregular warâ€ 23 characterized by endemic problems of information. In part, illegal state violence is
a symptom of partial state collusion. Between January 2008 and November
2012, more than 2,500 police officers and more than 200 military personnel
were murdered by criminal organizations. Yet in numerous instances, public
officials have collaborated with criminal organizations.24
Even if much of the violence seems to be taking place among and
against members of the so-called drug cartels, the actual logic of violence
on the ground appears to be more complicated. The boundaries between
private and public criminals are blurred and the logic of victim selection
is open ended. Armed actors kill for a broad variety of reasons, which are
often petty and opaque. As their crimes flourish in a climate of impunity, we
almost never get to know for certain who does what to whom and why.25
23. Stathis N. Kalyvas, Civil Wars, in Oxford Handbook of Comparative Politics 416, 423 (Carles
Boix & Susan Stokes eds., 2007).
24. Casualty figures from Molzahn, RodrÃguez & Shirk, supra note 19, at 30. On state abuse
and collusion, see, e.g., Amnesty International, Nuevos Informes de de Derechos Humanos A
Manos del EjÃ©rcito (2009), available at http://amnistia.org.mx/abusosmilitares/informe.pdf,
Amnesty Intâ€™l, Known Abusers, butVictims Ignored: Torture and Ill-Treatment in Mexico (2012),
available at http://www.amnestyusa.org/sites/default/files/mexico_torture_-_report_eng.
pdf, Amnesty Intâ€™l, Enfrentarse a una Pesadilla: La DesapariciÃ³n de Personas en MÃ©xico (2013),
available at http://amnistia.org.mx/nuevo/wp-content/uploads/2015/04/Enfrentarse_a_
19, Silencio forzado: El Estado, CÃ³mplice de la Violencia contra la Prensa en MÃ©xico: Informe
2011 (2012), available at https://www.ifex.org/mexico/2012/03/20/article19informe2011.
pdf, Article 19, Doble Asesinato: La Prensa Entre la Violencia y la Impunidad: Informe 2012
(2013), available at http://articulo19.org/informe2012/, Human Rights Watch, (HRW),
Uniform Impunity: Mexicoâ€™s Misuse of Military Justice to Prosecute Abuses in Counternarcotis
and Public Security Operations (2009); HRW, Neither Rights Nor Security: Killings, Torture,
and Disappearances in Mexicoâ€™s â€œWar on Drugsâ€ (2011); HRW, Mexicoâ€™s Disappeared: The
Enduring Cost of a Crisis Ignored (2013). On information problems in irregular wars, see
Stathis N. Kalyvas, The Logic of Violence in Civil War (2006).
25. Andreas Schedler, En la niebla de la guerra: Los ciudadanos ante la violencia criminal organizada
2016 The Criminal Community of Victims and Perpetrators 1049
C. The Boundaries of Violence
While the dynamics of organized criminal violence have been opaque and
expansionary, prevailing public discourse has treated it as transparent and
self-contained. In political debate and academic analysis inside as well as
outside the country, the epidemic of organized violence is conventionally
described as a â€œdrug warâ€ among criminal organizations that compete for
the control of transnational and local drug markets. It appears as a kind of
extraterritorial battle in which life-long members of a homogeneous, selfenclosed social group (called â€œcriminalsâ€ or â€œdelinquentsâ€) torture, kill, and
abduct each other. In this predominant perspective, which has been shaken
but not shattered in more recent years, violence is bounded. It is an exclusive
affair among professional criminals. In particular under the presidency of
Felipe CalderÃ³n (2006 to 2012), official discourse produced and reproduced
such a simple message of selective, self-contained criminal violence: the
war is about bad guys killing each other.
More than 90 percent of the homicides and executions, as we have been classifying them, derive from the fight of some cartels against others.26
More than 90 percent of the people who have died had connections with one
or the other band, be it that they distribute or transport drugs or because they
are retailers, people who have a small grocery shop, or repair tires, or are taxi
drivers, and who are known in their communities to be in the business of distribution, which is precisely the market those groups fight over.27
The notion of bounded criminal violence has not been exclusive to top
government circles. Lower-level officials have embraced it, too. Victim families have given countless testimonies of state officials who treated them with
disdain and refused to investigate their cases under the speculative suspicion
that their murdered or disappeared family member had been connected to
criminal groups.28 In the public sphere too, even media outlets critical of
the government like the weekly Proceso habitually describe the presumptive
26. Jorge Ramos, Muertes de Civiles son las Menos: fch, El Universal, 16 Apr. 2010, available at http://www.eluniversal.com.mx/notas/673331.html..
27. Felipe CalderÃ³n, El Presidente Felipe CalderÃ³n en la InauguraciÃ³n de la 72 ConvenciÃ³n
Bancaria MÃ©xico Ante la Crisis Financiera Mundial (19 Mar. 2009), available at http://
28. See, e.g., Amnesty Intâ€™l, Known Abusers, but Victims Ignored, supra note 24, at 12; Amnesty
Intâ€™l, Enfrentarse a una Pesadilla, supra note 24, at 6, John Gibler, Tinta Contra el Silencio, in Entre las Cenizas: Historias de Vida en Tiempos de Muerte 127, 139 (Marcela Turati &
Daniela Rea eds., 2012); Human Rights Watch (2013), supra note 24, at 2, 5; Marcela
Turati, Tras las pistas de los desaparecidos, in Entre las cenizas: Historias de vida en tiempos
de muerte 101, 107â€“08 (Marcela Turati & Daniela Rea eds., 2012).
1050 HUMAN RIGHTS QUARTERLY Vol. 38
victims of organized violence as criminal subjects killed by their criminal
peers. Treated as routine events (explicable and comprehensible), their
deaths are distinguished from those few (deplorable and exceptional) cases
of â€œinnocentâ€ passers who are hit by the random violence of stray bullets.29
Now, should we consider the notion of self-contained criminal violence
to be an ideological proposition or a faithful reflection of reality? Are we
talking about a distortive cognitive frame or about a set of plausible beliefs?
As a matter of fact it is both. The idea of bounded violence is grounded in the
widely accepted fact that a large share of organized violence in Mexico takes
place within and among criminal organizations. We do not know however,
how large that â€œlarge shareâ€ actually is. Herein lies the ideological core of
the frame: it posits certainty in the midst of opacity and it excludes from
its field of vision whole categories of victims (above all, civilians without
criminal affiliations) as well as whole categories of perpetrators (above all,
IV. CITIZEN RESPONSES TO VIOLENCE
Societal violence is a problem of public security that demands state action.
In addition, it is a problem of political justice that demands citizen action.
A. Political Injustice
The countless acts of murder, torture, and disappearance represent an atrocity
on a massive scale. But they also represent an injustice on a massive scale.
Even though they are not planned and executed by the central state, they
are tolerated by a state which has renounced the effective judicial prosecution of organized violence. The criminal war has escalated in a context of
near complete impunity. According to figures collected by Human Rights
Watch, between December 2006 and January 2011, Mexican authorities
counted 35,000 homicides they attributed to organized crime, of which
997 led to formal criminal investigations (2.8 percent), of which 343 led to
formal criminal accusations (0.9 percent), of which twenty-two led to firm
convictions (0.06 percent).30 For all practical reasons, the rate of successful
persecution is zero.
29. See Julieta Lemaitre, Civilization, Barbarism, and the War on Drugs: The Normalization
of Violent Death in Mexico and Colombia, 109th Annual Meeting, American Political
Science Association (APSA), Chicago, IL (29 Aug.â€“1 Sept. 2013).
30. HRW, Neither Rights Nor Security, supra note 24, at 15.
2016 The Criminal Community of Victims and Perpetrators 1051
The Mexican drug war thus appears as a rather extreme instance of a
political illness that afflicts most countries in contemporary Latin America:
a vicious circle of societal violence and judicial impunity. Homicide rates
in the region have been â€œconsistently and significantly higherâ€ 31 than those
recorded in Asia and Europe. Supplying 8 percent of the world population,
Latin America supplies 42 percent of global homicide victims.32 In 2010,
only five out of thirty-two countries of Latin America and the Caribbean
reported annual homicide rates below the level considered as endemic by
the World Health Organization (ten homicides per 100,000 inhabitants).33
Large territories of the region are closer to the rule of lawlessness than the
rule of law. In the Americas (which includes Canada and the US and thus
embellishes these figures), â€œless than a quarter of homicides lead to convictions.â€34 In Europe, more than 80 percent do.35
Endemic levels of homicide with impunity indicate a double failure
of the state: its initial failure to uphold the right to life and its subsequent
failure to persecute â€œthe worst of crimesâ€ 36 and bring perpetrators to justice. They indicate a failure of protection as well as a failure of justice. The
political injustice they impose differs from the political injustice committed
by repressive dictatorship or ethnic war. Responsibility for the commission
of violence and for the denial of justice is more diffuse. The motives and
the identity of victims are more opaque. Nevertheless, under conditions of
structural impunity, criminal societal violence calls out for citizen solidarity with victims in similar ways as political violence by state actors does.
B. Strategies of Survival
How have Mexican citizens responded to the epidemic of death and injustice
in their fledgling democracy? The direct and indirect victims of crime have
responded in many ways. Individually, they have mostly sought refuge in
exit strategies, such as changing their place of residence (internal and international migration), shutting down business operations in the face of extortion, or shutting themselves up in their homes.37 Collectively, their responses
31. United Nations Office on Drugs and Crime (UNODC), Global Study on Homicide 2013:
Trends, Contexts, Data 35 (2014).
32. MoisÃ©s NaÃm, La Gente mÃ¡s Asesina del Mundo, El PaÃs 15 Dec. 2012, available at http://
33. Antigua and Barbuda, Argentina, Chile, Haiti, and Uruguay. OAS, supra note 22, Table
34. UNODC, supra note 31, at 94.
36. Id. at 5.
37. Instituto Nacional de EstadÃstica y GeografÃa (inegi), Encuesta Nacional de VictimizaciÃ³n y
PercepciÃ³n sobre Seguridad PÃºblica (envipe) 2014: Principales Resultados 39 (2014), available
1052 HUMAN RIGHTS QUARTERLY Vol. 38
have been bifurcated. Some have turned to peaceful mobilization, others to
armed resistance. On the side of civil protest across the entire geography of
organized violence, numerous local civic movements and associations have
been formed by victims of violent crime and their families and friends.38
On the side of societal counter-violence, paramilitary self-defense forces
have risen in more than a dozen Mexican states. They have gained most
prominence in 2013 in MichoacÃ¡n, when the federal government launched
a large scale intervention attempting to regain territorial control through a
mixture of military action, institution building, and social policy.39
The question of political solidarity, however, does not concern the
victims of violence, but their co-citizens who have not (yet) been directly
affected by criminal civil war. The key test of political solidarity is not the
solidarity among victims, but solidarity towards victims. How then, have
ordinary citizens been responding to the civil war that has been unfolding
on their television screens? The picture has been mixed.
In essence, for most of the time over the past fifteen years, citizens have
adopted the role of passive onlookers. Overall, there has been â€œlimited organized civic engagement on the issue of security.â€ 40 Ordinary citizens have
been trying to accept the new realities of war as normal and to carry on their
daily lives as smoothly as possible. For most of the time we have seen few
public displays of sorrow or anger, little serious debate, and no sustained
pressure on authorities. Rather than a nation of concerned citizens mobilizing their energies towards the construction of a decent system of justice, we
have seen a nation of bystanders who have been quick to absorb the atrocious realities of civil war into ordinary language and statistical routines.41
38. Under the heading of â€œsites for peaceâ€ (sitios por la paz), the webpage of the Movement
for Peace with Justice and Dignity offers a collection of links to like-minded associations
39. See, e.g., Kimberly Heinle, Octavio RodrÃguez Ferreira, & David A. Shirk, Drug Violence in
Mexico: Data and Analysis through 2013, at 46â€“47 (2014).
40. Steven Dudley & Sandra RodrÃguez, Civil Society, the Government and the Development
of Citizen Security, San Diego: University of San Diego, Working Paper Series on Civil
Engagement and Public Security in Mexico 5 (2013).
41. On civil societyâ€™s responses to organized criminal violence in Mexico, see Dudley &
RodrÃguez, supra note 40; Vanessa Job, La Resistencia CibernÃ©tica, in Entre las Cenizas:
Historias de Vida en Tiempos de Muerte 147 (Marcela Turati & Daniela Rea eds., 2012); Lilian Paola Ovalle, ImÃ¡genes Abyectas e Invisibilidad de las VÃctimas: Narrativas Visuales
de la Violencia en MÃ©xico, 164 El Cotidiano 103 (2010); Ley GutiÃ©rrez, supra note 19,
ch. 5; Octavio RodrÃguez Ferreira, Civic Engagement and the Judicial Reform: The Role
of Civil Society in Reforming Criminal Justice in Mexico, San Diego: University of San
Diego, Working Paper Series on Civil Engagement and Public Security in Mexico (2013);
Reynaldo T. Rojo-Mendoza, From Victims to Activists: Crime Victimization, Social Support, and
Political Participation in Mexico (2013); Lauren Villagran, The Victimsâ€™ Movement in Mexico,
San Diego: University of San Diego, Working Paper Series on Civil Engagement and
Public Security in Mexico (2013). For a general argument on the causal role of crime
victimization in political participation, see Regina Bateson, Crime Victimization and
Political Participation, 106 Am. Pol. Sci. Rev. 570 (2012).
2016 The Criminal Community of Victims and Perpetrators 1053
C. Waves of Mobilization
However, there have been two notable instances of public mobilization
in support of victims of organized violence. In Spring 2011, existing local
victimsâ€™ movements acquired national visibility. After his son had been killed
by local police officers, poet Javier Sicilia founded the nationwide Movement
for Peace with Justice and Dignity. Several thousand citizens accompanied
him at the rallying points of his tours across Mexico. His movement drew
much media attention, led to first-time encounters between victims and the
president, and inspired the Law of Victims, an expansive, aspirational compilation of victimsâ€™ rights.42 Arguably, the movementâ€™s biggest achievement
was symbolic. By lending victims a public voice and by fighting for their
symbolic recognition as victims, it changed the terms of political discourse.
Thanks to the grieving poet, facile presumptions of guilt towards the victims
of violence have become hard to sustain in public.
As soon as Sicilia ended his series of marches, his movementâ€™s concerns
dropped from public debate. In the subsequent year, citizens voted Felipe
CalderÃ³nâ€™s National Action Party out of the presidential office. During the
entire election campaign though, parties and candidates had essentially
remained silent on the ongoing war. After taking office in December 2012,
the new government under Enrique PeÃ±a Nieto tried to push the problem
off the political agenda. However, in the autumn of 2014, it returned with
On 26 September 2014, in the town of Iguala, Guerrero, three hours
Southwest of Mexico City, the municipal police stopped protesting students
from the rural teachersâ€™ college of nearby Ayotzinapa and put them under
gunfire. Local hit men joined the shooting spree that left six persons dead. In
continuation, forty-three students were arrested by the police and according
to the official reconstruction of the crime, handed over to the local cartel,
murdered, mutilated, and burnt on a dumping place. In response, Mexico
experienced an unprecedented wave of solidarity. Teachers and students
went to strike. Huge protest demonstrations took place in major cities all
over the country.
It did not take very long though, for the country to get â€œback to normal.â€
At the beginning of 2015, the government tried to bring closure to the criminal investigation of the case. Although huge doubts still linger over basic
facts,43 protests have tended to fizzle out. More and more, they have been
42. The documentary film Luisa Riley, Javier Sicilia: En la Soledad del Otro (Canal 22, 2013),
available at https://www.youtube.com/watch?v=XujUvo2EduA, reconstructs the movementâ€™s cathartic first months.
43. In particular, it appears highly implausible that the cremation of forty-three bodies, more
than two tons of human flesh and bones, could have taken place under the conditions
described by official witness declarations. See, e.g., Centro Nacional de ComunicaciÃ³n
Social, CientÃficos Desmienten a PGR por Quema de Normalistas (11 Dec. 2014), available at http://www.cencos.org/comunicacion/cientificos-desmienten-a-pgr.
1054 HUMAN RIGHTS QUARTERLY Vol. 38
confined to the victimsâ€™ friends and family and to a hard core of disruptive
protesters in Guerrero and neighboring Oaxaca.
D. The Logic of Detachment
Even though injustice runs deep, the obstacles to solidarity with the victims
of organized violence are high. The sheer numbers of victims exceed human
emotional capacities. With few exceptions, both victims and perpetrators
have remained anonymous, with the media reporting aggregate figures,
rather than reconstructing individual cases. Responsibilities are diffuse and
opaque, which leaves potential protests without clear addressees. Under
the shadow of retaliatory violence, the risks of collective action are high. In
the face of unresponsive bureaucracies, its prospects of success are meager.
I do not deny the potential weight of these and other inhibiting factors.
Here though, I wish to explore one cognitive factor I expect to undermine
citizen sympathies towards victims: the idea of self-contained violence that
equates victims with perpetrators by placing both into the same category
of criminal actors.
Why should we expect the cognitive frame of bounded criminal violence
to undermine citizen solidarity towards the victims of violence? The notion
of â€œbad guys killing bad guysâ€ does not preclude human pity nor does it
invalidate demands for justice. As human beings, even victims who had been
cruel and unrepentant assassins deserve our pity. As citizens, all victims,
regardless of their criminal past, deserve the full protection of the law. As I
hypothesize though, the notion of bounded violence effectively undermines
citizen solidarity with victims by undermining its main cognitive prerequisite:
sympathy with victims. It presupposes that the boundary line that separates
combatants (criminals) from noncombatants (civilians) is crystal clear, while
the boundary line that separates perpetrators from victims is fuzzy. Criminals
form an imagined community distant from, or even outside of, society that
conceives itself as innocent. Both perpetrators and victims belong to a criminal community whose members are guilty of whatever happens to happen
to them. The armed conflict runs among criminal organizations who supply
the assassins and the corpses. Decent citizens have nothing to fear as long
as they stay out of their business. It is a war â€œamong them,â€ not â€œagainst
us.â€ By blurring one social boundary (between perpetrators and victims) and
deepening another one (between criminals and decent citizens), the idea of
bounded criminal violence undermines both sources of sympathy: the status
of victims as rightful members of the community as well as their status as
victims. In essence, they appear as voluntary victims who have chosen their
fate and placed themselves outside the community of decent citizens when
they entered the world of delinquency.
2016 The Criminal Community of Victims and Perpetrators 1055
V. EMPIRICAL SOURCES OF INDIFFERENCE
By framing the war as a kind of external war among the voluntary members
of the fraternal community of organized criminals, citizens are able to retreat
to a position of detached observers. To the extent that citizens believe in
the boundedness of criminal violence, they should be inclined to maintain
a detached attitude of indifference towards its victims.
I test these empirical expectations on the basis of the Mexican 2013
National Survey on Organized Violence (ENVO) that strives to reconstruct
citizen attitudes towards the main actors of organized violence under conditions of criminal civil war: perpetrators and victims, state and civil society.
ENVO is a nationally representative face-to-face survey which was carried
out in Mexico from 26 October through 30 November 2013 among adult
citizens (older than age 18). Its 2,400 interviewees were chosen through
multistage sampling based on election precincts as defined by the Federal
Electoral Institute (IFE).44 The national sample was stratified by five levels of
municipal violence (average municipal homicide rates from 2009 to 2011).
Designed by the author and jointly sponsored by the National Council of
Science and Technology (CONACYT) and IFE it was implemented by the
survey firm Data OpiniÃ³n PÃºblica y Mercados (OPM). The surveyâ€™s overall
margin of error is plus or minus 2 percent.45
Note that the survey was taken in between the two big waves of public
solidarity with victims. About two years before, Javier Siciliaâ€™s movement
had shaken up public assumptions about guilty victims. About a year after,
the mass murder of Iguala shook up public assumptions about the innocent
state. Had the survey been taken in early 2011 or before, it would probably
have shown more widespread beliefs in the bounded nature of criminal
violence and lower levels of sympathy with its victims. Had the survey been
taken in late 2014 or afterwards, it would probably show the inverse: less
widespread beliefs in the criminal selectivity of violence and higher levels
of sympathy towards its victims. However, even if aggregate levels of public
attitudes towards organized violence and its victims are time-bound, the
causal relationships I wish to test should hold nevertheless.
44. Since 4 April 2014: National Electoral Institute (INE).
45. The national population survey was complemented by an elite survey (N = 629) among
high-level representatives of six groups: government, parties, media, academia, civil
society, and business. The integrated dataset of both surveys is publicly available (in the
Spanish original as well as in English translation) at the CIDE Data Archive for Applied
Research in the Social Sciences (BIIACS), available at http://biiacs-dspace.cide.edu/
1056 HUMAN RIGHTS QUARTERLY Vol. 38
A. ATTITUDES TOWARDS VICTIMS
To capture the sympathies ordinary citizens harbor towards the victims of
organized violence, I use four survey items from ENVO :
1. Perspective taking: Are respondents able to adopt the virtual perspective of
victims? Can they imagine themselves falling victim of organized violence?
Or do bad things only happen to others? Do they picture victims as an alien
category of persons to which they themselves will never belong to? Survey
item: the subjective probability of organized crime sending someone to kill
the respondent or a family member.
2. Empathy: How close do citizens feel to the fate of victims? To what extent
have they allowed themselves to be emotionally touched by the fate of
individual victims of organized violence? Survey item: the ability to recall
a case of murder or disappearance the respondent found moving (â€œOutside
the circles of people you know personally, do you remember some person
whose [murder or disappearance] has moved you in particular?â€).
3. Silence: Do respondents care enough about the war and its victims to talk
about them? Or do they wish to push them outside the agenda of public
debate and private conversation? Survey item: â€œThere are so many good things
in Mexico. We should stop talking so much about violence.â€
3. Movement support: How do citizens view civic mobilization by victim
families who seek justice for their dead and knowledge about their disappeared? Survey item: â€œHow much do you identify with the victims who
Figure 3 displays the corresponding frequency distributions.46 By late
2013, almost half of Mexican citizens held it to be â€œvery unlikelyâ€ that
criminal organizations put them on their death lists (47.2 percent). The other
half, however, did assign positive probabilities to this gruesome eventuality.
Ten percent even held it to be â€œvery likelyâ€ (9.9 percent). At the same time,
only one in six respondents recalled some case of murder or disappearance
they found particularly moving (17.1 percent). Almost two thirds expressed
their desire to cover the war with the mantle of silence (63.6 percent).
Only one in fifteen citizens identified â€œvery muchâ€ with civic associations
formed by victims of criminal violence (6.7 percent), well over a third â€œnot
at allâ€ (37.9 percent).47 These data lends credence to the notion that ordinary citizens were (at least before the shocking events of Iguala) inclined to
keep the victims of organized criminal violence at a distance. Though many
could imagine falling victim to organized crime themselves, they remained
emotionally distant to the actual victims and politically distant to their civic
organizations. A solid majority of the citizenry preferred looking elsewhere.
46. For descriptive statistics, see Table B, Appendix.
47. Both variables are moderately correlated (r = .119, p = .000, N = 2241).
2016 The Criminal Community of Victims and Perpetrators 1057
Perspective taking: â€œHow likely do
you think it is that organized crime
sends someone to kill you or some
member of your family?â€
Empathy: â€œOutside your circle of
personal acquaintances, do you
remember some case of murder or
disappearance you found particularly moving?â€
Agenda setting: â€œThere are so many
good things in Mexico. We should
stop talking so much about violence.â€
Movement support: â€œHow much do
you identify with the victims who
Figure 3. Sympathy with victims
B. The Perceived Boundaries of Violence
In my theory of framing effects, the crucial â€œcausal mechanismâ€ that accounts
for variations in citizen attitudes towards victims is cognitive. I expect the
perceived boundaries of criminal violence to induce subjective distance
towards its victims. My initial question is descriptive: to what extent has the
hypothesis of bounded criminal violence found acceptance among Mexican
1058 HUMAN RIGHTS QUARTERLY Vol. 38
citizens? ENVO attempts to measure the perceived boundedness of criminal
violence in an indirect way. It asks respondents whether they believe they
can protect themselves by leading an honest life and staying out of the way
of criminal organizations: â€œTalking about murders attributed to organized
crime, how much do you agree with the following statement: As long as
you do not get involved with them, nothing happens to you.â€ 48
Violence is indiscriminate (unbounded) when it threatens to touch everybody regardless of what they do or who they are. Violence is selective
(bounded) when it targets those who say or do certain things, or omit saying
or doing certain things. Only in instances of bounded violence, citizens
can protect themselves by collaborating with the dominant group or by
presenting themselves as neutral in the battle between contending parties.49
In political civil war, neutrality is often not a viable option. In particular
in contested war zones, warring parties demand active collaboration, not
passive observation.50 In the face of organized crime by contrast, neutrality
appears as a reasonable self-protective strategy. To the extent that criminal
violence is selectively and exclusively committed by criminals against criminals, innocence should save citizens from its wrath. Living an ordinary life,
keeping their distance, staying clean, and staying out should protect them
from attracting the lethal attention of criminal organizations.
As Figure 4 shows, at the time of the survey, more than 60 percent of
respondents either showed full confidence (29.5 percent) or some confidence
(33.9 percent) in their capacity to protect themselves by not meddling with
the criminal world. About a fifth expressed some disagreement (21.1 percent). Only one eighth (12.9 percent) plainly rejected the idea that â€œnothing
happensâ€ as long as one keeps pretending that nothing happens. At that moment, the idea of bounded criminal violence thus was clearly majoritarian.
It was not consensual, though. It did show some variance which allows us
to explore the causal question about its consequences for citizen attitudes
C. The Proximity of Violence
How do people form their attitudes towards victims of criminal civil war?
Certainly, these attitudes derive from a complex process in which the â€œframes
of warâ€ constitute only one causal factor among others.51 The Mexican Na48. In Spanish, the item phrasing contains an ambiguity that is hard to translate. â€œMientras
uno no se meta con ellos, no pasa nadaâ€ implies two things: Nothing happens as long
as you do not join them . . . and as long as you do not get in their way.â€
49. See Kalyvas, supra note 23.
50. Id. at 226â€“32.
51. Butler, supra note 1.
2016 The Criminal Community of Victims and Perpetrators 1059
tional Survey on Organized Violence allows us to test for two sets of rather
obvious alternative explanations: the proximity of violence and the social
proximity of victims. The former has at least two components: personal
experiences of victimization and geographic proximity to violence.
Victimization: It seems reasonable to expect that personal experiences
of victimization by criminal organizations change personal stances towards
victims. Citizens who have experienced cases of assassination or disappearance inside their families, within their circles of friends, or acquaintances are
likely to be more sympathetic to victims than those who have been spared
the chilling touch of organized violence. To measure degrees of victimization
by organized crime, I constructed an aggregate â€œindex of victimizationâ€ that
adds experiences of victimization inside the family (extortion, murder, and
disappearance) as well as within the wider circle of friends and acquaintances
(murder, disappearance, orphanage, and emigration).52
Survey question: â€œTalking about murders attributed to organized crime, how much do you
agree with the following statement: As long as you do not get involved with them, nothing
happens to you.â€
Figure 4. Bounded criminal violence
52. For more precise descriptions of this as well as all other indices and variables, see Table
A, Appendix; for descriptive statistics, see Table B.
1060 HUMAN RIGHTS QUARTERLY Vol. 38
Distance to Violence: Organized violence in Mexico is not generalized, but territorially concentrated at entry and exit points and along the
transport routes by which drugs move transnationally. Between 2009 and
2011, less than 10 percent of Mexicoâ€™s 2,453 municipalities experienced
extreme levels of deadly violence (with average homicide levels above fifty
per 100,000 inhabitants). In more than a fifth of municipalities, not a single
person was murdered in these three years (22.4 percent), and more than
one eighth (13.9 percent) still enjoyed almost European levels of homicide
(less than 5 per 100,000).53
The objective proximity to criminal violence may have complex and
contradictory effects on public perceptions of violence. Yet overall, I expect
the same lineal relationship to hold as for victimization: objective proximity to violence is likely to generate subjective proximity to the victims of
violence. It should be harder to be indifferent to the fate of victims if they
get killed and kidnapped on your doorsteps.
To measure respondentsâ€™ geographic distance from the war, I constructed
an aggregate â€œindex of distance to violenceâ€ that adds three pieces of information: (a) objective municipal-level data on average annual rates of homicide
in their place of residence, (b) subjective sensations of local security (â€œHow
secure do you consider life in your locality?â€), and (c) subjective distance
from violence (â€œthings have been calm around here; violence occurs in other
regions of the countryâ€). To capture variance at the level of federal states,
I also estimate fixed state effects (whose results I do not to report but am
happy to make available upon request).
D. The Social Proximity of Victims
A substantial body of criminological literature argues that the perceived social
distance between citizens and criminals molds the â€œpunitive sentimentsâ€ the
former harbor against the latter.54 Social proximity seems to be regulating,
too, not just our antipathies towards perpetrators, but also our sympathies
with victims. A more disperse body of literature in history, psychology, and
politics suggests that citizens are able to watch the suffering of others with
perfect indifference, or even approval, if they are able to classify them as
53. Author calculations based on homicide data by the National Health Information System
(sinais) (www.sinais.salud.gob.mx) and population data from the 2010 national census by
the National Institute for Statistics and Geography (INEGI) (www.inegi.org.mx). Note that
the World Health Organization considers violence to be â€œepidemicâ€ once it surpasses
ten annual homicides per 100,000 inhabitants.
54. See, e.g., Mark D. Ramirez, Punitive Sentiment, 51 Criminology 329 (2013).
2016 The Criminal Community of Victims and Perpetrators 1061
distant others. â€œ[H]uman sympathy can be turned on or off depending on
how another person is categorized.â€ 55
The two most evident candidates for defining the social status of Mexicoâ€™s
victims of war are poverty and skin color. In the ethnically stratified societies
of Latin America, crime is often suspected to be ethnically stratified, too.
In the region, â€œthe most common image of criminals is of poor, nonwhite
menâ€ 56 and the same applies to the victims of violent crime. According to
one recurrent diagnosis, in Latin Americaâ€™s â€œviolent democracies,â€ 57 the
homicides tend to be â€œimpoverished, poorly educated, nonwhite adolescents
and young men,â€58 and they tend to recruit their victims from the same social stratum. Trigger-happy killers on the public payroll tend to share their
criteria of victim selection.59
Some critical observers have described the Mexican drug war in similar
terms, as â€œa war of the poor against the poor,â€ 60 â€œa new form of class war.â€ 61
If criminal violence indeed is and is perceived to be a domain of the poor,
with poor men abducting, torturing, and killing other poor men, we should
expect public opinion to reflect its social stratification. I take the reported
number of light bulbs in respondentsâ€™ dwellings as indicator of economic
status and classifications of facial skin color by interviewers (on the Latin
American Public Opinion Project [LAPOP] eleven-point color palette that
goes from pink to dark brown) as measure of phenotype.
Of course, objective respondent attributes need not translate into subjective attitudes. High social status does not necessarily produce negative
prejudice against subordinate classes and light skin color does not necessarily
produce racism. However, to the extent that (a) these objective attributes do
correlate with social and ethnic prejudice and (b) respondents conceive the
victims of organized violence as dark-skinned members from lower classes,
55. Steven Pinker, The Better Angels of Our Nature: Why Violence Has Declined 324 (2011).
Among many others, see also Hannah Arendt, The Origins of Totalitarianism (1951), Butler,
supra note 1; Grossman, supra note 13.
56. Enrique Desmond Arias, Drugs and Democracy in Rio de Janeiro: Trafficking, Social Networks,
and Public Security 12 (2006).
57. Violent Democracies in Latin America (Enrique Desmond Arias & Daniel M. Goldstein eds.,
58 Arias, supra note 56, at 1.
59. See, e.g., Daniel M. Brinks, The Judicial Response to Police Killings in Latin America: Inequality
and the Rule of Law (2008); Robert Gay, Toward Uncivil Society: Causes and Consequences
of Violence in Rio de Janeiro, in Violent Democracies in Latin America, supra note 57, at
201; Ruth Stanley, â€œLiving in a Jungleâ€: State Violence and Perceptions of Democracy
in Buenos Aires, supra note 57, at 133.
60. Daniela Rea, La justicia de Todos, in Entre las Cenizas: Historias de vida en Tiempos de Muerte
217, 230 (Marcela Turati & Daniela Rea eds.,2012).
61. JosÃ© RamÃ³n CossÃo DÃaz, Combate a la Delincuencia Como Lucha de Clases, El PaÃs,
14 July 2015, available at http://internacional.elpais.com/internacional/2015/07/14/
1062 HUMAN RIGHTS QUARTERLY Vol. 38
their social status and phenotype should be predictive of their sympathies
towards victims. The richer should be less and the darker more sympathetic.
E. Control Variables
Political Sophistication: Three standard variables in political survey researchâ€”
formal education, interest in politics, and media consumptionâ€”are likely to
mold citizen views on the war. They are all indicative of political sophistication. The more educated citizens are, the more interested in political affairs,
the more they keep themselves informed by watching, reading, and listening
to the news, and the more knowledgeable they should be about victims. If
information inhibits indifference, they should be more sympathetic as well.
To measure political interest, I use the respective standard item in ENVO:
â€œGenerally speaking, how much are you interested in politics?â€ The survey
also contains a battery of questions on news consumption: â€œHow frequently
do you follow the newsâ€ in different media (almost never, a couple of times
a month, a couple of times per week, almost daily)? I average the values for
television, radio, and newspapers.
Religiosity: Religion can justify anything. The big managers and killers
of the drug war are said to be deeply religious. Still, given the emphasis
the contemporary Catholic Church places on peace and solidarity (with
Catholicism still being the dominant religion in the country), religious belief
should lead Mexican citizens to sympathize with the plight of victims. As
measure of religiosity, I take respondentsâ€™ indications about the importance
of religion in their lives.
Sex and Age: Generally speaking, killing and being killed is a menâ€™s
business. At the global level, 95 percent of homicides as well as almost
eighty percent of their victims of homicide.62 We have little systematic
knowledge on perpetrators and victims in Mexicoâ€™s criminal war. Yet, the
familiar pattern of men killing men seems to hold. According to the â€œMemoriaâ€ dataset on organized violence in Mexico, assembled by the Justice
in Mexico project of the University of California, San Diego, for the years
2006 through 2013, â€œthe vast majority of victims were men. . . with just
9% of the victims identified as female.â€63
Mexico is no country for young men. Between 1998 and 2012, about
two thirds of victims of homicide with firearms have been younger than 40
years. The highest number of victims comes from the age group between
62. UNODC, supra note 31, at 13. On the average participation of men in the use of lethal
force in Mexico and Latin America, see, e.g., Bravo, Grau & Maldonado, supra note
22, at 79; OAS, supra note 22, Table 1.3; JosÃ© Ignacio Torreblanca, El VarÃ³n, Arma de
DestrucciÃ³n Masiva, El PaÃs, 26 Jan. 2014, at 29.
63. Heinle, supra note 39, at 33.
2016 The Criminal Community of Victims and Perpetrators 1063
20 and 29 years.64 If the so-called drug war carries a sex and age bias,
what follows for public attitudes towards victims? Psychological studies
find women to be more empathetic than men. Yet, if simple mechanisms of
social distancing work here, we should expect both women and people of
advanced age to be less concerned about victims.
F. Regression Results
I test for the causal effects of the frame of bounded violence by regressing it on dichotomous versions of all four measures of sympathy towards
victims. Table 1 shows the results of binary logistic regressions that include
my main explanatory variable, the two bundles of alternative hypotheses,
and all controls.65 The results first of all, confirm the relevance of cognitive
frames. Perceptions of bounded criminal violence carry both statistically and
substantively significant effects on each of our four indicators of sympathy
with victims. Even when controlling for everything else, each upward step on
the four-point scale of criminal selectivity depresses the odds of perspective
taking by 10 percent (eb
= .89), the odds of emphatic recall by 19 percent
= .81), and the odds of identification with victimsâ€™ movements by 15
= .85). It increases the odds of calls for silence by 26 percent (eb
= 1.26). The more firmly respondents believe in the self-contained nature
of criminal violence, the more likely they are to see themselves immune
to organized violence, to show themselves unmoved by the fate of victims
and alien to their organizing efforts, and to keep debates about the war off
the political agenda.
The competing hypothesis of proximity to violence turns out to be
complementary. As expected, personal experiences of victimization are a
strong source of sympathy towards victims. Inversely, personal distance to
organized violence works as a strong inhibitor of such sympathies (except
for assessments of victimsâ€™ movements, which are unrelated). Victims feel
closer to victims, and people closer to violence feel closer to victims, too.
All other predictors show less consistent patterns of covariation. Social
class only affects citizensâ€™ empathy and their willingness to talk about the
criminal war. Yet it does so against our theoretical expectations: poorer citi64. Bravo, Grau & Maldonado, supra note 22, at 80. See also JosÃ© Merino, Jessica Zarkin &
Eduardo Fierro, Marcado Para Morir, 427 Nexos 28 (2013), OEA supra note 22, at 21;
UNODC, supra note 33, at 14.
65. Rather than logistic regression coefficients, which are hard to interpret, the table reports
odds ratios (eb
), which indicate the extent to which the odds of occurrence of the outcome variable change if the independent variable undergoes a one-unit change. Close
to one they tell us that nothing changes, below one they indicate decreases, above one
increases in the dependent variable. The baseline odds for perspective taking are 0.37
(27/73), for empathy 0.2 (17/83), for silence 1.8 (64/36), and for movement support 0.49
1064 HUMAN RIGHTS QUARTERLY Vol. 38 TABLE 1. Determinants of sympathy with victims Perspective taking Empathy with victims Preference Movement for silence support Itâ€™s probable they I recall a We should I identify with get me killed moving case stop talking victim movements p eb p eb p eb p eb Bounded criminal violence .047 .896 .002 .814 .000 1.269 .005 .859 Victimization (0â€“5) .000 1.317 .000 1.764 .004 .837 .014 1.169 Distance to violence (0â€“11) .000 .875 .003 .890 .000 1.225 .930 1.003 Class (light bulbs) (1â€“4) .368 1.056 .037 1.169 .044 .893 .171 1.083 Phenotype (skin color) (1â€“10) .371 1.037 .026 1.116 .110 .942 .112 1.064 Education (0-8) .265 .967 .533 1.023 .724 .990 .102 1.047 Political interest (0â€“3) .182 1.085 .179 1.106 .818 1.013 .000 1.263 Media consumption (0â€“3) .009 1.237 .049 1.215 .408 1.064 .186 1.109 Religiosity (0â€“3) .357 1.060 .371 1.072 .077 1.108 .064 .895 Sex (0 = male, 1 = female) .125 1.187 .005 1.470 .900 .987 .440 1.085 Age .758 1.001 .523 .997 .813 1.001 .163 .995 Constant .003 .246 .000 .034 .205 .581 .000 .204 pac 74.3% 84.8% 68.6% 71.1% Snell & Cox R2 .123 .170 .128 .116 R2 Nagelkerke .179 .284 .176 .162 N 1983 2044 2044 1981 Note: Binary logistic regression. eb = odds ratio. pac = Percentage of accurate classification. Shaded cells indicated statistically significant coefficients (p â‰¤ .05). All models include state fixed effects (not reported). For descriptions of variables and descriptive statistics, see Tables A and B.
2016 The Criminal Community of Victims and Perpetrators 1065
zens are not more sympathetic, but less so. Skin color is unrelated to all but
one of our dependent variables: empathy. Each step along our eleven-point
scale of melanin (from pink to brown skin) augments the odds of empathic
memories by 76 percent (eb
= 1.76)! Three of our control variables prove
irrelevant: education, religiosity, and age. Political interest increases: identification with organized victims, the consumption of mass media news, the
subjective likelihood of getting killed, as well as the ability to recall moving cases. Men and women do not differ systematically in their attitudes
towards victims, with one exception: women, as gender prejudice suggests
and empirical evidence confirms, are more empathic than men (eb
Citizens who wish to deny solidarity to victims of injustice may activate two
age-old mechanisms: they may deny victims recognition as members of the
political community by conceiving them as aliens or enemies, and they may
deny them recognition as victims by blaming them for their own misfortunes.
In this paper, I analyzed a strategy of denial that fuses both mechanisms:
the equation of victims with perpetrators, their common identification as
members of the criminal community.
Whenever criminal violence seems to cluster within a certain category
of actors, be it a social class, an ethnic group, or an illicit organization,
citizens may detach themselves from it by conceiving it as an internal affair
among distant criminal actors. Situations of endemic societal violence are
usually opaque and messy. Framing them as situations of bounded violence
in which criminals kill criminals provides comfort. It imposes symbolic order
on a disordered reality. It creates a world with clear identities and a clear
separation of roles. As it blurs the line between victims and perpetrators, it
reinforces the boundary between victims and citizens.
The survey evidence reported in this paper confirms the causal relevance
of such cognitive frames in the context of Mexicoâ€™s so-called drug war. The
notion of self-contained criminal violence tends to erode sympathy towards
the victims of injustice and thus the attitudinal foundations of political
solidarity. It debilitates citizensâ€™ willingness to adopt the perspectives of
victims, to empathize with their fate, to keep them on the public agenda,
and to support the civic associations they have formed. Cognitive frames
are not carved in stone though. They are the contingent product of private
and public debate and reflection. As the events in Iguala indicate, they are
vulnerable to shocks. It remains to be seen however, whether the shock
waves of this and other such instances of joint public and private repression
produces a lasting impact on Mexican public opinion. Citizens may well
end up defining them as no more than deplorable exceptions from the rule
of safely self-contained violence within the criminal community.
1066 HUMAN RIGHTS QUARTERLY Vol. 38 TABLE A. Description of variables Dimension Survey questions Range / Categories Variable name Sympathy towards victims Perspective taking: â€œHow probable do you think it is that organized Not at all or a little (0), P6D_d crime orders to kill you or someone from your familyâ€ somewhat or very probable (1) Empathy with victims (in the context of questions on murder and No (0), yes (1) P31 disappearance): â€œOutside the circles of people you know personally, do you remember some person whose case has moved you in particular?â€ Silence: â€œThere are so many good things in Mexico. We should Disagree strongly or P10A_d stop talking so much about violence.â€ somewhat (0), agree somewhat or strongly (1) Support for victimsâ€™ movements: â€œGenerally speaking, which is your Not at all or a little, somewhat P59_d impression of these movements? How much do you identify with the (0.66) very much (1) victims who organize themselves?â€ Boundaries of violence â€œTalking about murders attributed to organized crime, how much do Disagree very much (0), P24B you agree with the following statement: As long as you do not get disagree somewhat (1), involved with them, nothing happens to you.â€ agree somewhat (2) agree very much (3). Victimization Additive index of victimization by organized violence within and 0â€“5 Index_vco outside family. Sum of five variables: Extortion: â€œOver the past years, has it happened to you or someone in No (0), yes (1) P26C your family that you were asked extortion money (â€˜derecho de pisoâ€™) to conduct your business or other activities?â€ Assassination or disappearance within family: â€œOver the past years, No (0), yes (1) P26DE some member of your family has been murdered or disappeared by organized crime?â€ Assassination or disappearance outside family: â€œSomeone among your No (0), yes (1) P30 friends or acquaintances has been murdered or disappeared by o organized crime?â€ Orphanage: â€œDo you know a child or teenager who was orphaned No (0), yes (1) P33 because criminal groups murdered their mother or father?â€
Emigration: â€œDo you know someone who migrated to the United States No (0), yes (1) P34
or some other country because of the violence?â€
2016 The Criminal Community of Victims and Perpetrators 1067 Distance to violence Additive index of subjective and objective distance to violence. Sum of 0â€“10 Index_d_viol three variables: Local security: â€œHow secure do you consider living in your locality?â€ Not at all (0), a little (1), P5 somewhat (2), very much (3) Subjective distance from violence: â€œAs a matter of fact, things have been Disagree very much (0), P10C calm around here; the violence occurs in other regions of the country.â€ disagree somewhat (1), agree somewhat (2) agree very much (3) Objective distance from violence: Five strata of civility (0â€“4) = inversion (0) = very high homicide Estrato_inv of survey sample strata of municipal violence, by municipal homicide rate (> 30), (1) = high (15â€“30), rates (annual number of homicides per 100,000 inhabitants, average (2) = medium (10â€“15), 2009â€“2011). (3) = low (6â€“10), (4) = very low homicide rate (< 6) Class Proxy for household wealth: number of light bulbs in place of residence. (1) = 1â€“3 bulbs, pk_ag (2) = 4â€“6 bulbs, (3) = 7â€“9 bulbs, (4) = 10 or more bulbs. Education Level of formal education of survey respondent 0â€“8 edu Phenotype Facial skin color of survey respondent, as assessed by interviewer at the 1 (pink) â€“10 (dark brown) enc1 end of the interview according to color palette developed by the Latin American Public Opinion Project (lapop). Political interest â€œGenerally speaking, how much are you interested in politics?â€ Not at all (0), a little (1), P1 somewhat (2), very much (3) Mass media news Frequency of news consumption in mass media: â€œHow frequently do Almost never (0), a couple of P2_prom_abc you follow the news on tv / on the radio / in the newspaper?â€ (average times a month (1), a couple of all three information sources) of times per week (2), almost daily (3) Religiousness Importance of religion in private life: â€œPlease, could you tell me, how Not at all (0), a little (1), pl important is religion in your life?â€ somewhat (2), very important (3). Dimension Survey questions Range / Categories Variable name
1068 HUMAN RIGHTS QUARTERLY Vol. 38 Sex Sex of survey respondent (binary) Male (0), female (1) Sexo Age Age of survey respondent (years) â‰¥ 18 Edad Source (in all Tables and Figures, unless otherwise indicated): Mexican National Survey of Organized Violence (envo). Dimension Survey questions Range / Categories Variable name Table A., cont.
2016 The Criminal Community of Victims and Perpetrators 1069
N Minimum Maximum Mean Standard
Perspective taking: They may 2264 0 1 .27 .446
get me killed
Empathy: I recall a moving case 2363 0 1 .17 .377
Silence: We should stop talking 2360 0 1 .64 .481
Support: I identify with victimsâ€™ 2273 0 1 .33 .471
Bounded criminal violence 2335 0 3 1.82 1.009
Index of victimization 2305 0 4 .53 .857
Index of distance to violence 2332 0 10 4.87 2.197
Class (light bulbs) 2361 1 4 2.49 .979
Phenotype (skin color) 2370 1 10 4.63 1.406
Education 2390 0 8 4.01 2.254
Political interest 2384 0 3 1.14 .958
Frequency of mass media news 2361 0 3 1.52 .740
Religiousness 2366 0 3 2.27 .895
Sex 2400 0 1 .51 .500
Age 2399 18 85 41.10 15.506
For descriptions of variables, see Table A.