The Logic of Violence in Civil War

Cambridge Studies in Comparative Politics (2006)

🎙️ Comps Prep (Oral Comprehensive Exam)

• If sovereignty fragments into overlapping “monopolies of violence,” then civilian victimization becomes strategic because armed actors must deter defection while solving an identification problem through local information. So what for strategy: intelligence + protection are the decisive terrain of “control,” not add-ons. (pp. 88–89)

• When an actor has dominant-but-incomplete control, then selective violence will peak there because civilians can more safely denounce individuals and selective violence is jointly produced through these transactions. So what for strategy: consolidate control and incentive-compatible information flows instead of relying on “front-line” coercion. (p. 13; p. 204)

• When individual-level information is unavailable, then actors shift toward indiscriminate violence, but it tends to backfire because civilians collaborate with the actor that best guarantees their security. So what for strategy: indiscriminate coercion often undermines control; compete on credible protection and governance. (p. 147; p. 168)

• Bridge: This aligns with Biddle on how information and constraints drive irregular methods, and it sharpens Mao/Galula-style population-control claims by specifying where control makes selective violence feasible (linkage is inference from themes, not an explicit comparison).

Online Description

Kalyvas develops a general theory explaining when and why civilians are targeted in civil war, emphasizing how territorial control, information, and civilian collaboration shape the use of selective and indiscriminate violence. He tests the theory with comparative evidence and a micro-level study of the Greek Civil War, showing that violence is patterned (not random) and closely linked to local control dynamics. (p. i; p. 13)

Author Background

• Arnold Wolfers Professor of Political Science at Yale University; Director of Yale’s Program on Order, Conflict, and Violence. (p. i)

• Previously taught at the University of Chicago, New York University, and Ohio State University; also served as a visiting professor at the Central European University. (p. i)

• Author of The Rise of Christian Democracy in Europe and recipient of major awards in political science (as listed in the book’s author bio). (p. i)

---

60‑Second Brief

• Core claim (1–2 sentences):

  Civil war violence against civilians follows a strategic logic rooted in control and collaboration: armed actors use violence to deter defection and shape civilian behavior under conditions of limited information. Selective violence is often “jointly produced” through civilian denunciation and is patterned by local control. (p. 13; p. 147)

• Causal logic in a phrase:

  Control → information & incentives → denunciation/defection → (selective vs indiscriminate) violence → shifts in control.

• Why it matters for IW / strategic competition (2–4 bullets):

  • “Control” is not just terrain; it is a social-political condition that changes what kinds of targeting are feasible (and effective). (pp. 120–121; p. 173)

  • Civilian behavior is not noise—civilian agency (especially via denunciation) is a core input into the violence process. (p. 13)

  • Indiscriminate violence is frequently self-defeating because it pushes civilians toward whichever side credibly provides security. (p. 147; p. 168)

  • Operational implication: map control gradients and information channels, not just enemy order of battle. (p. 204)

• Best single takeaway (1 sentence):

  In irregular war, who can credibly protect civilians and elicit usable information largely determines where violence concentrates and how control evolves. (p. 13; p. 168)

Course Lens

• How does this text define/illuminate irregular warfare?

  • Treats civil war violence as a form of irregular war rooted in fragmented sovereignty and competition to govern populations through coercion and information. (pp. 88–89)

  • Shifts the IW center of gravity from “battle” to control regimes, civilian incentives, and the identification problem. (pp. 88–89; p. 204)

• What does it imply about power/control, success metrics, and timeline in IW?

  • Power/control: not binary; control varies by zone and shapes which strategies (selective vs indiscriminate) are feasible. (pp. 120–121)

  • Success metrics: should track control gradients, defection/collaboration, denunciation flows, and civilian security—not only enemy attrition. (p. 204; p. 168)

  • Timeline: violence and control co-evolve; short-term coercion can undermine long-term control if it destroys incentives for collaboration or truthful information. (p. 147; p. 168)

• How does it connect to strategic competition?

  • Provides a micro-foundation for how rivals contest governance and legitimacy “from below”: control is built via protection, information, and selective coercion—often below the threshold of conventional war. (p. 13; p. 245)

---

Seminar Questions (from syllabus)

• Why is violence used in civil war?

• How does Kalyvas treat macro/meso/micro levels?

• What is selective violence? Where is violence most likely?

• Relationship between collaboration and control?

• How does this explain local population behavior?

• Why do civil wars show extreme violence?

✅ Direct Responses to Seminar Questions

• Q: Why is violence used in civil war?

  • A:

    • Violence in civil war is not analytically identical to the causes of civil war; it requires its own causal account. (p. 21)

    • In civil war violence (as distinct from genocide/extermination), armed actors often intend to govern civilians, so violence becomes part of a strategic interaction over compliance and defection. (p. 31)

    • A core logic is deterring defection under uncertainty: civilians can shift support/resources, so actors use coercion to shape behavior. (p. 13; p. 31)

    • Selective violence is frequently produced through transactions with civilians (e.g., denunciation) that supply the identification needed to punish defectors. (p. 13)

    • Indiscriminate violence emerges when actors lack reliable information, but the book argues it is often counterproductive because it shifts civilians toward the side that best guarantees security. (p. 147; p. 168)

• Q: How does Kalyvas treat macro/meso/micro levels?

  • A:

    • He explicitly argues for disaggregation: macro-level variables alone cannot reveal the mechanisms producing violence at the ground level. (p. 11; p. 386)

    • Macro: a “master cleavage” structures conflict, but it does not mechanically determine local violence patterns. (p. 365)

    • Meso: territorial control regimes (zones of control) and institutional settings shape information availability and incentives. (pp. 120–121; p. 353)

    • Micro: individual incentives and local rivalries matter because violence is “jointly produced” by armed actors and civilians (notably via denunciation). (p. 13)

    • Methodologically, he combines theory with comparative and microcomparative evidence to connect levels without collapsing them. (p. 245; p. 328)

• Q: What is selective violence? Where is violence most likely?

  • A:

    • Definition: “Violence is selective when there is an intention to ascertain individual guilt.” (p. 142)

    • Selective violence depends on information (who defected/collaborated), and the book emphasizes the role of denunciation as a key information mechanism. (p. 13; p. 173)

    • Core prediction: selective violence should concentrate where one actor has near-hegemonic / dominant-but-incomplete control, because civilians can safely provide information and the actor can act on it. (p. 13; p. 204)

    • The framework implies a non-linear relationship between control and selective violence: violence is not necessarily highest at the front line; contested zones can be relatively “quiet” because denunciation is too risky. (p. 173–174; p. 204)

    • Empirical finding (microcomparative): “areas largely but not totally controlled” were more likely to see selective homicidal violence than fully controlled areas. (p. 328)

• Q: Relationship between collaboration and control?

  • A:

    • Collaboration is defined behaviorally: actions that benefit one actor at the expense of its rival; defection is active collaboration with the rival. (pp. 92–93)

    • “Support” is conceptually distinct from collaboration: preferences do not map cleanly onto behavior under coercion and uncertainty. (p. 90)

    • Control is the capacity to set rules and deter defection; it shapes civilian incentives to collaborate, defect, and denounce because it structures the expected costs of retaliation. (pp. 120–121; p. 204)

    • Collaboration and control are mutually reinforcing and endogenously linked: control affects information and incentives; information enables selective violence; selective violence can deter defection and thereby strengthen control. (p. 13; p. 204)

    • The “identification problem” ties them together: without civilian information, control is hard to consolidate, pushing actors toward cruder coercion. (pp. 88–89; p. 171)

• Q: How does this explain local population behavior?

  • A:

    • Civilians are not passive; they are strategic actors whose behavior responds to security, incentives, and expected retaliation, making them central to the “joint production” of violence. (p. 13)

    • Many civilians collaborate with the actor that best guarantees their security (not necessarily the one they “prefer” politically). (p. 168)

    • In highly contested areas, civilians may avoid denunciation/collaboration because retaliation risks are high—contributing to the possibility of “oases of peace.” (p. 173–174; p. 204)

    • Denunciation can be political and personal: civil war “privatizes politics,” allowing private disputes to be reframed through the war’s categories and institutions. (p. 351; p. 365)

    • Because denunciation can be malicious, information flows are not automatically reliable; institutional settings shape incentives to denounce truthfully or falsely. (p. 353)

• Q: Why do civil wars show extreme violence?

  • A:

    • Kalyvas critiques “barbarism” explanations that treat violence as random breakdown; he evaluates multiple accounts (breakdown, transgression, polarization, technology of warfare). (p. 82)

    • He argues a “technology of warfare” / security-centered account best provides foundations for theorizing violence as strategic within irregular war conditions. (p. 82)

    • Extreme violence can reflect strategic messaging and coercion (even when it appears absurd), but its form varies with aims and local contexts. (p. 29; p. 168)

    • Civil war violence is shaped by strategic interaction and civilian choice (the option to switch support), which can escalate coercion as actors compete for control. (p. 31; p. 13)

    • The overall takeaway is that violence has logic and cannot be reduced to “opaque madness,” even when it reaches extreme forms. (p. 389)

---

Chapter-by-Chapter Breakdown

Chapter 0: Introduction (pp. 1–15)

• One-sentence thesis: Civil war violence is patterned and strategic, best explained by disaggregating to local control, collaboration, and information dynamics rather than treating violence as random “barbarism.” (p. 11; p. 13)

• What happens / what the author argues (5–10 bullets):

  • Poses “puzzles” about when/where violence occurs and why it varies across space and time. (p. 1)

  • States the book’s goal: explain violence against civilians with a general theory rooted in control and collaboration. (p. 11–13)

  • Emphasizes the need to link macro-level structures with micro-level mechanisms. (p. 11)

  • Introduces the idea that selective violence is produced through transactions with civilians and depends on control. (p. 13)

  • Previews the empirical strategy: comparative patterns plus micro-level evidence from the Greek Civil War. (p. i; p. 13)

• Key concepts introduced (0–5):

  • Disaggregation; joint production; control–violence link. (p. 11; p. 13)

• Evidence / cases used:

  • Road map points to comparative evidence and Greek microcomparative study. (p. i; p. 13)

• IW / strategy relevance (2–4 bullets):

  • Frames IW as competition for governance and information, not just kinetic exchanges. (p. 13)

  • Implies that “front line” intuitions can mislead; control gradients predict violence patterns. (p. 13)

• Links to seminar questions:

  • 1, 2, 3, 4, 5

• Notable quotes (0–2):

  • “Selective violence is the result of transactions between political actors and civilians.” (p. 13)

Chapter 1: Concepts (pp. 16–31)

• One-sentence thesis: Define the phenomenon precisely—civil war violence as multilateral coercion aimed at governing—then specify conceptual boundaries distinguishing it from repression and genocide. (p. 17; p. 31)

• What happens / what the author argues (5–10 bullets):

  • Defines civil war as armed combat within a recognized sovereign entity among parties subject to common authority at the outset. (p. 17)

  • Argues the causes of violence cannot be collapsed into the causes of civil war. (p. 21)

  • Builds a typology of mass political violence combining the actor’s aim (govern vs exterminate) and whether violence is unilateral vs multilateral. (p. 29)

  • Defines “civil war violence” as multilateral violence where at least one actor intends to govern, creating strategic interaction with civilians. (p. 31)

  • Emphasizes that civil war violence is shaped by civilians’ option to shift support, making violence consequential and interactive. (p. 31)

• Key concepts introduced (0–5):

  • Civil war (definition); typology of mass political violence; scope conditions. (p. 17; p. 29–31)

• Evidence / cases used:

  • Illustrative examples for typology (e.g., genocidal/reciprocal extermination contexts) and multilateral coercion dynamics. (p. 31)

• IW / strategy relevance (2–4 bullets):

  • Clarifies when “civil war violence” logic applies (governance competition) versus when extermination logic dominates. (p. 29–31)

  • Sets conditions for when civilian “choice” matters strategically. (p. 31)

• Links to seminar questions:

  • 1, 3, 6

• Notable quotes (0–2):

  • “The causes of violence in civil war cannot be subsumed under the causes of civil war.” (p. 21)

Chapter 2: Pathologies (pp. 32–51)

• One-sentence thesis: Much scholarship misreads civil war violence due to recurring biases and measurement problems; disaggregation is the corrective. (p. 49; p. 11)

• What happens / what the author argues (5–10 bullets):

  • Diagnoses interpretive biases that treat civil war violence as irrational, purely partisan, or purely “political” in a narrow sense. (p. 49–51)

  • Highlights an “urban bias” that overlooks rural/local dynamics where control and collaboration are negotiated. (p. 49)

  • Argues selection bias and overaggregation can produce misleading “macro” stories about violence. (p. 49–50)

  • Emphasizes that many core variables (e.g., control) are hard to observe and measure, pushing researchers toward proxies that can mislead. (p. 49–50)

  • Concludes that careful disaggregation is required before theorizing about macro-patterns. (p. 49–50)

• Key concepts introduced (0–5):

  • Bias/selection problems; overaggregation; data quality issues. (p. 49–50)

• Evidence / cases used:

  • Methodological critique drawing on patterns in existing civil war research. (pp. 32–51)

• IW / strategy relevance (2–4 bullets):

  • Warns strategists against “top-down” assessments that miss local control gradients and incentives. (p. 49–50)

  • Suggests operational metrics should be local and disaggregated to avoid self-deception. (p. 49–50)

• Links to seminar questions:

  • 2, 6

• Notable quotes (0–2):

  • TBD (none selected)

Chapter 3: Barbarism (pp. 52–86)

• One-sentence thesis: “Barbarism” is an inadequate catch-all; extreme violence is better explained by mechanisms tied to security and the technology of irregular war. (p. 82)

• What happens / what the author argues (5–10 bullets):

  • Reviews “breakdown” accounts that treat violence as collapse of norms/institutions. (pp. 52–86)

  • Examines “transgression” accounts focusing on escalation spirals and rule-breaking dynamics. (pp. 52–86)

  • Evaluates “polarization” accounts linking violence to ideological/ethnic sorting and intensified cleavages. (pp. 52–86)

  • Assesses a “technology of warfare” approach emphasizing insecurity and irregular war conditions as drivers of civilian targeting. (p. 82)

  • Argues this technology-of-warfare thesis provides the most appropriate foundation for the book’s theory of violence. (p. 82)

• Key concepts introduced (0–5):

  • Barbarism (as critique category); technology-of-warfare framing. (p. 82)

• Evidence / cases used:

  • Comparative discussion of civil wars and irregular conflicts used to assess competing explanations. (pp. 52–86)

• IW / strategy relevance (2–4 bullets):

  • Reframes “atrocity” as potentially strategic under insecurity/identification constraints—relevant for diagnosing adversary behavior. (p. 82)

  • Suggests prevention/mitigation must target security/information conditions, not only “messaging.” (p. 82)

• Links to seminar questions:

  • 1, 6

• Notable quotes (0–2):

  • TBD (none selected)

Chapter 4: A Theory of Irregular War I: Collaboration (pp. 87–110)

• One-sentence thesis: Civil war creates fragmented sovereignty and an identification problem, making collaboration/defection and information the central stakes of irregular war. (pp. 88–89)

• What happens / what the author argues (5–10 bullets):

  • Defines civil war as a condition of fragmented sovereignty with rival systems of rule. (p. 88)

  • Introduces the “identification problem”: actors cannot easily distinguish loyalists from defectors and need local, private information. (pp. 88–89)

  • Distinguishes support (preferences) from collaboration (behavior) to avoid conflating beliefs with observable actions. (p. 90)

  • Defines collaboration as behavior benefiting one side at the expense of the rival; defines defection as active collaboration with the rival. (pp. 92–93)

  • Categorizes defection types (e.g., individual/group, voluntary/involuntary) as groundwork for later control/violence logic. (p. 105)

  • Emphasizes the role of institutional settings in shaping collaboration opportunities and costs. (p. 102)

• Key concepts introduced (0–5):

  • Fragmented sovereignty; identification problem; support vs collaboration; defection typology. (pp. 88–90; p. 105)

• Evidence / cases used:

  • Comparative examples illustrating collaboration/defection under irregular war conditions. (pp. 87–110)

• IW / strategy relevance (2–4 bullets):

  • Puts “human intelligence” (and its distortions) at the center of IW outcomes. (pp. 88–89)

  • Clarifies why coercion alone cannot substitute for reliable identification and governance. (pp. 88–89)

• Links to seminar questions:

  • 1, 3, 4, 5

• Notable quotes (0–2):

  • TBD (none selected)

Chapter 5: A Theory of Irregular War II: Control (pp. 111–145)

• One-sentence thesis: Territorial control is the key independent variable shaping collaboration and the feasibility of selective vs indiscriminate violence through its effects on incentives and information. (pp. 120–121; p. 142)

• What happens / what the author argues (5–10 bullets):

  • Defines control as an actor’s capacity to enforce rules and deter defection across territory. (pp. 120–121)

  • Introduces a five-zone conceptualization of control (from full incumbent to full insurgent control, with intermediate zones). (pp. 120–121)

  • Explains how control structures collaboration choices by altering expected costs (retaliation) and benefits (protection). (p. 121)

  • Shows how control affects the “allocation of collaboration” and survival behavior among civilians. (pp. 111–145)

  • Links control to military constraints and feasible coercion options in irregular war. (p. 121)

  • Defines the selective/indiscriminate violence distinction and treats it as ideal-typical. (p. 142)

• Key concepts introduced (0–5):

  • Control zones; selective vs indiscriminate violence (conceptual distinction). (pp. 120–121; p. 142)

• Evidence / cases used:

  • Comparative illustrations of control regimes and their civilian consequences. (pp. 111–145)

• IW / strategy relevance (2–4 bullets):

  • Provides an operationalizable way to think about “control” beyond binary cleared/uncleared maps. (pp. 120–121)

  • Suggests violence patterns are diagnostic signals of control regimes and information availability. (p. 142)

• Links to seminar questions:

  • 1, 3, 4, 5

• Notable quotes (0–2):

  • “Violence is selective when there is an intention to ascertain individual guilt.” (p. 142)

Chapter 6: A Logic of Indiscriminate Violence (pp. 146–172)

• One-sentence thesis: Indiscriminate violence is often a substitute for missing information, yet it is usually counterproductive because it drives civilians toward the side that provides security. (p. 147; p. 168)

• What happens / what the author argues (5–10 bullets):

  • Defines and explains the incidence of indiscriminate violence in civil wars. (pp. 146–172)

  • Analyzes how information constraints and deterrence logic can generate indiscriminate violence. (p. 171)

  • Presents the core claim that indiscriminate violence is counterproductive in civil war (as a proposition). (p. 147)

  • Explains why, despite being counterproductive, actors still use indiscriminate violence (constraints, misperception, limited capabilities). (pp. 146–172)

  • Emphasizes the “association” mechanism: when specific identification is impossible, coercion targets categories/areas, shifting responsibility to civilians to police themselves. (p. 171)

  • Connects civilian behavior to security: civilians tend to collaborate with the actor that best guarantees their security. (p. 168)

• Key concepts introduced (0–5):

  • Indiscriminate violence as information substitute; protection–violence tradeoff. (p. 168; p. 171)

• Evidence / cases used:

  • Comparative examples of indiscriminate violence and its consequences; includes conceptual figure relating civilian behavior to violence/protection. (p. 167)

• IW / strategy relevance (2–4 bullets):

  • Strong warning against “collective punishment” logic as an IW strategy: it often undermines collaboration and control. (p. 147; p. 168)

  • Points to protection as the decisive competitive offering in contested governance environments. (p. 168)

• Links to seminar questions:

  • 1, 3, 5, 6

• Notable quotes (0–2):

  • “Proposition 2: Indiscriminate violence is counterproductive in civil war.” (p. 147)

Chapter 7: A Theory of Selective Violence (pp. 173–209)

• One-sentence thesis: Selective violence depends on information (often via denunciation) and is most likely where one actor has dominant-but-incomplete control—producing a non-linear violence–control relationship. (p. 173; p. 204)

• What happens / what the author argues (5–10 bullets):

  • States that selective violence is not restricted to the front line; it can be more likely in “safe” zones of control. (p. 173)

  • Argues the key constraint is information: identifying individual defectors is difficult without civilian participation. (p. 173–174)

  • Places denunciation at the center of selective violence production. (p. 173; p. 13)

  • Discusses how the logic operates (and differs) in ethnic civil wars. (p. 181)

  • Explores whether selective violence is feasible when retaliation risks and information problems are severe. (p. 174)

  • Presents a model linking control to defection and denunciation equilibria and derives predictions for where selective violence concentrates. (p. 204–205)

  • Flags caveats/limits to the model’s applicability. (p. 209)

• Key concepts introduced (0–5):

  • Denunciation equilibria; selective violence/control prediction; “oases of peace” logic. (p. 204)

• Evidence / cases used:

  • Theoretical model plus illustrative cases; sets up empirical testing in later chapters. (pp. 173–209)

• IW / strategy relevance (2–4 bullets):

  • Explains why “front-line heat maps” can mislead: violence can be highest where control is dominant but incomplete. (p. 173; p. 204)

  • Highlights the strategic centrality (and moral hazard) of denunciation networks in IW. (p. 173; p. 353)

• Links to seminar questions:

  • 1, 3, 4, 5

• Notable quotes (0–2):

  • “Selective violence is most likely to take place in zones that are safely controlled by one actor.” (p. 173)

Chapter 8: Empirics I: Comparative Evidence (pp. 210–245)

• One-sentence thesis: Comparative evidence supports the control-based theory by showing systematic links between control shifts and violence patterns across conflicts and contexts. (p. 245; pp. 219–221)

• What happens / what the author argues (5–10 bullets):

  • Develops measurement strategies for “control” and its shifts over time. (p. 210; p. 216)

  • Operationalizes the five-zone control framework and illustrates what control looks like on the ground. (pp. 219–221)

  • Uses comparative cases to examine violence under full control vs partial control vs contested control. (p. 221)

  • Highlights that “no control” is a meaningful condition for a given actor (control by the rival) and shapes violence opportunities. (p. 222)

  • Synthesizes control–violence relationships, reframing violence as an outcome of competition for control. (p. 245)

  • Suggests insurgency can be understood as alternative state-building (governance competition). (p. 245)

• Key concepts introduced (0–5):

  • Measurement of control; comparative operationalization of control zones. (p. 216; pp. 219–221)

• Evidence / cases used:

  • Illustrative cases for control mapping (e.g., Algiers, Mau Mau passive wing, China’s “red heart/white skin,” Maoist Nepal). (pp. 219–221)

• IW / strategy relevance (2–4 bullets):

  • Provides a framework to translate qualitative “control” descriptions into operational assessments. (pp. 219–221)

  • Reorients strategy toward governance competition and control consolidation as the core conflict dynamic. (p. 245)

• Links to seminar questions:

  • 2, 3, 4, 5

• Notable quotes (0–2):

  • TBD (none selected)

Chapter 9: Empirics II: Microcomparative Evidence (pp. 246–329)

• One-sentence thesis: A micro-level study of Greek villages validates the core predictions about control and violence while revealing important variation, mispredictions, and the limits of macro inference. (p. 328; pp. 323–325)

• What happens / what the author argues (5–10 bullets):

  • Lays out a research design for microcomparative study and why village-level analysis is essential. (p. 246)

  • Provides background on the Greek Civil War and the Argolid region; defines periods and actors. (pp. 246–266)

  • Measures violence and control over time and geography, distinguishing selective vs indiscriminate violence. (p. 267–279)

  • Uses quantitative tests (regressions) to assess predicted links between control zones and violence frequency/intensity. (p. 288–289)

  • Adds qualitative analysis to explain outliers and mispredicted cases (e.g., return to specific villages). (p. 305; p. 313)

  • Replicates logic in another Greek region (Almopia) and conducts out-of-sample tests across Greece. (p. 322–323)

  • Concludes that dominant-but-incomplete control is associated with selective homicidal violence, while full control is less violent; also stresses that forms of violence vary across time/space. (p. 328)

• Key concepts introduced (0–5):

  • Microcomparative method; village typologies; predicted vs observed violence mapping. (p. 281; p. 316)

• Evidence / cases used:

  • Greek Civil War (Argolid; Almopia; additional villages across Greece); mixed sources (archival + interviews). (pp. 246–329; Appendix A)

• IW / strategy relevance (2–4 bullets):

  • Demonstrates why “local politics” and civilian information channels are operationally decisive. (p. 328)

  • Shows that both violence and nonviolence can be strategic outcomes of control regimes—affecting targeting, governance, and legitimacy. (p. 323–325)

• Links to seminar questions:

  • 1, 2, 3, 4, 5

• Notable quotes (0–2):

  • “Areas largely but not totally controlled … were significantly more likely to see selective homicidal violence.” (p. 328)

Chapter 10: Intimacy (pp. 330–363)

• One-sentence thesis: Civil war violence is often intimate and locally mediated; denunciation connects politics and personal motives, and institutional settings shape the reliability and consequences of denunciation. (p. 351; p. 353)

• What happens / what the author argues (5–10 bullets):

  • Defines “intimate violence” and links it to the local social fabric of civil war. (p. 330)

  • Builds a sociology of denunciation: why civilians denounce and how denunciation becomes actionable information. (p. 333)

  • Explores the range of malicious denunciation and why civil war creates opportunities for private abuse of political categories. (p. 347)

  • Argues denunciation can be understood as the “dark face of social capital,” not only as political behavior. (p. 353)

  • Emphasizes that institutions (and their rules) shape denunciation incentives and outcomes. (p. 359)

  • Concludes that civil war is not only political; it privatizes politics and reshapes local disputes through war institutions. (p. 351)

• Key concepts introduced (0–5):

  • Intimate violence; malicious denunciation; denunciation as social capital’s “dark face.” (p. 351; p. 353)

• Evidence / cases used:

  • Cross-case examples of denunciation dynamics; micro evidence consistent with the Greek case logic. (pp. 330–363)

• IW / strategy relevance (2–4 bullets):

  • Warning for IW targeting: intelligence can be systematically biased by local rivalries and incentives. (p. 353)

  • Institutional design (procedures, review, accountability) matters for controlling the moral hazard of denunciation. (p. 359)

• Links to seminar questions:

  • 1, 4, 5, 6

• Notable quotes (0–2):

  • “Civil war also works the other way around: it privatizes politics.” (p. 351)

Chapter 11: Cleavage and Agency (pp. 364–387)

• One-sentence thesis: Civil wars transform and are transformed by local cleavages and agency; macro cleavages do not determine micro violence, making disaggregation analytically mandatory. (p. 365; p. 386)

• What happens / what the author argues (5–10 bullets):

  • Argues civil war links a “master cleavage” to local cleavages in a complicated, non-deterministic way. (p. 365)

  • Shows how local actors can shape the war’s trajectory by leveraging alliances and reinterpreting local disputes through the master cleavage. (p. 365)

  • Develops an agency-centered view (“Kto kovo?”) about who drives violence and political outcomes at the local level. (p. 367)

  • Explores center–periphery dynamics and how alliances structure local violence and collaboration. (p. 384–385)

  • Concludes with a strong methodological claim: macrovariables alone cannot reveal micromechanisms; disaggregation must come first. (p. 386)

• Key concepts introduced (0–5):

  • Master cleavage vs local cleavages; locus of agency; alliance dynamics. (p. 365; p. 384–385)

• Evidence / cases used:

  • Comparative illustrations plus integration with the Greek microcomparative evidence. (pp. 364–387)

• IW / strategy relevance (2–4 bullets):

  • Highlights that “local politics” can hijack national strategy; planners must treat local agency as a first-order variable. (p. 365)

  • Suggests coalition/partner dynamics are inherently local and incentive-driven, with violence as an instrument within those bargains. (p. 384–385)

• Links to seminar questions:

  • 2, 4, 5, 6

• Notable quotes (0–2):

  • “We cannot get at the micromechanisms of violence from the analysis of macrovariables.” (p. 386)

Chapter 12: Conclusion (pp. 388–392)

• One-sentence thesis: The book’s core contribution is showing that violence in civil war has a logic grounded in control, collaboration, and information—best understood through disaggregation and microfoundations. (p. 389; p. 386)

• What happens / what the author argues (5–10 bullets):

  • Synthesizes the theory: violence is strategic interaction among actors competing to govern and extract collaboration under uncertainty. (p. 389)

  • Reasserts the joint production of violence and the centrality of civilian agency and information. (p. 13; p. 389)

  • Emphasizes how control predicts variation in violence forms and where violence concentrates. (p. 204; p. 328)

  • Reinforces the methodological point: disaggregation precedes macro-interpretation. (p. 386; p. 389)

• Key concepts introduced (0–5):

  • Synthesis of control–collaboration–violence logic; disaggregation as method. (p. 386; p. 389)

• Evidence / cases used:

  • Wrap-up integrating comparative and Greek micro evidence. (pp. 388–392)

• IW / strategy relevance (2–4 bullets):

  • Encourages IW strategists to treat control, protection, and information as the core levers shaping violence and legitimacy. (p. 168; p. 204)

  • Suggests operational assessments must focus on local control and incentive structures—not just narratives or force ratios. (p. 386)

• Links to seminar questions:

  • 1, 2, 3, 4, 5, 6

• Notable quotes (0–2):

  • “This book has marshaled theory and evidence to demonstrate that there is logic to violence in civil war.” (p. 389)

Appendix A: Data Sources (pp. 393–411)

• One-sentence thesis: Documents and triangulates the study’s data sources, emphasizing the value and limits of combining oral and archival evidence for micro-level civil war analysis. (p. 393)

• What happens / what the author argues (5–10 bullets):

  • Describes oral sources (interviews) and archival sources used for coding violence and control. (p. 395–397)

  • Discusses source criticism: reliability, bias, and the value of cross-checking accounts. (p. 399–400)

  • Explains how micro-level data enable measurement of control and violence variation across villages. (p. 406)

• Key concepts introduced (0–5):

  • Data triangulation; micro-level coding foundations. (p. 399–400)

• Evidence / cases used:

  • Greek Civil War interviews and archives; lists of interviews. (p. 397)

• IW / strategy relevance (2–4 bullets):

  • Reinforces that granular mapping of control/violence requires mixed methods and attention to informant incentives. (p. 399–400)

• Links to seminar questions:

  • 2, 5

• Notable quotes (0–2):

  • TBD (none selected)

Appendix B: Coding Protocols (pp. 412–422)

• One-sentence thesis: Provides coding rules and variable definitions that operationalize control, violence types, and contextual factors for the Argolid and broader Greek village datasets. (p. 412)

• What happens / what the author argues (5–10 bullets):

  • Lists villages included and how ecological clusters and variables are defined. (p. 413–416)

  • Specifies independent variables and coding decisions supporting the statistical analysis. (p. 415)

  • Describes sampling across Greece for out-of-sample tests. (p. 416)

• Key concepts introduced (0–5):

  • Operational definitions and coding discipline. (p. 415)

• Evidence / cases used:

  • Argolid village list; broader Greek village survey list. (p. 413; p. 416)

• IW / strategy relevance (2–4 bullets):

  • Demonstrates how “control” can be operationalized for empirical testing—useful for designing IW assessment frameworks. (pp. 412–416)

• Links to seminar questions:

  • 2, 3

• Notable quotes (0–2):

  • TBD (none selected)

Appendix C: Timeline of Conflicts (pp. 423–426)

• One-sentence thesis: Supplies a timeline to situate conflicts discussed and to anchor comparative references used throughout the book. (p. 423)

• What happens / what the author argues (5–10 bullets):

  • Lists conflicts and key dates for comparative orientation. (p. 423–426)

• Key concepts introduced (0–5):

  • None (reference tool).

• Evidence / cases used:

  • Conflict timelines. (pp. 423–426)

• IW / strategy relevance (2–4 bullets):

  • Useful for quickly situating comparative references when tracing mechanisms across cases.

• Links to seminar questions:

  • 2

• Notable quotes (0–2):

  • TBD (none selected)

---

Theory / Framework Map

• Level(s) of analysis:

  • Macro (master cleavage / war-wide political structure), meso (local control regimes), micro (civilian incentives, denunciation, defection). (p. 11; p. 365)

• Unit(s) of analysis:

  • Political actors (incumbent vs insurgent), local communities (villages/territorial units), civilians as agents (collaborators/denouncers). (p. 13; pp. 246–252)

• Dependent variable(s):

  • Form and intensity of violence against civilians, especially selective vs indiscriminate violence; patterns of defection/denunciation. (p. 142; p. 204)

• Key independent variable(s):

  • Territorial control (five-zone scheme), information availability, institutional setting shaping denunciation incentives. (pp. 120–121; p. 353)

• Mechanism(s):

  • Control changes expected costs/benefits of collaboration and retaliation → shapes civilian behavior (including denunciation) → enables or constrains selective violence → violence deters defection and feeds back into control. (p. 13; p. 204)

• Scope conditions / where it should NOT apply:

  • Applies to “civil war violence” where at least one actor intends to govern (not exterminate/deport), and civilians can shift support. (p. 31)

  • Less applicable where the aim is group elimination (genocide/mass deportation) or where identities are fully observable and “identification” is not the main constraint. (p. 29–31; p. 181)

• Observable implications / predictions:

  • Selective violence peaks under dominant-but-incomplete control; full control is less violent; contested zones can show “oases of peace.” (p. 13; p. 204; p. 328)

  • Indiscriminate violence is more likely when identification is difficult, but tends to be counterproductive by pushing civilians toward the side that provides security. (p. 147; p. 168)

Key Concepts & Definitions (author’s usage)

• Civil war

  • Definition: “armed combat within the boundaries of a recognized sovereign entity between parties subject to a common authority at the outset of hostilities.” (p. 17)

  • Role in argument: Defines the universe of cases and the political condition within which violence is analyzed.

  • Analytical note: Distinguish from interstate war and from unilateral repression; identify common authority at onset.

• Civil war violence (scope condition)

  • Definition: Multilateral violence where at least one actor intends to govern rather than exterminate/deport; civilians can shift support. (p. 31)

  • Role in argument: Establishes when strategic interaction with civilians is central.

  • Analytical note: This is where “control,” collaboration, and information become decisive.

• Typology of mass political violence

  • Definition: Aims (govern vs exterminate) × production (unilateral vs multilateral) → ideal types (state terror; genocide/mass deportation; civil war violence; reciprocal extermination). (p. 29)

  • Role in argument: Clarifies what the book is and is not explaining.

  • Analytical note: Useful for diagnosing whether a case is drifting from “govern” to “exterminate” logics.

• Fragmented sovereignty

  • Definition: Civil war creates “rival and overlapping monopolies of violence” and “competing systems of rules and obedience.” (p. 88)

  • Role in argument: Explains why control is partial and contested; why civilian choice matters.

  • Analytical note: Operationalize through observable governance functions and enforcement capacity.

• Identification problem

  • Definition: Difficulty of distinguishing friends, foes, and defectors under irregular war; reliance on “local and private information.” (pp. 88–89)

  • Role in argument: Drives demand for civilian information and shapes violence type.

  • Analytical note: Expect greater reliance on denunciation and selective violence where identification can be solved.

• Support

  • Definition: Relative preferences over political actors; distinct from collaboration. (p. 90)

  • Role in argument: Prevents conflating attitudes with behavior under coercion.

  • Analytical note: In IW, observable “support” is often endogenous to security and fear.

• Collaboration

  • Definition: “behavior … that favors one actor to the detriment of the other.” (p. 92)

  • Role in argument: Central behavioral variable linking civilians to control and violence.

  • Analytical note: Treat as a spectrum and track it through observable acts (information, resources, enlistment).

• Defection

  • Definition: “active collaboration with the rival actor.” (p. 93)

  • Role in argument: The key behavior violence seeks to deter.

  • Analytical note: Defection risk is endogenous to control; measure via switching patterns and intelligence indicators.

• Control

  • Definition: An actor’s capacity to set rules and deter defection across territory; varies across zones. (pp. 120–121)

  • Role in argument: Primary causal driver shaping information and violence patterns.

  • Analytical note: Control is gradient; measure through enforcement, presence, predictability of retaliation, and governance penetration.

• Selective violence

  • Definition: “Violence is selective when there is an intention to ascertain individual guilt.” (p. 142)

  • Role in argument: Core outcome explained by information/control conditions.

  • Analytical note: Requires identification and therefore tends to co-occur with denunciation and dominant-but-incomplete control.

• Indiscriminate violence

  • Definition: Ideal-typical opposite of selective violence; used when individual guilt cannot be ascertained. (p. 142; p. 171)

  • Role in argument: Alternative violence strategy under information scarcity.

  • Analytical note: Often counterproductive; track association-based targeting and its effect on civilian alignment. (p. 147; p. 168)

• Denunciation

  • Definition: Civilian-provided information identifying individuals; central to selective violence production. (p. 13; p. 173)

  • Role in argument: The micro-mechanism enabling identification, punishment, and control consolidation.

  • Analytical note: High moral hazard: can be “dark face of social capital” and reflect private motives. (p. 353)

---

Key Arguments & Evidence

• Argument 1:

  • Violence in civil war has an autonomous logic and cannot be explained away as mere by-product of civil war onset; it must be theorized directly.

  • Evidence/examples:

    • Conceptual argument and scope conditions distinguishing civil war violence from genocide and unilateral repression. (p. 21; p. 29–31)

  • So what:

    • Strategists must analyze violence as a deliberate instrument tied to governance competition, not simply as “breakdown” or “hate.”

• Argument 2:

  • Control and information jointly determine where selective violence concentrates: it is most intense under dominant-but-incomplete control where denunciation is feasible.

  • Evidence/examples:

    • Theoretical prediction linking control to selective violence. (p. 13; p. 204)

    • Microcomparative confirmation in Greece: selective homicidal violence higher under largely-but-not-totally controlled areas. (p. 328)

  • So what:

    • Operational planning should prioritize consolidating control and building trustworthy information channels; “hot spots” may be diagnostic of partial control.

• Argument 3:

  • Indiscriminate violence is often a response to identification constraints but tends to be counterproductive because it shifts civilians toward whoever provides security.

  • Evidence/examples:

    • Proposition that indiscriminate violence is counterproductive. (p. 147)

    • Civilian behavior: collaboration flows to the actor guaranteeing security. (p. 168)

  • So what:

    • Collective punishment strategies often lose the intelligence and legitimacy war; measure backlash through collaboration/defection indicators.

• Argument 4:

  • Civil war violence is locally mediated and often “intimate”: denunciation and local rivalries privatize politics and bias information flows.

  • Evidence/examples:

    • “Privatizes politics” argument; denunciation as dark face of social capital. (p. 351; p. 353)

  • So what:

    • Partnered IW and targeting must anticipate malicious denunciation and design institutions to reduce false positives and prevent exploitation.

⚖️ Assumptions & Critical Tensions

• Assumptions the author needs:

  • At least one actor intends to govern (not exterminate/deport), making civilian choice consequential. (p. 31)

  • Civilians have agency and can provide information (denunciation) under certain control conditions. (p. 13; p. 204)

  • Control can be meaningfully conceptualized and approximated empirically at local level. (pp. 120–121; Appendix B)

• Tensions / tradeoffs / contradictions:

  • Endogeneity: violence affects control and control affects violence; disentangling causal direction is analytically and empirically hard (the book addresses via design, but tension remains). (pp. 245–246; Appendix B)

  • Ethnic vs non-ethnic wars: visibility of identity may change identification dynamics; the model addresses ethnic wars but may require different assumptions about information. (p. 181)

  • Information quality: denunciation enables selectivity but also introduces systematic bias (malicious denunciation), complicating the “information solves identification” story. (p. 353)

• What would change the author’s mind? (mark clearly as inference)

  • Inference: robust evidence across cases that selective violence is highest in fully contested zones (zone 3) rather than dominant-but-incomplete zones (2/4), controlling for actor capability and institutional settings, would undercut the core prediction. (p. 204; p. 328)

Critique Points

• Strongest critique:

  • Measuring “control” is notoriously difficult and potentially endogenous; even with careful coding, comparability across conflicts and time remains a challenge.

• Weakest critique:

  • The model may underweight emotion, ideology, and identity-based hatred as independent drivers—though the book explicitly tries to bound scope conditions and address ethnic wars.

• Method/data critique (if applicable):

  • Heavy reliance on micro-level historical reconstruction (interviews + archives) can embed recall bias and survivorship bias, despite triangulation. (Appendix A)

• Missing variable / alternative explanation:

  • External support/intervention and technology (ISR, surveillance, coercive capacity) may shift the identification problem and alter the predicted control–violence relationship in contemporary conflicts (inference; not a central variable in the book’s model).

---

Policy & Strategy Takeaways

• Implications for the US + partners:

  • Treat control as a governance-and-information condition: invest in protection and local institutional presence that changes civilian incentives. (p. 168; pp. 120–121)

  • Build intelligence systems that anticipate denunciation’s moral hazard; integrate verification, due process, and safeguards. (p. 353; p. 359)

  • Avoid indiscriminate violence and collective punishment approaches that push civilians toward the adversary. (p. 147; p. 168)

• Practical “do this / avoid that” bullets:

  • Do: map control gradients (not binary maps) and track them over time. (pp. 120–121; p. 216)

  • Do: design information incentives to reward truthful reporting and reduce retaliation risk. (p. 204; p. 359)

  • Avoid: equating expressed “support” with collaboration; behavior changes with security conditions. (p. 90; p. 168)

  • Avoid: “shock” tactics that rely on indiscriminate coercion to force compliance. (p. 147)

• Risks / second-order effects:

  • Denunciation systems can become instruments of private vendettas and state capture at the local level. (p. 351; p. 353)

  • Overreliance on coercion can degrade long-term control by destroying information channels and civilian trust. (p. 147; p. 168)

• What to measure (MOE/MOP ideas) and over what timeline:

  • Control indicators (zone-like gradients): patrol persistence, dispute resolution, taxation/extortion patterns, and predictability of retaliation (months). (pp. 120–121; p. 216)

  • Civilian security indicators: local homicide/civilian casualty patterns by actor and type (weekly/monthly). (p. 328)

  • Information health: denunciation volume, false-positive rates, verification lag, retaliation incidents (weekly/monthly). (p. 353; p. 359)

  • Collaboration/defection proxies: tips, recruitment flows, compliance with rules, switching patterns (monthly/quarterly). (pp. 92–93; p. 168)

⚔️ Cross‑Text Synthesis (SAASS 644)

Connect to other course texts (only what fits):

• Where this aligns:

  • Biddle (Nonstate Warfare): Both emphasize that irregular outcomes are shaped by constraints, information, and incentives more than “irrationality”; Kalyvas adds a control-based microfoundation for civilian behavior and violence patterns.

  • Mao / Galula / Algeria texts: The intuition “control the population” is sharpened: Kalyvas specifies where partial control produces the fiercest selective violence and why contested zones may not be the most violent.

• Where this contradicts:

  • Simplistic “barbarism” / pure-hatred framings (common in some narratives of civil war): Kalyvas insists violence is often strategic interaction and locally mediated, not simply norm-collapse. (p. 389; p. 353)

• What it adds that others miss:

  • A precise linkage between control gradients and forms of violence (selective vs indiscriminate) plus the role of denunciation as the actionable micro-mechanism. (p. 204; p. 353)

• 2–4 “bridge” insights tying at least TWO other readings together:

  • Biddle + Kalyvas + Mao: Biddle explains why many actors fight irregularly; Kalyvas explains how irregular control regimes generate information problems and selective violence; Mao’s “political work” can be reinterpreted as building the control/information conditions that make selective coercion feasible.

  • Peterson/Galula (Algeria) + Kalyvas: Classic COIN stresses separating insurgents from population; Kalyvas predicts the most intense selective violence in dominant-but-incomplete control zones—precisely where “separation” efforts and intelligence campaigns are most contested and morally hazardous.

  • Simpson + Kalyvas: Simpson emphasizes war’s political/narrative character; Kalyvas shows politics can become “intimate” and privatized through denunciation and local rivalries—suggesting narrative strategies must account for micro-incentives and local social fractures, not just messaging themes. (p. 351; p. 353)

---

❓ Open Questions for Seminar

• If selective violence is “jointly produced,” how should strategists design institutions that harvest information while minimizing malicious denunciation and false positives? (p. 13; p. 353)

• How do modern surveillance technologies (ISR, biometrics, digital exhaust) alter the identification problem and therefore the predicted control–violence relationship? (inference beyond text’s primary scope)

• When do conflicts shift from “govern” logics to “exterminate” logics, and what indicators show that a case is moving from civil war violence toward genocide/mass deportation? (p. 29–31)

• Does Kalyvas’s “oasis of peace” logic in contested zones generalize to conflicts with high mobility/urban density where zones are less stable? (p. 204)

• How should we think about partner forces whose coercion is effective tactically but indiscriminate strategically—what levers change behavior without collapsing the coalition? (p. 147; p. 168)

• In ethnic civil wars where identity is partially observable, when does denunciation matter most versus when does violence become more “categorical”? (p. 181)

• What does “success” mean in a control-driven theory—reduced violence, increased compliance, or durable governance—and how do these trade off? (p. 245; p. 147)

✍️ Notable Quotes & Thoughts

• “Selective violence is the result of transactions between political actors and civilians.” (p. 13)

• “The causes of violence in civil war cannot be subsumed under the causes of civil war.” (p. 21)

• “Violence is not merely state terror multiplied by two.” (p. 31)

• “Violence is selective when there is an intention to ascertain individual guilt.” (p. 142)

• “Proposition 2: Indiscriminate violence is counterproductive in civil war.” (p. 147)

• “Selective violence is most likely to take place in zones that are safely controlled by one actor.” (p. 173)

• “Civil war also works the other way around: it privatizes politics.” (p. 351)

• “The practice of denunciation may be thought of as the dark face of social capital.” (p. 353)

• “We cannot get at the micromechanisms of violence from the analysis of macrovariables.” (p. 386)

• “This book has marshaled theory and evidence to demonstrate that there is logic to violence in civil war.” (p. 389)

Exam Drills / Take‑Home Hooks

• Likely prompt 1: Explain Kalyvas’s theory of selective violence and where violence is most likely.

  • Outline:

    1. Define selective violence and why information is the binding constraint. (p. 142; p. 173)

    2. Explain control zones and the non-linear prediction (dominant-but-incomplete control → peak selective violence; contested zones can be quieter). (pp. 120–121; p. 204)

    3. Strategic implications: protection + intelligence institutions; manage denunciation moral hazard. (p. 168; p. 353)

• Likely prompt 2: Why is indiscriminate violence often counterproductive, and why do actors still use it?

  • Outline:

    1. Define indiscriminate violence as response to identification limits. (p. 171; p. 142)

    2. Show counterproductive effect via civilian security alignment. (p. 147; p. 168)

    3. Explain persistence: constraints, misperception, limited capabilities, short-term incentives. (pp. 146–172)

• Likely prompt 3: How does Kalyvas connect macro cleavages to micro violence, and why does disaggregation matter?

  • Outline:

    1. Macro “master cleavage” frames conflict but does not determine local outcomes. (p. 365)

    2. Micro mechanisms: denunciation, local rivalries, agency, institutional settings. (p. 13; p. 353)

    3. Method claim: macrovariables can’t reveal micromechanisms; disaggregation is necessary. (p. 386)

• If I had to write a 1500‑word response in 4–5 hours, my thesis would be:

  Civil war violence is best explained as a strategic, locally mediated process in which territorial control shapes information and incentives, producing patterned selective and indiscriminate violence through civilian collaboration and denunciation.

  • 3 supporting points + 1 anticipated counterargument:

    • Support 1: Violence has analytic autonomy; it cannot be derived from civil war onset causes. (p. 21)

    • Support 2: Control zones predict where selective violence concentrates; empirical evidence supports dominant-but-incomplete control pattern. (p. 204; p. 328)

    • Support 3: Indiscriminate violence is often self-defeating because civilians align with providers of security; coercion without protection loses control. (p. 147; p. 168)

    • Counterargument (anticipated): Some civil wars are driven by ideology/identity and extermination aims where identification is easy and denunciation is less central; response is that Kalyvas’s scope conditions and typology distinguish these dynamics and bound the theory’s domain. (p. 29–31; p. 181)