H-Diplo Roundtable XXI-33
Amelia Hoover Green. The Commander’s Dilemma: Violence and Restraint in Wartime. Ithaca: Cornell University Press, 2018. ISBN: 9781501726477 (hardcover, $49.95).
23 March 2020 | https://hdiplo.org/to/RT21-33
Roundtable Editors: Thomas Maddux and Diane Labrosse | Production Editor: George Fujii
Contents
- Introduction by Elisabeth Jean Wood, Yale University.. 2
- Review by Michael Boyle, La Salle University.. 6
- Review by Jessica Trisko Darden, School of International Service, American University.. 8
- Review by Stathis N. Kalyvas, University of Oxford.. 12
- Review by Jonah Schulhofer-Wohl, Leiden University.. 15
- Response by Amelia Hoover Green, Drexel University.. 18
Introduction by Elisabeth Jean Wood, Yale University
Given the many opportunities and incentives to engage in violence against civilians during war, why do some armed organizations—states and non-state actors—engage in narrow repertoires, eschewing many forms of violence, while others engage in nearly all forms of violence? In particular, why do some engage in only lethal violence, while others deploy a much wider repertoire?
In her recent book, Amelia Hoover Green analyzes what she terms “the commander’s dilemma,” namely, how to create a force that is willing to engage in violence but remains under control of the group’s leadership. Building on literature in social psychology, sociology, and history, Hoover Green suggests that whatever their initial training, the wartime processes of wielding and witnessing of violence create a willingness among many combatants to engage in a wide repertoire of violence against civilians.
Thus the central puzzle, she suggests, is that of restraint: how do some armed organizations maintain a narrow repertoire of violence even after years of combat? The answer, she argues, lies not in patterns of recruitment and discipline but in the socialization of combatants.
More specifically, Hoover Green focuses on institutions for political education, by which she means those institutions that purposefully indoctrinate combatants with a particular identity and purpose, aligning individual preferences with those of commanders. Armed groups that continuously inculcate a strong sense of purpose, teaching and enforcing a code of conduct that emphasizes that some forms of violence undermine that purpose, have narrower repertoires than those that do not.
To assess her theory, Hoover Green focuses on the civil war in El Salvador, a conflict in which repertoires varied sharply across organizations and for which she could gather, with significant effort to recreate lost archives, relatively high-quality violence data. She draws on multiple-systems estimation (also known as capture-recapture) to establish the patterns of violence by state forces and the Frente Farabundo Martí para la Liberación Nacional (FMLN). She also draws on dozens of interviews with former combatants (both soldiers and insurgents), an original survey of Salvadoran military veterans, and documents from the archives of the FMLN, major human rights groups, and the National Security Archive to analyze their institutions and practices.
Hoover Green shows that the FMLN’s repertoire was consistently narrower than that of state forces, and that her theory accounts for the difference: the FMLN engaged in ongoing indoctrination of its combatants to its socialist ideology and overarching strategy of limited violence, while the state made little effort to indoctrinate soldiers. Her analysis of changes in their repertoires over time also confirms her theory, as do the contrasting repertoires of different units within the insurgent group. Hoover Green also shows that her theory accounts for variation in repertoire across several rebel organizations in the civil wars in Sierra Leone and Liberia, a difficult test, she explains, because those organizations had access to diamonds and so were less dependent on civilians and held territory much less continuously.
Among the book’s central contributions are that it brings into scholarly focus variation in repertoires of violence, particularly the challenges of restraint; draws across disciplines for appropriate theoretical insights; develops an original theory of combatant socialization for restraint; and offers a mixed method research design that leverages cutting edge statistical methods with a wide range of qualitative data.
The roundtable’s four reviewers concur that the book makes original and important contributions to our understanding of wartime violence. Stathis Kalyvas describes it as a “rich, nuanced and important book worth reading closely” with empirical material “as rich as it is compelling.” Both Jonah Schulhofer-Wolf and Jessica Trisko Darden note implications for policy that follow from the argument.[1]
Both Kalyvas and Trisko Darden emphasize Hoover Green’s forthright discussion of her argument’s limitations. And Kalyvas notes that the discussion of ethics in the appendix is a “must read” for those researching violence. All the reviewers, particularly Jonah Schulhofer-Wohl, emphasize the rich research agenda that follows from the central argument, including whether political education varies with the type of organization, the role of ideology in commander choice of institutions and the content of combatant socialization, and the degree to which ties between insurgent organizations influence institutions and the content of socialization.[2]
The reviewers also question some aspects of the argument and evidence. Two reviewers note that the content of political education need not be toward restraint. Michael Boyle asks: What if political education emphasizes gratuitous violence? Kalyvas similarly asks: What if it emphasizes restraint only towards some civilians? Kalyvas also suggests that there are other pathways to restraint other than the one Hoover Green emphasizes.
Both Kalyvas and Trisko Darden find Hoover Green’s definition of violence as physical violence against persons or immediate necessities such as homes or food supplies too narrow, as it excludes other attacks that may also cause extensive suffering. Trisko Darden raises a concern that Hoover Green’s empirical approach underestimates civilian casualties in some categories, and suggests that the United States may have had some role in the narrowing of the repertoire of violence on the part of state forces in the later years of the war.
Finally, Kalyvas finds a recurrent and, he suggests, perhaps inevitable tension between repertoire breadth and the level of violence, which he reads as arising from a presumption that narrow repertoires imply low levels of violence—and thus restraint.[3]
In her response, Hoover Green notes that inclusion of violence that she did not measure would be very unlikely to change her principal empirical finding of the sharp contrast in violence repertoires between the Salvadoran military and the insurgents. She notes while repertoires can indeed be narrow while levels of violence are high, as reviewers suggest, that pattern is consistent with her theory: narrow repertoires reflect strong internalization of acceptable forms of violence whatever the level of violence. In addressing the extent to which commanders can in fact choose institutions and norms, Hoover Green pushes back against the suggestion that such choices are structurally determined, but notes the need for data on institutions for a broader set of organizations to study the origins of commander choice. Similarly, Hoover Green agrees that the evidence available in the case of El Salvador does not allow her to test all the implications of her theory, but robustly defends multiple systems estimation against the claim that the method lacks rigor or depth, noting its use in quantitative disciplines such as demography and public health, citing a recent coauthored article.[4]
The roundtable demonstrates the importance of the book’s theory and findings, as well as its power to set the agenda for researchers working on patterns of violence against civilians in war.
Participants:
Amelia Hoover Green (Ph.D. Political Science, Yale University, 2011) is Associate Professor in the Department of Politics at Drexel University, Philadelphia, PA, and a Consultant to the Human Rights Data Analysis Group. Her recent publications include “Civilian killings and disappearances during civil war in El Salvador” (Demographic Research 2019, with co-author Patrick Ball), and the chapter “Successful Fieldwork for the Fieldwork-Hater,” in Krause and Szekely, eds., Stories from the Field (Columbia University Press, 2020). Her current projects include a simulation-based approach to assessing the impact of biased and incomplete violence data, and a cross-national data collection on armed groups’ internal institutions
Elisabeth Jean Wood is the Crosby Professor of the Human Environment and Professor of Political Science, International and Area Studies at Yale University. She is the author of Forging Democracy from Below: Insurgent Transitions in South Africa and El Salvador (Cambridge University Press, 2000) and Insurgent Collective Action and Civil War in El Salvador (Cambridge University Press, 2003), and co-editor with Morten Bergsmo and Alf B. Skre of Understanding and Proving International Sex Crimes (Oslo: Torkel Opsahl Academic Epublisher, 2012). She is currently writing a book on wartime sexual violence. A fellow of the American Academy of Arts and Sciences, she teaches courses on comparative politics, political violence, collective action, agrarian studies, and qualitative research methods.
Michael J. Boyle is an Associate Professor of Political Science at La Salle University and is the author of Violence After War: Explaining Instability in Post-Conflict States (Baltimore: Johns Hopkins University Press, 2014).
Jessica Trisko Darden is Assistant Professor at American University’s School of International Service and Jeane Kirkpatrick Fellow at the American Enterprise Institute. She is the co-author (with Alexis Henshaw and Ora Szekely) of Insurgent Women: Female Combatants in Civil Wars (Georgetown University Press, 2019) and author of Aiding and Abetting: U.S. Foreign Assistance and State Violence (Stanford University Press, 2020). She is also the author of Tackling Terrorists’ Exploitation of Youth (American Enterprise Institute, 2019). Her research examines the political economy of conflict with an emphasis on U.S. foreign assistance.
Stathis N. Kalyvas is Gladstone Professor of Government and fellow of All Souls College at Oxford. Until 2018 he was Arnold Wolfers Professor of Political Science at Yale University, where he also directed the Program on Order, Conflict, and Violence and codirected the Hellenic Studies Program. Kalyvas obtained his BA from the University of Athens (1986) and his M.A. and Ph.D. from the University of Chicago (1993), all in political science, and the European University Institute. He is the author of The Rise of Christian Democracy in Europe (Cornell University Press, 1996), The Logic of Violence in Civil War (Cambridge University Press, 2006), Modern Greece: What Everyone Needs to Know (Oxford University Press, 2015), the co-editor of Order, Conflict, and Violence (Cambridge University Press, 2008) and the Oxford Handbook on Terrorism (Oxford University Press, 2019), and the author of over fifty scholarly articles in five languages, as well as several books and edited volumes in Greek. His current research focuses on global trends in civil conflict and political violence with an additional interest in the history and politics of Greece.
Jonah Schulhofer-Wohl is Assistant Professor of Political Science at Leiden University. He specializes in the conduct of civil wars, with an empirical focus on the Middle East. Current projects study the interaction between civil war belligerents at three levels of analysis: quamire as the macro-level result of the interaction between the warring parties; warfighting choices, focusing on alliance behavior and the operational goals of fighting, both meso-level behaviors; and, at the micro-level, the behavioral determinants of individual actions in situations of group conflict. His work has been published in International Security, the Journal of Conflict Resolution, and Science. His book, Quagmire in Civil War, is forthcoming from Cambridge University Press.
Review by Michael Boyle, La Salle University
One of the central puzzles of recent research on civil wars concerns the variations in violence within it. Scholars have highlighted a number of explanations for why violence might vary within a civil war, including the control of territory, the degree to which combatants have information about the loyalties of the civilian population, and the degree to which combatants are incentivized to use indiscriminate violence by loot and other material rewards.[5] Alongside this work has been a new wave of scholarship on how specific types of violence such as rape have also varied within wars.[6] In this valuable new volume, Amelia Hoover Green tackles both questions and offers a theoretical framework to explain variations in both the intensity and types of violence present within civil war.
Her explanation for this variation begins with an important, but often overlooked, observation: that wartime commanders have an incentive to get their fighters to act violently, but not so violently that their actions produce a backlash and jeopardize their larger goals within the war itself. In other words, they face a Goldilocks-style problem: they need violence, but just enough. Avoiding both the oversupply and undersupply of violence is what Hoover Green describes as the “Commander’s Dilemma.” She acknowledges that punishments and rewards are one way to modulate the violence, but makes a compelling case that fostering norms of restraint among the soldiers themselves is equally if not more important. While this process of internalizing norms happens through a number of mechanisms, including recruitment and training, Hoover Green highlights the role played by political education in encouraging soldiers to limit their use of violence, particularly against the civilian population. She shows in detail how different groups (both government and rebel groups) deployed political education in El Salvador’s civil war and how this affected the levels and types of violence that they use.
There is much to like and praise about this book. Perhaps one of the more attractive elements is its deployment of the concept of repertoires of violence, defined here as the variety of forms of violence (for example, assault, kidnapping, rape) that are present in a particular conflict. Hoover Green conceptualizes this as either narrow or broad depending on the number of forms of violence present in a conflict. This is a helpful way to think about it, though it is notably a different conceptualization of repertoires than that offered by Charles Tilly and his-co-authors about a decade ago. They focused more on exploring repertoires of contention (both non-violent and violent) and seeing how they emerged from one another. While Hoover Green makes some references to Tilly’s work, she leaves aside his Politics of Collective Violence (2003) and Dynamics of Contention (published in 2001 with Sidney Tarrow and Doug McAdams), both of which make a different case for considering a wider account of repertoires.[7] On some level, this is entirely defensible: her scope and aim is more limited (i.e., only violence in civil wars) than theirs. But it would have been fascinating to pair some of their insights – particularly that non-violent contentious interactions can produce violent ones and that they may share causal mechanisms—with the rich narrative evidence of violence in El Salvador that Hoover Green has assembled here. This might be a possible extension for Hoover Green to explore in future work, though some conceptual refinement would need to be done on converting Tilly’s approach into testable hypotheses.
The emphasis on the role of political education that Hoover Green identifies is an important one and contrasts with other explanations for the variation in violence, such as ones based on institutions and political wings of rebel groups.[8] With detailed evidence from interviews with a variety of combatants in El Salvador, and briefer comparative evidence from Liberia and Sierra Leone, Hoover Green makes a good case that the political education about the purpose of the fighting and the need to respect civilians inculcated norms of restraint among combatants, which in turn produced narrower repertoires of violence. This argument is plausible and is supported well by the rich detail provided here, but as she notes it is hard to improve the causal pathway beyond a reasonable doubt. Moreover, it is nevertheless important to be careful about the applicability of a general argument about political education. As Hoover Green acknowledges and addresses in depth, the evidence is truncated because obviously not every act of wartime violence was recorded and memories are sometimes faulty. More generally, it is also important to see how political education varies across a variety of cases of different combatants and conditions within the civil wars. There may be something more persuasive about the success of political education provided by rebel groups that are looking more likely to win their conflict than that provided by groups whose futures are uncertain. Also, ideology might matter here. Marxist and/or Maoist political education that encourages restraint (as Mao Zedong indeed did with his writings on guerilla warfare) might have a different effect than other types of political education. The comparative evidence that Hoover Green presents from Liberia and Sierra Leone deals with this objection to an extent, but the argument would be stronger if were based upon the investigation of some possibly negative cases of similar Marxist-Leninist groups (for example, Sendero Luminoso in Peru) that were deeply invested in political education but nevertheless were gratuitously violent.
Another way in which the argument could be advanced would be to examine cases where there is a divergence of preferences and/or behavior between the commanders and the soldiers. In other words, are there cases where political education can be shown to rein in soldiers from expanding their repertoire of violence even when they have a commander egging them on to be particularly violent against civilians? Such a case, admittedly perhaps rare and hard to find, would offer a stronger test of the theory. Hoover Green leaves commanders aside (207) from much of the theory, in part for evidentiary reasons, and in part because she finds standard principal-agent theories lacking in a number of other respects (38). But given the evidence suggesting that commanders can produce an enabling environment for atrocities, using event-level data and accounts of specific atrocities within the El Salvador conflict to show how political education produced a different outcome than might be expected given the commander’s preferences would strengthen her case.
In the end, the overall argument and its applicability to El Salvador and her other cases is persuasive and well-constructed. Hoover Green has provided a compelling alternative explanation for the variation in violence in civil wars and made a strong case that the field could advance by considering repertoires of violence rather than just body counts. She has also produced a skillful blend of field work, interviews, and quantitative data which shows both rigor in its attempt to deal with messy, imperfect data and a real sensitivity to the lived experiences of those on the ground. This book will thus be of great interest to scholars and graduate students who want to use mixed methods and produce a theoretically sophisticated book without sacrificing the nuances of the civil war itself.
Review by Jessica Trisko Darden, School of International Service, American University
The leadership of both state and nonstate armed groups face a challenge: after successfully socializing individuals to engage in violence, how do they establish control over the types of violence that their soldiers commit? One aspect of this challenge—which Amelia Hoover Green terms the Commander’s Dilemma—is easily understood. In a civil war, if groups engage in too much violence, they risk alienating the local populations on which they rely for support. If groups employ too little violence, they risk being swiftly defeated. But Hoover Green focuses on something different: what forms of violence are adopted by an armed group as its repertoire.
I focus on the contributions that The Commander’s Dilemma: Violence and Restraint in Wartime makes to three main areas, though readers will likely identify many more. First, this book is fundamentally about the use of violence against civilians in war. Hoover Green offers a compelling argument that increased complexity and nuance in the ways that we study the use of violence, and in particular the repertoires of violence that armed groups adopt, change our understanding of conflict dynamics. She does so by disaggregating violence into lethal and nonlethal forms and then further disaggregating the latter category. Second, and integral to the use of violence, the book explores political education and how it is manipulated—or not—by armed groups to establish discipline and control within their ranks. Third, Hoover Green is keenly aware that bringing a gender analysis to these first two issues yields new insights and questions that ultimately lead back to why we see variation in the use of violence in war.
The Uses of Violence in Civil War
The Commander’s Dilemma is one among several recent studies of the dynamics of El Salvador’s civil war (1980-1992).[9] Many scholars examine the main nonstate actor in the conflict, the Frente Farabundo Martí para la Liberación Nacional (FLMN), as a unitary actor even though it was initially formed as an umbrella organization for five revolutionary groups. As a fighting force, the FMLN’s guerrillas were once the largest nonstate armed actor in Latin America, numbering between 10,000 and 12,000 in late 1983.[10] A prominent minority of FLMN members were women, with estimates ranging from 15 to 30 percent of FMLN cadres.
Hoover Green expands our understanding of the FMLN by comparing its two major factions: the Revolutionary Army of the People (Ejército Revolucionario del Pueblo, ERP) and the Popular Forces of Liberation (Fuerzas Populares de Liberación Farabundo Martí, FPL). Doing so allows her to leverage more variation across different parties to the conflict, in addition to variation over time and across space.
Hoover Green’s discussion of multiple systems estimation and variations in the available data on El Salvador’s civil war violence makes a significant contribution to our understanding of the data’s insights and limitations. In analyzing this data, she convincingly argues that “examining only one type of violence can dramatically misrepresent the dynamics of violence, especially as repertoires evolve over time” (204). She finds that FMLN forces committed vastly lower levels of violence against civilians than state forces, which is in line with most existing research. But in the process, she neglects to discuss some prominent violent FMLN actions against civilians, such as a multiyear campaign of targeted assassinations of mayors.[11]
By intentionally including nonlethal violence, Hoover Green seeks to expand the types of data researchers collect as well as our interpretation of what available data currently exists. Yet, Hoover Green defines nonlethal violence in a somewhat circumscribed way as physical acts directed at civilians. Examples of nonlethal violence include torture, kidnappings, assault, and sexual violence (147). This definition is similar to widely used definitions of physical (or personal) integrity rights violations, with the exception that Hoover Green’s approach is ambiguous with regard to whether the actions should still be coded as nonlethal if they led to an individual’s eventual death.[12] More importantly, this approach neglects a significant portion of the violence that the FMLN did commit—attacks on physical infrastructure, most commonly public infrastructure such as bridges and electrical grids but also farms, factories, and other private property.[13] The implicit assumption is that the FMLN’s economic violence was directed at the state and not at civilians and is therefore outside of the study’s scope. However, economic violence clearly impacted civilians’ daily lives by creating widespread blackouts and disrupting the cotton industry. The FMLN’s campaign of economic sabotage is estimated to have cost between $1 and $1.5 billion in direct repair costs and over $2 billion in economic damages.[14]
Hoover Green’s emphasis on identifying and explaining nonlethal violence intersects with the U.S. government’s provision of nonlethal assistance to El Salvador, which began under the Jimmy Carter administration. From the American perspective, nonlethal assistance included tear gas, crowd control equipment, and riot gear. (It was unclear whether hand grenades and grenade launchers were in fact nonlethal and the United States ultimately decided not to provide them).[15] The nonlethal state violence that the United States sought to support in El Salvador diverges from that discussed in The Commander’s Dilemma, which focuses extensively on sexual violence, including rape (see 103, 169). In this area, Hoover Green adds to the work of Miranda Alison, Dara Kay Cohen, Myriam Denov, Michele Leiby, Meredith Loken, Megan MacKenzie, Laura Sjoberg, Elizabeth Wood, and others, on wartime sexual violence.[16] But other forms of nonlethal violence receive limited attention in comparison.
All armed branches of the Salvadoran state were implicated in the civil war’s violence—including the National Guard, the Treasury Police, and the National Police. Hoover Green finds that state violence grew less lethal over the course of the war, but only somewhat less common (165). However, this observable change in the state’s repertoire of violence is not directly attributed to any cause. The ‘American-imposed’ National Plan (1983-1986) and the strategic changes it brought to the Salvadoran military are mentioned (95), but U.S. military training teams were operational in El Salvador beginning in 1980.
In addition, small military-backed civilian militias conducted joint operations with and received logistical support from branches of the military.[17] Yet the role played by these state-aligned militias or ‘death squads’ is relegated to the ‘other/unknown state forces’ category in the book’s analysis. A better understanding of how death squads operate (including their command structure), how they select targets, and their relationship to the state remains an important area of research.
The United States’ influence on violence in El Salvador is not directly addressed in The Commander’s Dilemma, outside of its role in Salvadoran military training. This is relevant because the United States may have served as a restraint on some forms of violence in El Salvador (and as an enabler of others—the United States subsidized the Salvadoran air force’s jet fuel, for instance).[18] Both the Carter administration and Congress pushed for reductions in civilian killings. President Ronald Reagan sent Vice President George H.W. Bush to El Salvador in December 1983, where he condemned the violence being perpetrated by the FLMN and government-aligned militias. Hoover Green’s fine-grained data provides an opportunity to test whether these efforts actually impacted the behavior of armed groups. Research by Hal Brands suggests that death squad violence did decline considerably after Bush’s visit even as U.S. military assistance reached its zenith.[19] The lengthy discussion of U.S. violence in Vietnam in the conclusion could have offered Hoover Green an opportunity to explore these issues.
Political Education and Armed Group Behavior
For Hoover Green, the key to solving the Commander’s Dilemma is political education. Through semi-structured interviews and a survey of 360 ex-combatants, Hoover Green seeks to establish the roles that socialization, training, and indoctrination played in shaping the repertoires of violence of Salvadoran armed groups. Broadly, she finds that by emphasizing limited violence, group loyalty, and a positive combatant identity, the FMLN was more successful in limiting violence against civilians. She contends that the armed forces of El Salvador lacked comparable political education despite the fact that they received highly routinized military training. This, more than the extensive differences in the size of the combatant forces (at peak 12,000 versus 60,000) or their military capabilities, explains why so much violence against civilians was perpetrated by the state.
To be clear, Hoover Green does not argue that differences in tactics or capabilities are irrelevant. But following a brief discussion of the air war, she concludes that “in general, it appears, violence against civilians by the Salvadoran Army had much more to do with location than strategy or tactics. In interviews, a number of Salvadoran Army veterans stated that harassing civilians was, in effect, a way for soldiers to entertain themselves…Neither new weapons nor new strategic postures significantly altered the state’s repertoire of violence” (166). This claim is difficult to accept or to reject in part because the data that Hoover Green uses may underestimate the toll on civilians of the war’s aerial bombardments (137-140).
The focus on political education is novel and linked to a burgeoning literature on the institutional dimensions of nonstate armed groups. One area where the analysis of the FMLN’s political education could be expanded is its connection to the political education and agendas of other leftist organizations in Guatemala, Nicaragua, and Cuba with which the FMLN had ties. The influence of these groups on the FLMN’s political education, and vice versa, is one of many avenues for future research that The Commander’s Dilemma opens up. El Salvador also offers an additional case regarding how competition or outbidding between ideologically similar nonstate armed groups can shape their political agenda. Competition between nonstate armed groups in Colombia can help explain the expansion of groups’ gender and LGBTQ rights platforms, for instance.[20]
That gender is taken seriously as an influence on the behavior of armed groups is among the many laudable aspects of this book. Attention to how gender relates to armed group practices including socialization, discipline, and combat are found throughout The Commander’s Dilemma (for examples, 100, 102, 125, 164). The explicit focus on sexual violence, noted above, adds to the complexity that characterizes the analysis. Gender, however, is relatively underexplored in the content of the FMLN’s political education. For instance, Figure 3.1 displays two pages of FMLN political education (92, for a translation, 91). A woman is clearly depicted educating men. An armed male FMLN member is shown holding a child as a women stands next to him. In both instances, there are no markers to indicate that these women are part of the FMLN or that they bear arms. What does this mean for a group that had notable female participation, including as combatants? In what ways were women included and excluded from the group’s political education? As more research is done on women’s roles in nonstate armed groups, questions like these should be at the forefront.
Conclusion
It is no small feat that The Commander’s Dilemma opens the door to so many nuanced research questions across multiple academic disciplines. The book’s approach is complex and rich. It is also a refreshingly honest book—one that reflects on its own biases and limitations as well as how it relates to existing research. Few scholars would be willing to, or indeed could, write with the degree of self-awareness and acuity that Hoover Green shows. All of this makes The Commander’s Dilemma an important contribution to the study of wartime violence and to our understanding of violence and restraint in civil war.
Review by Stathis N. Kalyvas, University of Oxford
Violence exercised against civilian populations, in and out of the context of war, is a topic of enduring fascination—and, in the last fifteen years, a subject of rigorous scholarly investigation. In particular, understanding violence against noncombatants in civil war contexts has attracted considerable attention from multiple conceptual angles and methodological approaches.
Amelia Hoover Green proposes her own perspective, one that is both original and valuable. Her most important contribution takes the form of a conceptual move focusing on ‘repertoires of violence,’ i.e. the combination of all different types of violence exercised by an armed group, both as a collective actor and by its individual members. For example, an individual fighter might take part in killing and abusing noncombatants in a variety of ways and loot property—on top of engaging in actual fighting: this is a broad repertoire. Repertoires range from narrow (few, tightly controlled forms of violence) to broad (many forms) and may vary across individuals or subgroups, across geographical space and time.
The concept of ‘repertoire’ presents two advantages over existing approaches. First, most studies up to now have focused on lethal violence, primarily for reasons of empirical convenience; Hoover Green’s book is able to move beyond this, since the concept of repertoires allows for a fuller range of violent practices. Second, unlike studies that focus on specific forms of violence, lethal or non-lethal, such as rape or forced mass displacement, she is able to examine combinations of types of violence as opposed to single types. Both of these moves are welcome, yet they come with their own problems. I highlight three.
First, tradeoffs are inevitable. If the focus on death counts leaves much violence out, so does the concept of repertoires, which, by being focused on physical violence, leaves out all sorts of non-physical forms (6). Likewise, her critique of the exclusive focus on lethal violence (based on the claim that ignoring nonlethal violence either implies a theory of its dynamics or the assumption that it is random) applies equally to her choice of ignoring non-physical violence and highlights the reason as to why this is perhaps too strong a critique.
Second, repertoires impose extremely stringent empirical demands on researchers, since counting all forms of violence exercised by various actors is hard. Indeed, things can very quickly become extremely complicated: besides the number of forms of violence regularly used, estimating a group’s repertoire of violence might call for measuring quantities such as the proportion of episodes of violence comprising a particular form of violence, the ratio of lethal to nonlethal violence, or the extent of uniformity across individuals and sub-groups comprising particular groups. Hoover Green does a good job of trying to measure these repertoires, primarily in El Salvador, but her method, the so-called Multiple Systems Estimation, ultimately falls short, as she herself recognizes.
Third, despite repeated claims that repertoires and levels of violence are analytically separable, they seem to be extremely close, and sometimes indistinguishable. Broad repertoires of violence end up connoting unrestrained violence, whereas narrow repertoires are more or less equated to instances of restrained violence. In other words, Hoover Green’s ultimate aim is to measure and explain restraint, but the book approaches it obliquely, through the concept of repertoires. This decision is not without problems. For instance, it is counterintuitive to equate ‘narrow repertoires’ with ‘restrained violence,’ since a high number of killing can be in theory simultaneously an instance of a narrow repertoire and of unrestrained violence. Things can become even more complicated when repertoires are associated with other dimensions of violence, for example, with indiscriminate violence (5). Hoover Green defends her choice vigorously on empirical grounds, claiming that narrow repertoires tend to be associated with lower levels of killings; yet her claim both remains to be established empirically and is not supported theoretically (on Table 0.1 for example, the cell of low levels and broad repertoires is deemed to be exceptionally rare). Conversely, although narrow repertoires are compatible with wildly varying levels of violence, this distinction ends up being downplayed and is not really explored in depth. While going over the book’s empirical part, I kept wishing for a comparison of the Farabundo Martí National Liberation Front (FMLN) with the Liberation Tigers of Tamil Eelam (LTTE) in Sri Lanka or the Peruvian Shining Path, both of which appear to have had strong systems of political education in place yet implemented both broad repertoires and high levels of violence. In short, we are left with the feeling that narrow repertoires are overall preferable to broad ones because they tend to be associated with lower levels of violence. Compounding this problem, the language shifts very often to “low levels of violence” either exclusively (for example, on page 20) or in conjunction with narrow repertoires (for example, on page 45). The conclusion takes us back to the contrast between “kinds of violence” versus “how much violence” (200) with an emphasis on the former. This confusing back-and-forth between repertoires and levels is inevitable given the uneasy relationship between the two and our natural (though not necessarily healthy) fascination with levels over forms of violence, but it is not always helpful.
Another problem comes into sharp focus during the examination of the civil war in El Salvador which constitutes the empirical backbone of the book. For instance, following U.S. pressure on the government of El Salvador, the regime’s forces lowered their level of lethal violence. A naïve focus on lethal violence could lead us to conclude that the regime became nonviolent, clearly a misleading claim, as a large body of credible evidence shows that arbitrary detentions, torture, etc., continued. In contrast, a ‘repertoires of violence’ approach, helps us sidestep this pitfall, but at the same time introduces a problem of false equivalence. Because the regime’s repertoire of violence remained broad after the U.S. intervention (8), we might be tempted to (misleadingly) conclude that nothing really changed. However, the reduction in lethal violence by the regime was both real and important.[21]
In the end, and despite their central position, repertoires of violence turn out not the book’s main analytical goal, but merely a means of explaining restraint in violence. Restraint is puzzling, because of what Hoover Green aptly terms the Commander’s Dilemma. This dilemma is the expression of the contradiction exercised by two opposite incentives faced by military commanders: making their fighters more likely to unleash violence when needed, but also making them less likely to do so without explicit orders. In Hoover Green’s words, restraint “occurs only when commanders exercise consistent discipline and fighters internalize messages that valorize limited, controlled violence” (17). Note, however, that the minimum that fighters need to do is valorize their commanders’ orders, which is something altogether different. Come what may, the solution to the Commander’s Dilemma, according to the book, is the provisioning of a robust program of political education which leads to more restrained repertoires of violence—or, in a more precise formulation, “When restraint does appear … it is likely because consistent discipline is accompanied by some mechanism for internalizing norms of restraint. Usually that mechanism is political education of some sort” (79).
Unlike much recent research in political science, Hoover Green’s book devotes a considerable amount of attention to the causal pathway leading from political education to restrained behavior, emphasizing social psychological processes; and she effectively distinguishes political education from parallel processes, such as recruitment, discipline, and military training. Although this is one of the strongest parts of the book, I was left with a key question: what if fighters receive a type of political education that stresses restraint vis-à-vis some civilians but not others (i.e. landowners, bourgeois, reactionaries, non-coethnic, etc.)?[22] I also wondered if narrow repertoires were indeed an externality of the commanders’ need to solve their Catch-22 dilemma, or whether political education is implemented more directly in order to produce restraint. If it is the latter, then political education rather than be a full-fledged theory of restraint may be instead a missing causal mechanism of existing theories of restraint, such as those put forward by Hyeran Jo or Jessica Stanton.[23]
The empirical part of the book is as rich as it is compelling, consisting primarily of a thorough study of violence during the civil war in El Salvador, based on interviews, archival material, and a quantitative analysis of death counts. This is complemented by a less extensive study of the violence that took place in the context of the civil wars in Sierra Leone and Liberia (the “Mano Rivers wars”), which is based on secondary sources along with a correspondingly heroic data analysis effort. The analysis makes a plausibly strong case for a meaningful variation of violence between armed actors (primarily in El Salvador, less so in the Mano River wars) and supplies credible evidence that this variation is linked to the presence or absence of programs of political education (again, primarily in El Salvador). Things are more muddled in the Mano River wars empirically, as these conflicts lack some of the key strategic features of classic irregular wars, moving them beyond the argument’s scope conditions (198). Nevertheless, a militia in Sierra Leone, the Kamajors (or CDF—Community Defense Forces) are found to have exhibited higher levels of restraint because they socialized their members through ritualistic norms of community service derived from hunter societies and other traditional groups. At a general level, these norms share a logic that resonates with the logic of political education, but clearly, the two cannot be conceptualized as instances of political training without stretching this concept.
Obviously, the book’s core finding (assuming it holds more broadly and generalizes, which future research will clarify), begs a tantalizing question: what explains variation in the implementation of political education programs by armed groups? The author speculates about the possible role of the Cold War and the character of communist insurgencies but does not further address the issue. Whether this is driven by ideology, learning of some kind, or some other process remains an open question—one well worth exploring systematically.
In the end, I found Hoover Green’s concept of repertoires of violence to be more intriguing than fully convincing, and her theoretical argument about the role of political education to be plausible, though somewhat underspecified, in that it is hard to distinguish from norms of restraint that obtain via different channels and possibly connected to broader processes (ideology, traditions, etc.) that have yet to be researched thoroughly.
Some of these critiques are anticipated and discussed in an impressively clear-headed and frank section of the conclusion (“Missing Pieces,” 204-209) which ought to be an exemplar for writers (the appendix on “Ethical Considerations” is also a must-read for anyone planning to research these vexing questions). There is no question in my mind that Amelia Hoover Green’s book has made an important contribution to the study of violence in civil war contexts. The author writes with extreme clarity and sharpness, unusual intellectual maturity, and a very strong sense of what she can and what she cannot show: there is no over-claiming and she looks very hard at her own claims. I find her call to move beyond lethal violence and be more rigorous with how to measure violence to be absolutely on target. I also second her call for placing descriptive inference and good measurement before causal inference (37), since good causal inference based on bad data takes us back to the well-known problem of ‘garbage in, garbage out.’ I also endorse the way she combines different methods together and weaves conceptual, theoretical, and empirical analysis. This is a rich, nuanced, and important book worth reading closely.
Review by Jonah Schulhofer-Wohl, Leiden University
The act of violence is far from central, even relegated off-stage, in many social scientific studies of war. This absence, if strange, is not surprising. Scholars are steeped in the debates and controversies of a discipline, engage in ongoing—if sometimes arcane—conversations with their colleagues’ research findings, and strive to enhance the scientific quality of their work. But with these necessary underpinnings, few studies probe the visceral activities that comprise war, even within the growing micro-level research agenda on civil war, an area that is well equipped to do so. Quite fortunately, then, Amelia Hoover Green’s The Commander’s Dilemma is an unusual book.
Hoover Green studies perplexing variation in repertoires of violence, advancing a compelling explanation based on whether an armed organization undertakes the political education of its members. The book employs a sound comparative research design, advanced statistical methods, extensive field research on the ground-truth of civil war in El Salvador, and includes informative case studies of Liberia and Sierra Leone to boot. But it is also more than a strong work of social science. Hoover Green’s expert theoretical exploration of the act of violence opens powerful avenues for future research on armed organizations. The book’s findings should also stimulate an important policy debate regarding command responsibility for war crimes.
Repertoires of violence, defined by Hoover Green as the “forms of violence frequently used by an actor, and their relative proportions” (5) vary across organizations, over time, and across conflicts. It is clear that they are a key manifestation of wartime behavior. And yet as outcome, they have hardly been studied. In the book’s introduction, Hoover Green lucidly explains the profound gap that this omission creates. Theoretical and empirical analyses tend to focus on individual forms of violence. Such binaries, for example, the presence versus the absence of killing, might give scholars analytic leverage but impede the study of actors’ wartime choices. If something deters combatants from killing civilians, it may not change the lethality of their behavior if instead of killing they destroy shelter, livestock, and stores of food. And murder is not the only way that armed organizations can irreparably alter the course of civilians’ lives; sexual violence, torture, and long-term detention all take a severe toll on individuals, families, and communities. For combatants, commanders, and civilians, then, the full set of possible wartime experiences matter. Any research agenda that studies behavior in civil war would be well served by the investigation of repertoires of violence.
To understand repertoires, Hoover Green interrogates the act of violence. The questions, assumptions, and empirical focus of research on war often rest on understandings of what constitutes ‘baseline’ behavior, of individual combatants or of armed organizations. But putative baselines tend to derive from anecdotal evidence. And studies rarely explicate the logic behind them or confirm them using systematic empirics. The result is that presumed understandings can be fundamentally wrong. Based on a careful reading of civil wars in comparative and historical perspective, Hoover Green points out that while violence against civilians may be abhorrent, empirically it is far from aberrant. To understand repertoires of violence against civilians, we should ask not ‘why violence?’ Hoover Green contends, but what generates combatant restraint?
Building from a rich literature in military sociology and psychology, Hoover Green argues that combatants must be socialized in order to be able to carry out violence. To produce forces capable of fighting, then, a military organization must ensure that its troops become disinhibited with respect to violence. The “Commander’s Dilemma” comes from the tension between organizations’ two core military needs: to deploy capable (read disinhibited) fighters, and to manage these same combatants, specifically to direct and limit their use of violence, so as to successfully wage war. The combat environment adds to the problem as it also “increases combatant predispositions to all types of violence, ordered and unordered, strategic and antistrategic” (34). An important scope condition of the research is that to the extent that limiting and directing violence is not necessary to armed organizations’ success, the dilemma does not bind. The book’s argument, then, applies in the main to “irregular wars in which soldiers interact directly with civilians” (17-18).
Hoover Green argues that punishment and disciplinary systems are not sufficient to produce restraint. This is due to the changes in incentive structures produced by combatants’ place in a military organization and the nature of combat. In particular, “In wartime, calculation of rewards and punishments beyond the present moment goes out the window” (39). Restraint is possible only if combatants internalize norms that support a narrow repertoire. Chapter 1 explains the deficiencies of several possible channels for socialization into norms of restraint—recruitment, discipline, and military training. Instead, thorough, ongoing political education, defined as “formal instruction that explains specific social or political purposes of a particular conflict, and connects conflict purposes to specific behavioral norm” (41), produces combatant-internalized norms that produce restraint.
The empirical material against which The Commander’s Dilemma tests its argument is impressive. The empirical heart of the book is a multifaceted study of El Salvador, grounded in long interviews and survey research with former combatants. Chapter 2 introduces the war. Chapter 3 examines the inner workings of state and opposition armed organizations. Hoover Green then assesses the argument via an elegant investigation of the two largest organizations operating under the opposition umbrella of the Farabundo Martí National Liberation Front (FMLN) in Chapter 4, showing the evolution of repertoires of violence by each one over time and comparing and contrasting the two, and using subnational data on six forms of violence in Chapter 5 (killing, sexual violence, detention/abduction, torture, other, property crimes, other nonfatal assault) to study state and opposition repertoires (13 hypotheses based on the theory and a further 17 alternative hypotheses based on the literature are considered; see 151-152).
One among many reasons the El Salvador study stands out is Hoover Green’s careful measurement of violence. She fluently and concisely explains the problem that unreported events pose for the analysis of violence and how to use Multiple Systems Estimation (MSE) to fill in otherwise-missing data (51-53). The Commander’s Dilemma is proof of concept that MSE civil war researchers can use MSE fruitfully. High data availability requirements (multiple, overlapping sources that record violence) may have limited MSE’s use in the past. But the book’s example will help researchers to see the value in designing new data collection efforts with an eye to using MSE. A persistent problem is that data sufficient to permit MSE for the purposes of granular analyses rarely exist. But Hoover Green demonstrates how MSE can guide or supplement granular analyses, an important role for this method.
Hoover Green shows the broad relevance of the El Salvador findings via Chapter 6, which presents case studies of multiple armed organizations involved in civil wars in the 1990s and early 2000s in Sierra Leone and Liberia (three in the former, four in the latter). The armed organizations in these wars are a hard test of the findings since some observers question the extent to which they had incentives to limit violence against civilians. As post-Cold War conflicts and ones that took place outside a superpower sphere of influence, the Mano River wars also differ significantly from El Salvador along lines that might affect repertoires of violence.
The extensive empirical evidence in The Commander’s Dilemma is at odds with alternative accounts and supports Hoover Green’s argument thoroughly. With this solid foundation, the book sets the agenda for research on repertoires of violence.[24] The book suggests several intriguing areas for inquiry going forward. First, to what extent are the choices and strategies of an opposition regarding restraint, whether through the use of political education or otherwise, determined by the government’s prior choices, or vice versa? The interaction between enemy organizations, each one’s anticipation of the other’s moves, may influence the range of options available. Second, might political education look fundamentally different according to the type of armed organization? Can status quo organizations, like government forces, rely on existing information already disseminated within society, concerning the purpose of the state and the penalties for challenging its sovereignty? Third, does political education as a restraining force on repertoires challenge, reinforce, or otherwise interact with performative logics of violence?[25] Fourth, if some effective war-fighting strategies or paths to victory in irregular war do not involve restraint,[26] does political education still restrict repertoires of violence? Can it also be used to broaden them?
More broadly, The Commander’s Dilemma also enters into the long-standing debate about violence and human nature. Hoover Green’s work belongs to the side of this debate that contends that violence and killing in war require socialization or are due to social pressures on the individual. The findings that support this come especially from military sociology. But that body of research has limits in its focus on the United States military. In other words, they apply to a military that is subordinate to civilian control within a democracy, extremely well-resourced, casualty-averse, and technologically sophisticated. How well do these findings apply to other militaries, against which the U.S. military is an outlier? How well do they apply to opposition armed groups created specifically to fight a civil war? And how well do they apply to combatants within a civil rather than international war? The other side of the debate acknowledges the importance of socialization but also underscore the elements of human behavior that tend naturally towards violence. In What It is Like to Go to War, for example, Karl Marlantes argues that “In combat you are already over some edge. You are in a fierce state where there is a primitive and savage joy in doing in your enemy” [27] Research has yet to address the questions about the applicability of the socialization findings, and social forces can certainly coexist with intrinsic ones. So it remains important to investigate the other side of the debate. To what extent do natural tendencies to violence, whether on the part of a subset of individuals or produced by situational factors in war, also help to inform our understanding of repertoires of violence?
An added highlight of the book is that Hoover Green distills a striking policy implication of the research. The link between political education and restrained repertoires of violence against civilians means that responsibility for war crimes falls on leaders should they fail to invest in this type of education for the rank-and-file (215-216). Looking to the possibility that states and the international community can undertake war crimes prosecutions on this basis, Hoover Green refers specifically to criminal responsibility. This is a controversial claim because it can be seen as privileging structural conditions over the commission of the criminal act itself—or at the very least putting them on an equal footing with it. The Commander’s Dilemma presents persuasive evidence that responsibility for violence against civilians due to the absence of political education rests, from a causal perspective, with leaders. And moral responsibility follows readily from that causal responsibility. But a full debate is warranted concerning whether criminal responsibility follows too. If criminal responsibility is adjudicated based on a full analysis of the roots of behavior, particularly structural ones, we might well find ourselves in an entirely different jurisprudential world than the one we currently inhabit. Importantly, this would be the case not only with respect to the outcomes of judicial proceedings, but also with respect to the principals upon which systems of justice are based.
Regardless of how the debate on its policy implications develops, the necessity of that debate is a testament to the insightfulness of Hoover Green’s book. For political science, The Commander’s Dilemma is also an important reminder. Developments in the discipline mean that the modal researcher invests the most attention and resources in a project’s causal identification strategy. But in this book, Hoover Green also shows that three other elements are integral to outstanding research. Scientific progress follows readily when researchers take care in measuring outcomes of interest, delve into the process that led to those outcomes, and investigate the actions of the humans involved.
Response by Amelia Hoover Green, Drexel University
Sustained attention to one’s work is a gift. I thank the four reviewers, both for their generosity in recognizing what has gone well and for offering critiques that push the broader work forward. I am also indebted to Elisabeth Wood for the productive synthesis her introduction offers. Given the space constraints of the roundtable format, below I consider just a few key themes that emerged across the reviews.
Conceptualization
Three of four reviewers discuss the book’s conceptualization of repertoires and restraint, and so I begin my discussion there. Jessica Trisko Darden examines the conceptualization of violence, generally (on which the concept of repertoire is necessarily premised). She notes that in the Salvadoran case I “neglect[] a significant portion of the violence that the FMLN did commit,” particularly violence against infrastructure and the campaign of targeted assassinations against mayors. This is certainly true; I also do not include violence that is institutional, structural, psychological, or emotional. One question that arises here is to what extent including indirect forms of violence would have altered my analysis or conclusions. In the specific case of the FMLN, campaigns against mayors are included (as civilian killings); sabotage is not. Neither campaign would change my view of FMLN violence as highly regimented, carefully planned, and controlled from the top of the organization. However, as Stathis Kalyvas also notes, the decision to exclude some forms of violence recapitulates some of the very problems with existing studies of civil war violence that prompted my focus on repertoires. Michael Boyle articulates a similar concern, emphasizing that I borrow the term “repertoire” from Charles Tilly[28], without hewing to Tilly’s conceptualization, which emphasized the interplay between violent and non-violent repertoires of contention. Moving forward, I hope that both my own research and that of others can more richly consider repertoires.
Kalyvas, despite acknowledging the potential benefits to the concept of repertoires of violence, is particularly forceful about their drawbacks: “[D]espite repeated claims that repertoires and levels of violence are analytically separable, they seem to be extremely close, and sometimes indistinguishable.” And, “it is counterintuitive to equate ‘narrow repertoires’ with ‘restrained violence,’ since a high number of killing can be in theory simultaneously an instance of a narrow repertoire and of unrestrained violence.” This passage highlights a key confusion for many readers regarding my conceptualization of ‘restraint.’ I define restraint only in passing, as low levels of violence, narrow repertoires of violence, or both (17; 13-16 for the discussion). Not emphasizing this definition was a mistake; in addition, the better word is probably ‘control,’ as in ‘controlled violence.’ To clarify: I would not consider a high level of killings, committed as part of a narrow repertoire of violence, to be “simultaneously an instance of a narrow repertoire and of unrestrained violence,” because that is not how I define restraint.
Armed groups that use restrained repertoires of violence frequently commit restrained (low) levels of violence, but they need not do so. Cases in which we observe restrained repertoires with high levels of violence—such as the Shining Path and the Liberation Tigers of Tamil Eelam (LTTE)—are particularly fascinating, as Kalyvas and Boyle note. Regarding the Shining Path in particular, I note (in a discussion of the Lucanamarca massacre): “Much of the violence committed by Shining Path cadres could be described as ‘overkill’—violence in excess of that required to kill, including performatively gruesome acts…Yet in most times and in most areas, Shining Path cadres committed very few acts of sexual violence relative to government forces” (3-4). The ritualistic nature of Shining Path violence can also be read as signifying control.
Theory
Jonah Schulhofer-Wohl asks several vital questions that test the boundaries and assumptions of the theory presented in The Commander’s Dilemma. I focus on just two here: (1) whether rebel strategies regarding restraint are “pre-determined by government actions,” i.e., whether institutions themselves are really doing causal work; and (2) whether “natural tendencies to violence” play a role. With respect to the first question, it is certainly clear that incumbent strategies shape insurgent strategies. It is less clear to me that rebel decisions (or “decisions”) about violence and restraint are sufficiently constrained by government actions to be “pre-determined.” (Although one might read the Salvadoran war itself as pre-determined by government actions—the state wiped out advocates for non-violent reform so effectively that the only opposition remaining was a nascent insurgency.) My sense of the (rather limited) evidence from outside my own cases is that variation in institutions, while constrained on several dimensions, nevertheless is seldom determined by structural factors.
Relatedly, in discussing the policy implications of The Commander’s Dilemma, Schulhofer-Wohl notes that my emphasis on institutions “can be seen as privileging structural conditions over the commission of the criminal act itself.” In my view this is true only if armed group leaders are extremely tightly constrained as they build institutions. In most circumstances, particularly in rebel groups, I do not believe this is so. However, a broader test of the relationship between structural factors outside the group and armed group institutions awaits the creation of more comprehensive cross-national data on such institutions. Several datasets track one or more institutional characteristics, but particularly for internal institutions such as political education, discipline, and advancement, the data are sparse and difficult to verify.
The second question obliquely challenges a key assumption of the theory, namely, that armed groups must make combatants violent. In some ways, this question echoes the work of Jeremy Weinstein (2007), whose theory suggests that “good” (restrained) types and “bad” types are born, not made.[29] I agree that there are individual, cultural, and group differences in propensity to violence. But the broader point of the social psychology is that “baseline” military training and socialization, when combined with combat exposure, ultimately swamp the effects of individual differences in propensity to violence. I view this as an important assumption underlying both the shift to understanding restraint as the puzzle to be explained and the concept of “restraint” as I have defined it, to include not only low overall levels of violence but narrow repertoires.
Evidence
Several of the reviewers suggested that the evidence supporting the theory falls short in places. In my view, much of the difficulty lies with my insistence on attempting a quantitative repertoire measure. To return to the conceptualization issues discussed above: an ideal conceptualization of “repertoires of violence” could never be adequately measured, and the measurable aspects of repertoire may not line up with an ideal conceptualization. In The Commander’s Dilemma I attempted to resolve this difficulty by using several measures, including survey data, interview evidence (when it was offered), and multiple systems estimates. But, as all the reviewers noted, there are places where the available evidence simply cannot adequately test the full range of hypotheses derived from the theory (summarized in Table 5.4). Trisko Darden draws attention to some key shortcomings in my discussion of the broad differences between the state and the FMLN: the possibility that different tactics of killing produce different reporting rates, and my failure to control for relative force size. These critiques are well taken.
Among the very few passages with which I truly take issue, it is important to note one in particular, since the implications of the critique stretch beyond my work to a broader research community that remains largely outside political science. Kalyvas writes, “Hoover Green does a good job of trying to measure these repertoires, primarily in El Salvador, but her method, the so-called Multiple Systems Estimation, ultimately falls short, as she herself recognizes.” There are a couple of questionable claims embedded in this sentence. The first is that the only method I employ to examine repertoires is Multiple Systems Estimation (“her method”); the second is that the method rather than (for example) the data or the application “falls short”; the third is that Multiple Systems Estimation is an untested, ad hoc, or questionable method (“so-called”). None of these claims is correct, although it is true that even the 60,000+ episodes of violence recorded across my four datasets were insufficient to produce MSE estimates for most non-lethal violence. Given the space constraints here, I refer readers to the more in-depth discussion of MSE (known as capture-recapture or multiple-recapture analysis in other contexts), its intellectual pedigree, and its broad acceptance among public health and demography scholars in a recent article.[30]
In closing, I would like to thank Elisabeth Wood and Tom Maddux for shepherding this roundtable, and Michael Boyle, Stathis Kalyvas, Jonah Schulhofer-Wohl, and Jessica Trisko Darden for their close and careful attention to The Commander’s Dilemma. As I move forward, their comments and critiques will inform attempts to gather broader, more systematic data as well as closer reflections on some key cases.
Amelia Hoover Green (Ph.D. Political Science, Yale University, 2011) is Associate Professor in the Department of Politics at Drexel University, Philadelphia, PA, and a Consultant to the Human Rights Data Analysis Group. Her recent publications include “Civilian killings and disappearances during civil war in El Salvador” (Demographic Research 2019, with co-author Patrick Ball), and the chapter “Successful Fieldwork for the Fieldwork-Hater,” in Krause and Szekely, eds., Stories from the Field (Columbia University Press, 2020). Her current projects include a simulation-based approach to assessing the impact of biased and incomplete violence data, and a cross-national data collection on armed groups’ internal institutions.
Notes
[1] The recent International Committee of the Red Cross book, The Roots of Restraint (Geneva: ICRC, 2018), cites her work on restraint at key points in the text.
[2] In an earlier article, Hoover Green finds (controlling for various other factors) that Communist rebels engage in lower levels of rape than other rebels even though lethal violence levels are similar. “The Commander’s Dilemma: Creating and Controlling Armed Group Violence,” Journal of Peace Research 53:5 (2016): 619-632, DOI: https://doi.org/10.1177/0022343316653645.
[3] See Francisco Gutiérrez Sanín and Elisabeth Jean Wood, “What Should We Mean by ‘Pattern of Political Violence’? Repertoire, Targeting, Frequency, and Technique,” Perspectives on Politics 15:1 (2017): 20-41, DOI: https://doi.org/10.1017/S1537592716004114.
[4] Amelia Hoover Green and Patrick Ball, “Civilian Killings and Disappearances During Civil War in El Salvador (1980-92),” Demographic Research 41:27 (2019): 781-814, DOI: https://doi.org/10.4054/DemRes.2019.41.27.
[5] Stathis Kalyvas, The Logic of Violence in Civil War (Cambridge: Cambridge University Press, 2006); Jeremy M. Weinstein, Inside Rebellion: The Politics of Insurgent Violence (Cambridge: Cambridge University Press, 2007).
[6] Dara Kay Cohen, Rape During Civil War (Ithaca: Cornell University Press, 2016).
[7] Charles Tilly, The Politics of Collective Violence (Cambridge: Cambridge University Press, 2003); Doug McAdam, Sidney Tarrow, and Charles Tilly, Dynamics of Contention (Cambridge: Cambridge University Press, 2001).
[8] Elizabeth Jean Wood, “Armed Groups and Sexual Violence: When is Wartime Rape Rare?” Politics & Society 37:1 (2009):131-161; Jessica A. Stanton, Violence and Restraint in Civil War: Civilian Targeting Under the Shadow of International Law (Cambridge: Cambridge University Press, 2016).
[9] See, inter alia, Erik Ching, Stories of Civil War in El Salvador: A Battle over Memory (Ashville: University of North Carolina Press, 2016); Russell Crandall, The Salvador Option: The United States in El Salvador (New York: Cambridge University Press, 2016); Jocelyn Viterna, Women in War: The Micro-Processes of Mobilization in El Salvador (New York: Oxford University Press, 2013); Elisabeth Wood, Insurgent Collective Action and Civil War in El Salvador (New York: Cambridge University Press, 2003).
[10] Belisario Betancur and Reinaldo Figueredo Planchart, From Madness to Hope: The 12-Year War in El Salvador: Report of the Commission on the Truth for El Salvador (Washington, D.C.: United States Institute of Peace, 2001).
[11] Central Intelligence Agency (CIA), “Assassinations and Kidnappings of Local Officials,” 2 January 1991, https://www.cia.gov/library/readingroom/docs/DOC_0000049393.pdf; and CIA, “El Salvador: Rebels Target Mayors,” 24 February 1989, https://www.cia.gov/library/readingroom/docs/DOC_0000049379.pdf.
[12] See, David L. Cingranelli and David L. Richards, “The Cingranelli and Richards (CIRI) Human Rights Data Project,” Human Rights Quarterly 32:2 (2010): 401-424; Mark Gibney, Linda Cornett, Reed Wood, Peter Haschke, Daniel Arnon, and Attilio Pisanò, The Political Terror Scale 1976-2017 (2018), http://www.politicalterrorscale.org.
[13] Raymond Bonner, “Salvador Bridge: Its Loss is Serious,” New York Times, 9 November 1981, A3; Alma Guillermoprieto, “Salvadoran Rebels Said to Gain Territory,” Washington Post, 10 November 1981, https://www.washingtonpost.com/archive/politics/1981/11/10/salvadoran-rebels-said-to-gain-territory/4d70835b-2e38-4df1-81c7-bed52b68ce89/.
[14] William Stanley, The Protection Racket State: Elite Politics, Military Extortion, and Civil War in El Salvador (Philadelphia: Temple University Press, 1996), 238.
[15] U.S. Department of State, “Memorandum from Secretary of State Vance to President Carter,” in Foreign Relations of the United States, 1977–1980, vol. XV, Document 421, https://history.state.gov/historicaldocuments/frus1977-80v15/d421.
[16] For instance, Miranda Alison, “Wartime Sexual Violence: Women’s Human Rights and Questions of Masculinity,” Review of International Studies 33:1 (2007): 75-90; Dara Kay Cohen, Rape during Civil War (Ithaca: Cornell University Press, 2016); Laura Eramian and Myriam Denov, “Is It Always Good to Talk? The Paradoxes of Truth-Telling by Rwandan Youth Born of Rape Committed during the Genocide,” Journal of Genocide Research 20:3 (2018): 372-391; Michele Leiby, “Wartime Sexual Violence in Guatemala and Peru,” International Studies Quarterly 53:2 (2009):445-468; Meredith Loken, “Rethinking Rape: The Role of Women in Wartime Violence,” Security Studies 20:1 (2016): 60-92; Megan H. MacKenzie, Female Soldiers in Sierra Leone: Sex, Security, and Post-Conflict Development (New York: New York University Press, 2012); Laura Sjoberg, Women as Wartime Rapists: Beyond Sensation and Stereotyping (New York: New York University Press, 2016); Elizabeth Wood, “Armed Groups and Sexual Violence: When is Wartime Rape Rare?” Politics & Society 37:1 (2009): 131-161.
[17] Stanley, 120-121.
[18] On how U.S. support affected state violence in El Salvador, see Jessica Trisko Darden, Aiding and Abetting: U.S. Foreign Assistance and State Violence (Stanford: Stanford University Press, 2020).
[19] Hal Brands, Latin America’s Cold War (Cambridge: Harvard University Press, 2012), 203.
[20] Jessica Trisko Darden, Alexis Henshaw, and Ora Szekely, Insurgent Women: Female Combatants in Civil Wars (Washington, D.C.: Georgetown University Press, 2019), 59-60.
[21] Note, by the way, that the use of various forms of nonlethal violence by Salvadoran troops includes instances of both strategic violence (arrests, disappearances, etc.) which substituted for killings and opportunistic violence, two logics of violence that diverge but tend to be compressed into one.
[22] Note that this is not the same as asking whether or why some commanders might ask their fighters for maximal violence (207).
[23] Jo Hyeran, Compliant Rebels: Rebel Groups and International Law in World Politic (New York: Cambridge University Press, 2015); Jessica A. Stanton, Violence and Restraint in Civil War: Civilian Targeting in the Shadow of International Law (New York: Cambridge University Press, 2016).
[24] See Francisco Gutiérrez-Sanín and Elisabeth Jean Wood, “What Should We Mean by ‘Pattern of Political Violence’? Repertoire, Targeting, Frequency, and Technique.” Perspectives on Politics 15:1 (2017): 20-41.
[25] See Lee Ann Fujii, “‘Talk of the Town’: Explaining Pathways to Participation in Violent Display.” Journal of Peace Research 54:5 (2017): 661-673.
[26] Yuri Zhukov, Yuri. 2007. “Examining the Authoritarian Model of Counter-insurgency: The Soviet Campaign Against the Ukrainian Insurgent Army.” Small Wars & Insurgencies 18:3 (2007): 439-466.
[27] Karl Marlantes, What It is Like to Go to War. (New York: Atlantic Monthly Press, 2011), 30.
[28] Charles Tilly, From Mobilization to Revolution (Reading: Addison-Wesley, 1978).
[29] Jeremy Weinstein, Inside Rebellion: The Politics of Insurgent Violence (New York: Cambridge University Press, 2007).
[30] Amelia Hoover Green and Patrick Ball. 2019. “Civilian Killings and Disappearances during Civil War in El Salvador (1980-1992),” Demographic Research 41:27 (2019): 781-814.
0 Replies