1. Mitton 2008, 202.

2. This definition draws directly from Parkinson and Zaks 2018.

3. Kay Cohen 2013; Saul 2013; Gilligan, Samii, and Mvukiyehe 2012; Schwartz 2019; Stedman 1997; Fortna 2003; Collier, Hoeffler, and Söderbom 2008; Daly 2016.

4. Martin, Piccolino, and Speight 2022; Themnér 2015; Daly 2016.

5. Allison 2010; Marshall and Ishiyama 2016; Matanock 2017a.

6. Zaks 2024.

7. On dismantling combat units, see de Zeeuw 2008, 12, and Söderberg Kovacs and Hatz 2016. On reformulating decision-making structures, see Manning 2004, 2007; de Zeeuw 2008; and Sindre 2016b. On restructuring finances, see Manning 2007, 256. On recruiting candidates, see Manning 2004, 59. On building party structures, see de Zeeuw 2008, 13. On adapting to the postconflict environment, see Manning 2007, 255.

8. Close and Prevost 2007; de Zeeuw 2008.

9. According to Michael T. Hannan and John Freeman, “the process of dismantling one structure and building another” threatens internal legitimacy and destabilizes organizations (1984, 149, 159). Even scholars who are more optimistic about the prospects of organizational change would express skepticism when the new requirements depart sharply from existing skill sets (Haveman 1992, 1993a).

10. Hannan and Freeman 1984, 156. See also Amburgey, Kelly, and Barnett 1993.

11. Hannan and Freeman 1984.

12. For example, Gyda M. Sindre notes that “former rebel parties have to adjust to the procedural rules of multi-party democracy” (2016a, 196). Similarly, Manning asks, “What factors influence [rebel] parties’ decisions to adapt to or subvert the rules?” (2007, 55). See also Acosta 2014.

13. Staniland 2014, 38.

14. As James G. March argues, organizations “rarely change in a way that fulfills the intention of a particular group of actors” (1981, 563). And when they try, we often observe “weak relations between intentions and outcomes” (Hannan and Freeman 1989, 22–23).

15. de Zeeuw 2008; Söderberg Kovacs and Hatz 2016; Acosta 2014; Matanock 2018.

16. Allison 2006; Manning and Smith 2019.

17. Allison 2006; Klapdor 2009.

18. Daly 2022.

19. On prewar party experience, see Wade 2008, 46; de Zeeuw 2008; and Manning and Smith 2016, 984. On centralization and cohesion in the organization, see Söderberg Kovacs 2007; Rudolph 2008; Klapdor 2009; Ishiyama and Batta 2011a; and Holland 2016. On convertible capabilities, see Dresden 2017. On wartime territorial control, see Allison 2010; Ishiyama and Widmeier 2013. On resource endowments, see Manning 2002, 108; Allison 2006; Wade 2008; Klapdor 2009.

20. Nelson and Winter 1982, 134.

21. As Scott Somers argues, while managers need control over their employees, “highly bureaucratized, command-and-control style structures impede creativity and adaptive behavior” (2009, 13).

22. Levitsky 2003, 28.

23. Allison 2010; Ishiyama and Widmeier 2013; Dresden 2017.

24. Hannan and Freeman 1989, 6.

25. Sarah Zukerman Daly (2016) observes similar problems with resources vis-à-vis conflict recurrence—namely, a group’s latent mobilizational capacity mediates the effectiveness with which resources are leveraged should the group wish to reassemble.

26. For other examples of this approach, see Staniland 2014 and Parkinson 2016, 2022.

27. Sobre los Problemas Mas Agudos en la Estructura de las MPL Según Apreciaciones de María del Organismo del PPL (FPL, approx. 1984e).

28. Wantchekon and Neeman 2002.

29. Hannan and Freeman 1977, 957.

30. Panebianco 1988, 3; Parkinson and Zaks 2018.

31. For a review of the organizational trajectory in civil war research, see Parkinson and Zaks 2018.

32. On patterns of violence and restraint, see Hoover Green 2018. On wartime resilience, see Parkinson 2013, 2022. On fragmentation, see Bakke, Cunningham, and Seymour 2012. On conflict resurgence, see Daly 2012, 2016. On rebel-to-party transformation, see Lyons 2016; Ishiyama and Widmeier 2019.

33. The first three questions draw from Barnett and Carroll 1995.

34. Haveman 1993b.

35. On governing civilians, see Huang 2016b; Stewart 2021. On providing social services, see Heger and Jung 2016.

36. Parkinson 2013; Staniland 2014.

37. This distinction is first drawn explicitly in Zaks 2024.

38. Huang 2016b.

39. These propositions draw on Anna Grzymala-Busse’s exploration of how the legacies of communist successor parties affect their post-democratization party structure (2002, 21).

40. Of course, some transnational groups may have local offshoots, whose aim is confined to a specific country. In these cases, the local offshoot would fall within the bounds of the theory.

41. This criterion may seem overly simplistic, yet at least two datasets with rebel-to-party variables omit this condition. As a result, groups like al-Qaeda Somalia are coded as rebel-to-party failures even though party formation was illegal in Somalia for the years the dataset covers (Matanock 2016).

42. Drawing on the Uppsala Conflict Data Program’s definition of armed conflict, a civil war is “a contested incompatibility that concerns government and/or territory where the use of armed force between two parties, of which at least one is the government of a state, results in at least 25 battle-related deaths in one calendar year” (Gleditsch et al. 2002; Petterson 2023).

43. de Zeeuw 2008; Klapdor 2009; Söderberg Kovacs and Hatz 2016; Holland 2016.

44. Kenneth A. Bollen observes a similar problem with measures embedding “stability” in “democracy” (1980, 375).

45. Then, of course, there are groups on the margins, like the Sandinistas in Nicaragua. Here, the Sandinista National Liberation Front (FSLN) only briefly transitioned into a one-party state following a military victory. Soon thereafter, however, they opened up the political system and successfully competed in local and national elections.

46. This approach takes inspiration from what Amelia Hoover Green calls “institutional biographies,” which focused on education and disciplinary institutions within the movement (2018, 19).

47. Parkinson 2021, 59. Parkinson develops this concept to refer to the disjuncture between how political scientists conceive of and ask about ideology and how people in militant organizations conceive of it outside of the simplified left-right spectrum.

48. Regarding who joins rebellions and why, see Berman et al. 2011; Humphreys and Weinstein 2008. Regarding forcibly recruiting participants, see Eck 2014. For how and when rebels deploy different forms of violence, see Weinstein 2007; Gutiérrez-Sanín and Wood 2014. On when and why rebels engage in behaviors beyond violence, see Mampilly 2011; Arjona 2016.

49. Huntington 1968; Panebianco 1988; Katz and Mair 1993; Grzymala-Busse 2002; Levitsky 2003; Weinstein 2007; Daly 2012.

1. Parkinson and Zaks 2018.

2. de Zeeuw 2008, 13.

3. Gonzalez 2018, 4.

4. The PCS was banned from elections after the insurrection in 1932, and since political organizing was banned, the group was forced into a clandestine existence for nearly five decades (Montgomery 1995).

5. Bracamonte and Spencer 1995; Montgomery 1995, 52; Viterna 2013, 250.

6. Veredas de la Audacia: Historia del FMLN (Cienfuegos 1986). Emphasis added.

7. For accounts of the group’s emergence and the political context surrounding it, see Pearce 1986; Bracamonte and Spencer 1995; Montgomery 1995; Byrne 1996; Binford 2004; Wood 2000, 2003; Hoover Green 2018; and Sprenkels 2018.

8. Bracamonte and Spencer 1995; de Zeeuw 2008, 9.

9. Hoover Green 2018, 59.

10. McClintock 1998, 90.

11. Montgomery 1995, 105; Binford 1997, 57.

12. Binford 1997, 57. In a document titled Aproximación a la Historia del PRTC, Nidia Díaz (2006), one of the founders of the PRTC, describes the organization in these terms. This was a turn of phrase she used both during and after the war to describe the FMLN leadership (Viterna 2013, 250).

13. Aldrich 1995.

14. Panebianco 1988, 6.

15. This definition draws on and modifies the Uppsala Conflict Data Program definition of “opposition organizations” (2021).

16. This condition does not exclude rebels who hold extralegal elections in captured territory. While rebel-held elections may indicate groups’ willingness to become legal parties, these organizations are distinct from groups operating as militants and parties concurrently.

17. Sartori 1976, 54.

18. Shugart 1992, 122; de Zeeuw 2008, 5; Acosta 2014, 671; Manning and Smith 2016, 973; Söderberg Kovacs and Hatz 2016, 7; and Matanock 2016, Appendix, 2, all use this definition.

19. John Ishiyama and Anna Batta argue that “durable peace settlements require the active involvement and cooperative engagement of these political groups” (2011a, 369). Marshall and Ishiyama argue that “the inclusion and participation of former rebel parties in national government” is the mechanism by which transition fosters stability (2016, 1009). Manning and Smith ask “under what conditions are rebel groups successfully incorporated into democratic politics?” (2016, 972). Matanock notes the importance of “bringing armed actors into normal politics” (2018, 656).

20. On rebel groups, see Petersen 2001; Weinstein 2007; Pearlman 2011; Parkinson 2013; Balcells and Justino 2014; Staniland 2014. On parties, see Huntington 1968; Panebianco 1988; Katz and Mair 1993; Mair 1994; Grzymala-Busse 2002; Levitsky 2003.

21. Mampilly 2011, 6.

22. Díaz 1982–1984.

23. Wendy Pearlman, for example, argues that fragmented movements “lack leadership, institutions, and collective purpose,” whereas cohesive movements “enjoy the organizational power to mobilize mass participation, enforce strategic discipline, and contain disruptive dissent” (2011, 2). See also Staniland 2012; Daly 2016, 16; and Balcells 2017, 64.

24. Ishiyama and Batta 2011a.

25. Hannan and Freeman 1984, 155; Haveman 1993b; Somers 2009.

26. Mintzberg 1979, 12.

27. Söderberg Kovacs 2007; Klapdor 2009; Ishiyama and Batta 2011a; Pearlman 2011; Bakke, Cunningham, and Seymour 2012; Staniland 2014.

28. Jeremy M. Weinstein, for example, asks “under what conditions rebel organizations will take on tasks of governance” (2007, 38). Additionally, Reyko Huang’s theory of civilian mobilization argues that when rebels “depend heavily on civilian inputs” during the war, they “lay down extensive roots to draw on material and moral support,” which results in popular mobilization and, eventually, democracy (2016b, 29). Even more directly, Cunningham, Huang, and Sawyer begin their article with the following: “Rebel actors engage in a number of behaviors beyond violent conflict, including social service provision, diplomacy, and establishing local governance” (2021, 81).

29. Zachariah Cherian Mampilly’s definition of governance system is one of few that acknowledge the duality of organization and behavior. He defines rebel governance as “not only the structures that provide certain public goods but also the practices of rule insurgents adopt” (2011, 4).

30. Barnett and Carroll 1995.

31. de Zeeuw 2008.

32. Mintzberg 1979, 2.

33. A few notable exceptions are Petersen 2001, 2010; and Parkinson 2013. Roger D. Petersen observes that “a great deal of variation exists in the types of roles that individuals come to play during sustained rebellion” (2001, 8). Abdulkader H. Sinno, too, notes that “specialized subgroups within organizations” are one of the core components of structure (2010, 35). One of the core contributions of Parkinson’s work is the compound insight that the variety of noncombat roles and relations present in rebel groups are not only more elaborate than what is typically acknowledged, but also that the subdivisions dedicated to logistical and support tasks often become the linchpin of organizational survival in times of crisis. Finally, as Ralph Sprenkels notes, “In-depth research demonstrates that insurgencies tend to covertly tie together multiple groups and individuals with an amalgam of roles, interweaving political, social, military, and economic aspects” (2018, 31).

34. Parkinson and Zaks 2018, 273.

35. I use subdivisions and wings interchangeably throughout the manuscript to refer to collective divisions within organizations. Though some do, I do not use institution as a synonym for subdivisions. I reserve the concept of institution to refer to the rules that structure relations and behavior within organizations.

36. Reyko Huang draws a crucial parallel distinction in her discussion of wartime extortion, arguing that “sporadic acts of extortion” should not be expected if rebels rely primarily on local populations for support (2016b, 32). Instead, we should expect them to build institutions “to regularize the process.”

37. Parkinson and Zaks 2018, 274.

38. With few exceptions, the nature of these connections is neither specified nor questioned. The most notable counterexamples are Paul Staniland’s work, which critically examines the role of prewar social networks in the formation of militant organizations (2014), and Sarah Parkinson’s work, which finds that organizational resilience and other crucial outcomes are attributable to kinship relations, rumor networks, and other quotidian connections between individuals that tend to be overlooked in analyses of conflict (2013, 2016).

39. Wood (2003), Mampilly (2011), and Parkinson (2013) are some of the earlier examples of scholarship that go beyond such conventional portrayals. More recently, the rebel governance literature implicitly takes up questions of external relations. See Arjona 2016; Huang 2016a; Mampilly and Stewart 2021. On the production and type of violence between rebels and civilians, see Kalyvas 2006; Humphreys and Weinstein 2006; Weinstein 2007. On the lack thereof, see Hoover Green 2018.

40. See Parkinson 2013, 2016; Christia 2014; Staniland 2014; Huang 2016b; and Dresden 2017.

41. Horne and Orr 1998, 30; Sutcliffe and Vogus 2003.

42. Sewell 1992; Haveman 1993b; Sutcliffe and Vogus 2003.

43. Vogus and Sutcliffe 2007, 342; Somers 2009, 13; Minkoff 1999.

44. Katz and Mair 1993, 594–595.

45. Katz and Mair 1993, 596–597.

46. Scott and Davis 2007, 35.

47. Stinchcombe 1965.

48. Allison 2010; Ishiyama and Widmeier 2013; Dresden 2017.

49. To be clear, I make a distinction between the origins of rebels’ organizational structures and the antecedents to civil war. I’m not asking why rebel groups emerge, I’m asking why they are structured the way they are—given that they have incentives to emerge. Of course, some of these might be related; and some scholars treat them as the same question. James Fearon and David Laitin (2003), for example, attribute both the mobilization of the insurgency and its wartime structure to the environmental and political conditions of the state in which it emerges.

50. Notable exceptions are Gates 2002; Weinstein 2007; Parkinson 2013; and Staniland 2014.

51. On ideology, see Selznick 1952; Huntington 1968; Eck 2007; Thaler 2012; Kalyvas 2015; Mampilly 2011; Stewart 2021. On state context, see Skocpol 1979. Theda Skocpol argues that state breakdowns are necessary conditions for peasant revolution. Revolutionaries and scholars alike argue that geological features are critical to sustaining rebellion (see, for example, Guevara [1961] 2006 and Gates 2002). Similarly, Fearon and Laitin argue that state characteristics shape “the political and military technology of the insurgency” (2003, 81). Indeed, FMLN leaders noted that some of their initial difficulty in securing funding was because no one believed a flat country like El Salvador could sustain a revolution. According to Fermán Cienfuegos, early negotiations always turned to the question, “Where are the mountains?” (Cienfuegos 1986). On resource endowments, see Weinstein 2007; Humphreys and Weinstein 2008; Lidow 2016. On prewar social networks, see Kaufman 1985. Herbert Kaufman notes the broad importance of social networks in shaping organizational structure (94). Doug McAdam, Sidney Tarrow, and Charles Tilly (2001) argue that mobilization is highly contingent on whether “would-be” activists are able to “socially appropriate” existing networks in service of organizational goals. See also Daly (2012, 474) and Staniland (2014, 23–33).

52. On patterns of violence, see Gutiérrez-Sanín and Wood 2014; Maynard 2019. On recruitment practices, see Gutiérrez-Sanín and Wood 2014; Wood and Thomas 2017. On quotidian interactions and socialization, see Moro 2017; Parkinson 2021. On claim making, see Gutiérrez-Sanín and Wood 2014. On adopted institutions, see Schubiger and Zelina 2017.

53. Moro 2017, 945. The corollary, of course, is that early decisions have powerful and lasting impacts. See Pierson 2000, 253.

54. Staniland 2014, 23.

55. Consider, for example Newtonian mechanics. Newton’s laws (and the variables constituting them) work well much of the time, but they break at both the subatomic level and as matter approaches the speed of light because the variables operate differently at these different levels.

56. Though used for more specific purposes, Suykens (2015, 139) advances a similar argument in his discussion of “governance ideology,” which he defines as “how rebel groups understand the nature of their rebellion, their relations to the territory they operate in and to the civilians present in it.” As such, Suykens is one of few who explicitly integrate local context with broad principles to explain how rebels organize.

57. Parkinson 2021, 52.

58. In the early days of the rebellion, the nascent movement “rallied a number of radical intellectuals with a Green Book background” (Richards and Vincent 2008, 83).

59. Interview quoted in Peters 2011, 93.

60. Abdullah and Muana 1998.

61. Staniland 2014.

62. Weinstein 2007, 37.

63. Mampilly 2011; Parkinson 2013.

1. This figure draws on “Veredas de la Audacia: Historia del FMLN” (Cienfuegos 1986) as well as Dunkerley 1982, 302–3; Montgomery 1995, 102; McClintock 1998; and Sprenkels 2018, 55.

2. On mass organizations, see Wood 2003, chap. 6. On OP-Ms, see Bracamonte and Spencer 1995; McClintock 1998, 52.

3. In an early PRTC document, En Cerros de San Pedro, the author (Ernesto Flores) describes an event in which the RN “conquered” some PRTC members and stole cattle and other supplies from their trade routes (Flores 1983).

4. The first move in this vein came in 1856 with a law mandating that two-thirds of all communal lands (both tierras comunales and ejidos) be planted with coffee; failure to comply would result in state seizure of the land (Montgomery 1995, 30). Then, in the early 1880s, the state outlawed tierras comunales and ejidos while simultaneously ramping up enforcement of vagrancy laws (Wood 2000, 28).

5. Wade 2018, 397; Paige 1997, 107.

6. Unsurprisingly, estimates vary widely. Wood (2001, 31) and others contend that the number is about seventeen thousand; Montgomery (1995, 37) puts the number closer to thirty thousand. The ten thousand to forty thousand range is cited in a number of journalistic and encyclopedic accounts, but without references.

7. Wood 2000, 31.

8. Wood 2000, 32. Moreover, between 1950 and 1960, the membership of the National Guard swelled at an estimated rate of 3,500 recruits per year, which amounts to 1 recruit for every 714 citizens (Walter and Williams 1993, 47). To put this number in context, the recruitment rate of the US armed forces amounts to 1 recruit for every 4,050 citizens.

9. Montgomery 1995, 39; Wood 2000, 32.

10. As of the mid-1960s, the PCS was estimated to have approximately two hundred members (Benjamin and Kautsky 1968, 122).

11. Staniland 2014, 21.

12. Hammond 1999, 70.

13. Montgomery 1995, 87.

14. Hammond 1999, 74.

15. Montgomery 1995, 89. Campesino is a self-referential term referring to El Salvador’s rural poor. They often work in the agricultural sector, but as others have noted, the term itself does not translate well (Wood 2003, 5). As such, I use campesino throughout to refer to the people, and eventually, the collectives they built.

16. Hammond 1999, 74.

17. In 1975, 48.9 percent of men and 57.2 percent of women in rural areas were reported to be illiterate (Statistical Abstract of Latin America 1987).

18. Hammond 1999, 72.

19. According to Montgomery, more than fifteen thousand campesinos were trained as community leaders through church programs (1995, 89).

20. Bracamonte and Spencer 1995, 2.

21. Carpio’s objective in coining this term was to highlight the balance between political and military work in what he called “a new type of Marxist-Leninist party.” He wanted the balance between political and military engagement to be as explicit as possible (Harnecker 1983, 87).

22. Harnecker 1983, 84. Marcial was Carpio’s nom de guerre. In the interview transcript, all leaders are referred to by their pseudonyms.

23. Bracamonte and Spencer 1995, 2.

24. Interview with the PCS leader Schafik Hándal (Harnecker 1983, 73).

25. Bracamonte and Spencer 1995; Montgomery 1995, 52.

26. Montgomery 1995, 110. The General Command—originally called the Unified Revolutionary Directorate—was the five-man decision-making body of the FMLN, which comprised the leaders of each OP-M.

27. Selznick 1952, 15; Huntington 1968, 336.

28. This fact gives rise to an additional and highly consequential problem. For groups that espouse other, less organizationally focused ideologies, mapping their belief systems onto structures will thus be a considerably more difficult task. Thus, even if it performs well here, ideology likely has heterogeneous utility for explaining organizational outcomes.

29. Gates 2002, 113; Fearon and Laitin 2003, 81; Arjona 2014; Wickham-Crowley 2015.

30. However, making more fine-grained predictions about suborganizational variation requires that we identify and define structural variation within state borders.

31. Horowitz, Perkoski, and Potter 2018, 141.

32. Weinstein 2007; Humphreys and Weinstein 2008.

33. Bracamonte and Spencer 1995, 16.

34. According to Bracamonte and Spencer (1995, 177), Farid Hándal (the brother of the PCS leader Shafik Hándal) went on “a world tour” securing funds from “the United States, the Soviet Union, Czechoslovakia, Bulgaria, East Germany, Algeria, Libya, Ethiopia, and Vietnam.”

35. Sprenkels 2018, 366.

36. Hoover Green 2016, 26.

37. Chalatenango and Morazán are departments—roughly comparable to states—in El Salvador. Both are quite poor, both are dominated by agricultural labor, and both were primary targets of liberation theology in prior decades.

38. The organizational salience of the Christian base communities was as apparent to the rebels as it was to the scholars who studied them. Countless internal memoranda speak to the central role of religious sectors. In an FPL document detailing the growth and functioning of the movement, the author writes, “In the first place, Catholic groups are the sector to which we have the greatest access,” and they go on to detail how (FPL 1984a).

39. Staniland 2014.

40. Staniland 2014, 25.

41. See Zaks (2017) on how relationships among “competing” explanations do not demand mutually exclusive relationships.

42. Once again, specifying the outcome in organizational terms sheds light on where existing explanations fall short. Namely, while each purports to account for organizational structure, they rely on different or incomplete notions of what structure entails. As a result, we encounter explanations of traits like cohesion without specification of what exactly is cohering.

43. Dunkerley 1982, 102.

44. Carpio 1982a.

45. Interview with Joaquín Villalobos (Comandancia General del FMLN 1985a). Emphasis added.

46. The Central Command (or Co-Cen) was a seven-person circle at the top of the FPL hierarchy (Carpio 1982b, 1).

47. FPL Central Command 1975.

48. McClintock 1998, 253.

49. Harnecker 1983, 352.

50. Roberto Roca, quoted in McClintock 1998, 57.

51. Bracamonte and Spencer 1995, 15. Carpio elaborates: “If we really want to promote revolution, it is necessary to build a Marxist-Leninist Party that looks after the interest of the working class, the poor peasantry, and the rest of the people. . . . Without the true Communist Party, any war will lead to partial results” (1981, 8).

52. Comisión Nacional de FPL 1981, 3.

53. On material provisions, see Tarrow 1998; Wood 2003; Mampilly 2011; Staniland 2014. On recruitment, see Kalyvas 2006. On discretion, see Wood 2003; Lewis 2020.

54. Carpio 1982b.

55. FPL, approx. 1972a; FPL 1972b; Carpio 1983.

56. Comisión Nacional de FPL 1981. Emphasis added.

57. Cienfuegos 1986.

58. Arnado 1972.

59. Arnado 1972; see also ERP, n.d.; and Villalobos 1989.

60. Montgomery 1995, 104. Foquismo theory hails primarily from the Cuban revolutionary Che Guevara, who argued that armed guerrilla cadres were necessary to foment a socialist revolution (Guevara [1961] 2006). Guevara further describes the approach as the logical application of Marxism-Leninism to the Latin American condition, which makes foquismo itself a prime example of organizational ideology.

61. Bracamonte and Spencer 1995, 14.

62. Bracamonte and Spencer 1995, 14.

63. Byrne 1996, 37.

64. Díaz 2006.

65. Dunkerley 1982, 102.

66. Díaz 2006. In his treatise On War, Carl Von Clausewitz famously quipped, “War is the continuation of politics by other means” (87).

67. In the document Sobre el Plan de Contrainsurgencia, more of their strategic principles focus on the importance of political and diplomatic work than the military work (PRTC early 1980s, 9).

68. Staniland 2014.

69. As Yvon Grenier argues, universities in El Salvador “offered distinct mobilizational resources” due to their “autonomy, the organizational configuration of the university community, and their specific disposition to act as society’s vanguard” (1999, 103).

70. Due likely to the PRTC’s modest size (five hundred to eight hundred members, compared to the three thousand to four thousand boasted by the FPL and ERP) and the fact that it did not hold anywhere near the same amount of territory throughout the war, information on the PRTC’s social base beyond university students is sparse. Thus, I focus more extensively on the FPL and ERP for this discussion. However, because the FPL and ERP differ so considerably in their organizational ideologies, the countless similarities across their social bases is the more salient information here.

71. One of Jenny Pearce’s interviewees called Chalatenango “the most punished zone” (1986, 178). Leigh Binford (1998) similarly documents horrific patterns of government repression in Morazán.

72. Pearce 1986; Montgomery 1995; Hammond 1999; Binford 2004.

73. Pearce 1986, 128, 202; Binford 1998; Consalvi 2010, xxviii.

74. Pearce likewise notes the ideological tension between Marxism and the role of the church in the revolution (1986, 128).

75. FPL Central Command 1975. Emphasis added.

76. Viterna 2013, 250.

77. Viterna 2013, 250. These estimates come from a declassified document from the Salvadoran military archives. They are likely biased downward, for a few reasons. Most notably, they are specifically estimates of the groups’ armed factions. It is thus unlikely that they captured the full extent of participation, since a great deal of revolutionary work took place outside the combat realm.

78. Hoover Green 2018, 26.

79. Flores 1983.

80. Flores 1983.

81. Staniland 2014.

82. Staniland 2014, 25.

83. Weinstein 2007.

84. Prior to 1977, the ERP still exhibited divisiveness in its leadership, leading to the formation of the RN in 1974 and the PRTC in 1976. See Hoover Green (2018) on the effectiveness of disciplinary and control institutions across the FMLN.

85. Comisión Nacional de FPL 1981, 3.

86. Carpio 1982b. Yes, the FPL (as well as the PRTC) had a wing of its organization dedicated to organization.

87. FPL 1972b.

88. Carpio 1982b.

89. FPL, n.d.b.

90. Carpio 1982b.

91. The plans to create an organizational arm to accomplish this task are articulated in Materiales Basicos de las FPL (FPL 1972b), and both Montgomery (1995) and Wood (2000) confirm that the BPR functioned as intended.

92. Comandancia General del FMLN 1985a, 38. Emphasis added.

93. Carpio 1982b, 4. See also FPL 1984d.

94. Pearce 1986, 242, 294.

95. Pearce 1986; Hammond 1999.

96. ERP, n.d.

97. Dunkerley 1982, 95.

98. Vigil 1991, 141.

99. The ERP commander Joaquín Villalobos made frequent mention of this in early documents and interviews: “Communication,” he argues, “. . . will become a fundamental part of our plans” (Harnecker 1983, 107).

100. Vigil 1991; Consalvi 2010, xxxiv.

101. From Madness to Hope: The 12-Year War in El Salvador: Report of the Commission on the Truth for El Salvador (1993).

102. See note 17 in this chapter.

103. Recall that Horowitz, Perkoski, and Potter (2018) touted the importance of state-centric factors for determining organizations, arguing that government repression would lead rebel organizations to be more tactically diverse

104. Vigil 1991, 3. Emphasis added.

105. Vigil 1991; Ching 2010, xxxviii.

106. I elaborate on their expansion in chapter 3.

107. Montgomery 1995, 151.

108. Interview with Joaquín Villalobos (Harnecker 1983, 113). Of course, propagandists always argue that their content is the real story, but in this case Villalobos was not stretching. In fact, it was well documented at the time that “even [the] enemies tuned in because its news reports were more reliable than the pro-government sources” (Consalvi 2010, xxxix).

109. Álvarez and Orero 2014, 678. Francisco Jovel would eventually leave the ERP to join the PRTC.

110. According to Leigh Binford, “Although the guerrillas functioned as a quasi-state and dominated through force when necessary, attainment of their strategic objectives was tied to the development of a modicum of hegemony over civilians, who served as a recruitment poll and supplied food, labor, information, and other forms of assistance crucial to the struggle” (1998, 4).

111. Interview with Joaquín Villalobos (Harnecker 1983, 98).

112. Resistencia Nacional 1975.

113. ERP, n.d. See also Montgomery 1995, 120; Binford 1998, 7.

114. Viterna 2013, 249.

115. Manning 2008a, 117.

116. Montgomery 1995, 107.

117. Montgomery 1995, 107.

118. ERP, n.d.; Montgomery 1995, 121.

119. Viterna 2013, 250.

120. This argument finds reasonable support among organizational sociologists (Blau 1970; Pugh 1973; Haveman 1993b; Minkoff 1999).

121. Kasfir 2015, 24.

122. Recall the complexity and diversity of the organizational diagram that motivated figure 1.1.

123. Díaz 2006.

124. Díaz 2006.

125. In the document Sobre el Plan de Contrainsurgencia, more of the group’s “strategic principles” highlight the importance of political and diplomatic work than focus on militancy (PRTC early 1980s, 9).

126. This page is one of countless organizational depictions from the wartime diary of Nidia Díaz (1982–1984). Throughout her diary, the noncombat divisions take priority.

127. PRTC, n.d.a.

128. El Secretario Central de PRTC, n.d.

129. Pearce 1986, 133.

130. Resistencia Nacional 1975.

131. Resistencia Nacional 1975.

132. Montgomery 1995, 105; Ching 2010, xxx.

133. Montgomery 1995, 105; Bracamonte and Spencer 1995, 14.

134. Montgomery 1995, 107.

135. Montgomery 1995, 107.

136. Hándal interview (Harnecker 1983, 73).

137. Montgomery 1995, 105.

138. Hándal interview (Harnecker 1983, 73).

139. Bracamonte and Spencer 1995, 3.

140. On strength of disciplinary institutions, see Weinstein 2007. On ties to local populations, see Staniland 2014. On differences in tactical approaches, see Horowitz, Perkoski, and Potter 2018.

141. Recall that the ERP took years to build a mass front and once it did, its members viewed it more as a logistical support system than an autonomous source of revolutionary politics.

142. To be sure, the “most important” feature will change depending on the analysis.

143. Montgomery 1995, 107.

144. Weinstein 2007.

145. Bracamonte and Spencer 1995; Montgomery 1995; Viterna 2013.

146. Hoover Green 2018.

147. Wood 2003, 193.

148. Wood 2003, 193.

149. Moreover, Staniland’s theory helps us get traction on an ostensible paradox within the FMLN. On the one hand, we observe severe fragmentation at the top of the organization. Yet, “fragmented” poorly describes the group, since we also observe strong cohesion and effective command-and-control structures within each OP-M. By differentiating among horizontal and vertical ties, Staniland’s typology exemplifies the analytic purchase of specifying the different levels at which organizational traits apply to a given group.

150. Staniland 2014, 25.

151. Montgomery 1995, 109.

152. Bracamonte and Spencer 1995; Montgomery 1995, 52; Viterna 2013, 250.

153. Bracamonte and Spencer 1995, 16.

154. Bracamonte and Spencer 1995, 19.

155. Bracamonte and Spencer 1995, 18.

156. Although ¡Radio Venceremos! was transmitting, the signal at the time was too weak to reach enough of El Salvador for listeners to rely on that broadcast alone.

157. Montgomery 1995, 113.

158. Montgomery 1995, 113; Montgomery 1995, 113; Bracamonte and Spencer 1995, 19, 20.

1. Weinstein 2002, 2; Kalyvas 2006; Humphreys and Weinstein 2008; Kalyvas and Balcells 2010; Horowitz, Perkoski, and Potter 2018.

2. On conflict strategies, see Kalyvas and Balcells 2010. On patterns of violence, see Gutiérrez-Sanín and Wood 2017.

3. de Zeeuw 2008, 8.

4. Manning 2004, 59.

5. Manning 2007, 57; de Zeeuw 2008, 13; Close and Prevost 2007, 7.

6. John Ishiyama (2019), for example, demonstrates that the majority of ex-rebel parties do not change their names significantly when they transform into parties, and that when they do, the name change has no discernible effect on their electoral performance. Furthermore, in a parallel exploration of party organizational change, Anna Grzymala-Busse (2002, 73) contends that one of “the first two tasks” communist successor parties faced at the end of the Cold War was to “break with the past decisively—by changing the party’s name, program, symbols, and public representatives.”

7. Manning and Smith 2016.

8. See, for example, Hannan and Freeman 1977, 1984, 1989; Haveman 1992, 1993b; Amburgey, Kelly, and Barnett 1993; Barnett and Carroll 1995; and Greve 1998. Of course, they focus on different traits and harbor varying levels of optimism (or pessimism) for the prospects of change in the face of upheaval. However, none denies the importance of looking at the structures and traits the organization has in place in the run-up to transformation.

9. Grzymała-Busse 2002, Levitsky 2003, and Parkinson 2013, 2022, are notable exceptions. Though they focus on different political organizations—communist successor parties, labor movements, and militant organizations, respectively—each masterfully demonstrates that organizational flexibility is key to adaptation and survival.

10. Hannan and Freeman 1984, 1989. Grzymała-Busse (2002, 23) echoes these challenges in her discussion of the organizational legacies of communist successor parties.

11. Haveman 1992.

12. Hannan and Freeman 1977, 1984; Haveman 1992, 1993b.

13. Barnett and Carroll 1995.

14. On rebel organizations, see Weinstein 2007; Pearlman 2011; Daly 2012; Parkinson 2013; Staniland 2014. On party organizations, see Janda 1983; Schonfeld 1983; Panebianco 1988; Levitsky 2003; Van Biezen 2005.

15. Panebianco 1988, 6–8.

16. Panebianco 1988; Levitsky 2003.

17. Key 1942; Huntington 1968; Janda 1983; Schonfeld 1983.

18. The quote is from Key 1942, 315. Similarly, Katz and Mair argue that aggregating traits to describe the party organization as a whole is problematic because “parties often have several separate bureaucracies” (1993, 595). Panebianco, too, characterizes parties as “mixed organizations” with multiple “bureaucratic elements, which are at times contradictory and at times harmonious” (1988, 201). Finally, Lars Svåsand (1994) identifies a similar set of structural elements in his exploration of Norwegian parties.

19. Key 1942, 377.

20. Katz and Mair 1993, 594–595.

21. Indeed, as Panebianco (1988, 20) notes, not only do subdivisions within the party have different goals, but those goals are sometimes incongruous with one another.

22. Katz and Mair 1993, 596–597.

23. Katz and Mair 1993, 596–597.

24. V. O. Key (1942, 316) makes a similar distinction in his disaggregation of party organizations.

25. Katz and Mair 1993, 598–599

26. Indeed, Katz and Mair (1993) disaggregate party subdivisions in the first place to demonstrate that parties are not in decline; rather, parties are de-emphasizing one subdivision in favor of dedicating more resources to another.

27. Katz and Mair (1993, 594) explicitly argue a full organizational analysis of a given party would require scholars to further disaggregate each “face” and specify the type and arrangement of its various components.

28. Manning 2004, 59; de Zeeuw 2008.

29. On governance, see Mampilly 2011; Arjona, Kasfir, and Mampilly 2015; Arjona 2016; Stewart 2021. On social services, see Flanigan 2008; Mampilly 2011; Heger and Jung 2016; Stewart 2018. On natural resources, see Humphreys 2005; Lujala 2010; Ohmura 2018. On political education, see Pearce 1986; Hoover Green 2016.

30. Arjona, Kasfir, and Mampilly (2015, 3) define rebel governance in terms of behavior, or actions.

31. Security and policing also fall under “governance,” but I do not expect structures that redirect the coercive apparatus of war to the local level to have major implications for rebel-to-party transformation, as these tasks do not require the type of political or administrative skills associated with other governance activities. For excellent accounts on the origins, varieties, and implications of rebel governance, see Mampilly 2011; Arjona, Kasfir, and Mampilly 2015; Arjona 2016; and Stewart 2018.

32. Silverstein 2007; Flanigan 2008.

33. For perhaps the most extensive and impressive account of the PPLs, see Pearce (1986, 241–273).

34. Mampilly 2011, 62.

35. Katz and Mair 1993, 596.

36. Flanigan 2008, 505; Mampilly 2011, 63.

37. Monsma 1996; Chernov Hwang and Schulze 2018.

38. This assumption is most succinctly captured in the title of Humphreys and Weinstein’s 2008 piece, “Who fights? The Determinants of Participation in Civil War,” in which fighting and participation are treated as synonymous. For a more extensive account of how participation in rebellion varies, see Wood 2003.

39. Wood 2003; Parkinson 2013.

40. Amelia Hoover Green provides what is likely the most comprehensive account of internal political education in the FMLN. She notes that nearly all ex-combatants received formal political education and they can speak at length about the reasoning underlying the war (2018, 90–91).

41. Hannan and Freeman 1984.

42. Ramzipoor 2015.

43. Manning 2007, 2008a; de Zeeuw 2008.

44. Manning argues, “Electoral politics require a different set of skills than those demanded during wartime” (2004, 59).

45. Close and Prevost 2007, 7.

46. Katz and Mair 1993, 597.

47. Mampilly 2011; Arjona 2016; Stewart 2021.

48. Minkoff 1999, 1671.

49. As Kaufman (1985, 46) argues, organizational change “entails reasoned assessments” of both the environment and the organization.

50. Tan 2020, 80.

51. Zachariah Cherian Mampilly briefly remarks on the importance of organizational “feedback mechanisms” in explaining rebel governance outcomes (2011, 17). For his theory, the feedback mechanisms relate specifically to fostering civilian participation, yet we would do well to think more broadly about how this trait manifests and varies from one organization to the next.

52. Padgett and Powell 2012. For an example of this dynamic during war, see Parkinson’s analysis of organizational plasticity in the Palestine Liberation Organization (2013, 2022).

53. Minkoff 1999, 1671

54. Close and Prevost 2007, 9.

55. Parkinson 2013, 422. Emphasis added.

56. Grzymala-Busse 2002.

57. Grzymala-Busse 2002, 29.

58. Reno, n.d., 265.

59. More broadly, the lack of coordination among FMLN factions—though rooted ideologically—was made worse by its limited capacity for communication. The movement had not yet established the radio communications system it would come to rely on later in the war. As a result, even the cadres that did participate failed to coordinate the simultaneous attacks that the original plan called for (Bracamonte and Spencer 1995, 19).

60. Moreover, knowing that they lacked the military strength to overthrow the regime themselves, the ERP leaders believed success was contingent on mass mobilization in the form of a national strike. That facet of the plan, however, was built on an assumption that revolutionary fervor was stirring throughout all sectors of the population. Dominated largely by the ERP’s reasoning that conspicuous shows of force would ignite these grievances, sending people flooding into the streets in solidarity with the insurrection, the FMLN “felt no need to prepare for this event” (Bracamonte and Spencer 1995, 19).

61. Bracamonte and Spencer 1995, 19.

62. This argument comes directly from Cynthia McClintock’s analysis (1998, 90). Bracamonte and Spencer 1995, Montgomery 1995, and Manning 2008a corroborate this point, arguing that the FMLN’s internecine rivalries were a greater obstacle to success than the Salvadoran state.

63. Bracamonte and Spencer 1995, 116.

64. Bracamonte and Spencer 1995, 116.

65. Bracamonte and Spencer 1995, 116.

66. Bracamonte and Spencer 1995, 21.

67. The ERP leader Villalobos considered the insurrection less a failure than some of the others, which comports with the ERP’s guiding principles. He writes, “The fact that most analyses [of the insurrection] started from a political vantage, they reached the conclusion that since we did not fully seize power, January 10th could be considered a failure. This situation did not allow others to evaluate the event as a momentous change in the development in the war and the quality of our military forces” (Villalobos 1981, 88).

68. Bonasso and Gómez Leyva 1992, 35.

69. Villalobos 1986, 8.

70. Dweck and Sorich 1999.

71. In the FPL document Cuadernos de Formación No. 3 the authors recount shifting their approach to organizational development in response to a series of misguided approaches (“un serie de enfoques equivocados”) (Carpio 1982b, 2). A collection of early ERP documents is rife with explicit references to how they shifted organizational tactics on the basis of autocrítico (ERP 1977). The founding document of the RN is itself called Balance Auto-Crítico (Resistencia Nacional 1975). Throughout her diaries, Nidia Díaz routinely scheduled “reunions de autocrítico”—meetings for self-critique (Díaz 1982–1984). Finally, in multiple interviews, Schafik Hándal of the PCS routinely highlighted the importance of autocrítico for the FMLN to move forward (Harnecker 1983, 1988).

72. For example, see ¡Revolución o Muerte: El Pueblo Armado Vencera! (Carpio 1983, 21) and Con el Tiempo a Nuestro Favor (Comandancia General del FMLN 1985a, 35; among many others).

73. On paper, the FMLN General Command officially operated according to the Leninist principle of democratic centralism, which required consensus from all five leaders (Cienfuegos 1986).

74. Bracamonte and Spencer 1995, 8.

75. Bracamonte and Spencer 1995, 116.

76. Harnecker 1983, 95.

77. Ching 2010, xxxii.

78. Harnecker 1983, 95.

79. Dunkerley 1982.

80. Binford 1997, 61; Ching 2010, xxxii. Villalobos describes the training conditions in the aftermath of the retreat: “The offensive put the army on the defensive . . . which gave us a few months of peace and allowed us to create seven strategic fronts. [Quickly], however, the enemy’s offensives ramped up and became a combat-preparation school. We were forced to solve the problem of tactical and combat training in real time, by facing the enemy” (Harnecker 1983, 95).

81. In an interview with Marta Harnecker, Villalobos wavers at the term “liberated zone” because the ERP “was not fully self-sufficient” (Harnecker 1988).

82. Binford 1997, 61.

83. According to these arguments territorial control provides the opportunity to build proto-party structures. For example, John Ishiyama and Michael Widmeier argue, “Rebels that control territory are more likely to establish order and set up ‘bush bureaucracies’” (2013, 533). Similarly, Paul Staniland argues that “vertical ties [to local populations] make it possible for leaders to quickly establish institutions for local control . . . [and] . . . share their ideology with the people at the local level, facilitating political education” (2014, 27).

84. Huang 2016b, 39.

85. Ventura 1990, 8.

86. As Leigh Binford recounts, “Within the areas in which it exercised nominal day-to-day control . . . attainment of [the ERP’s] strategic objectives was tied to . . . civilians, who served as a recruitment pool and supplied food, labor, information, and other forms of crucial assistance to the struggle” (1998, 4).

87. Harnecker 1983, 98.

88. Binford 1997, 62.

89. Harnecker 1983.

90. Binford 1997, 61.

91. Montgomery 1995, 173; Wood 2003, 157.

92. Montgomery 1995, 173; Wood 2003, 157.

93. Montgomery 1995, 120.

94. Bracamonte and Spencer 1995, 15.

95. Ching 2010, xxxii.

96. Comisión Nacional de FPL 1981, 3; Carpio 1982a, 1982b.

97. Pearce 1986, 242.

98. Comandancia 1981.

99. Comandancia 1981, 1.

100. Pearce 1986, 242.

101. Linea de Masas—FPL was written in March of 1981, just a few weeks after the January 10 offensive.

102. Comisión Nacional de FPL 1981, 3.

103. For a detailed description and analysis of PPLs, see chapter 8 in Pearce 1986.

104. Pearce 1986, 242. In Linea de Masas, the first objective for rural territories is “to participate actively in the construction, development, and consolidation of ‘farabundista’ popular power” (Comisión Nacional de FPL 1981, 5). Jenny Pearce confirms the FPL’s guiding hand in her analysis of the structures, describing them as playing an “orienting role” and shaping the political priorities of the PPLs throughout their growth (1986, 248).

105. FPL, n.d., 1; FPL 1984c, 1.

106. Pearce goes on to describe the structure in detail:

“Each PPL was elected by and responsible to 400–500 people . . . a ‘locality.’ The entire area of Chalatenango under guerrilla control was divided into three sub-zones[;] . . . sub-zone one had seven PPLs and a sub-regional government. The highest power in each locality was the popular assembly, a general meeting of the whole population. Between popular assemblies, power rested with the junta or council of the PPL, made up of a president, a vice president, and secretaries for production and popular economy, for social affairs (health and education), for legal affairs, for political education, and for defense. . . . Every eight days there would be a meeting of the general secretaries of the mass organizations in the locality who represented the consejo of the PPL and whose task was to mobilize the population behind the decision of the PPL” (1986, 243–244).

107. Pearce 1986, 250.

108. Pearce 1986; Binford 1997.

109. Pearce 1986, 242. Emphasis added.

110. Cienfuegos 1986, 84. Emphasis added.

111. PPL member, quoted in Pearce 1986, 244.

112. While official estimates place illiteracy rates at 43 percent at the time, Hammond and Portillo argue that the true figure is closer to 70 percent (1991, 91–92).

113. In an analysis of “popular education” in El Salvador, Hammond and Portillo argue that education is “viewed as a political act and an obligation of both the teacher and the learner” (1991, 91).

114. As Jenny Pearce observed, “Learning to read and write was one of [the population’s] expectations of the revolution, a way to ‘break our silence’” (1986, 261). In another interview, a former teacher describes an “anxiety to learn” among adults in the community (Pearce 1986, 261).

115. Pearce 1986, 261.

116. ANDES (the National Association of Salvadoran Educators) was a progressive teachers’ union that came under increasingly heavy scrutiny and oppression in the early 1980s (Hammond and Portillo 1991, 93). While many members fled the country and taught literacy programs in the refugee camps, many others stayed behind and joined sects of the FMLN. For example, the FPL’s second-in-command, Mélinda Anaya Montes (a.k.a. Ana María), held a PhD from the University of El Salvador and was one of the founders of ANDES (McClintock 1998, 256). The FPL’s use of its international wing in service of its educational subdivision also sheds important and interesting new light on the nature of international patronage. While we normally think about rebel state sponsors as supplying money, guns, and training to insurgents, the FPL implored its donors to send notebooks and pencils for its students (Pearce 1986, 267).

117. Interview excerpt from Pearce 1986, 266–267.

118. Freire 1970.

119. FPL 1984b, 1, 2, 5.

120. FMLN 1980s.

121. CONEPI 1984, 2–4.

122. CONEPI 1984, 4.

123. Wood 2003, 28.

124. The US aid to the Salvadoran military was not exactly an enforcement of human rights. In response to Senator Edward Kennedy’s request for an impact statement on US military assistance to El Salvador, Joseph E. Kelley of the National Security and Affairs division of the General Accounting Office prepared a report that includes the following details. US advisers to El Salvador told the government to “do what you need to stop the commies, just don’t get caught” (Kelley 1991, 138). More horrific, a former US intelligence officer suggested that the death squads needed to leave less visual evidence of human rights violations. Specifically, they should “stop dumping bodies on the side of the road” because “they have an ocean and they ought to use it” (Kelley 1991, 139).

125. Cynthia McClintock (1998, 118) reports that the number of civilians abducted by government and paramilitary forces dropped from 535 to 53 over the course of a year.

126. Montgomery 1995, 185.

127. Viterna 2013, 63.

128. Viterna 2013, 63.

129. Schwartz 2019.

130. Bracamonte and Spencer 1995, 23.

131. José Angel Moroni Bracamonte and David E. Spencer argue that the FMLN’s early embrace of tactical diversity imbued the organization with “a flexibility rarely seen in previous guerrilla groups . . . allowing them to pick and choose between what worked and what didn’t” (1995, 8).

132. Byrne 1996, 98. According to Bracamonte and Spencer (1995, 23), the General Command’s new directive pushed the organization to “adopt a new strategy loosely based on that of the FPL, with elements of the RN and ERP lines of thought incorporated.”

133. FMLN 1987, 23.

134. Byrne 1996, 132; McClintock 1998, 83.

135. Montgomery 1995, 198.

136. Montgomery 1995, 198.

137. According to Sarah Miles and Bob Ostertag, organizing politically “became the centre of the rebel plan during this period” (1991, 222).

138. The General Command explicitly acknowledged the importance of these early structures when it wrote, “The experience that has been accumulated in some areas of the Eastern Front is very positive and [moving forward] we must generalize that approach” (Comandancia General del FMLN 1986, 9–10).

139. Wood 2000, 49.

140. FMLN 1987, 26.

141. Comandancia General del FMLN 1986, 9–10. See also Binford 1997, 13.

142. For a rich analysis of the FMLN’s internal political education structures, see Hoover Green 2018. For example, documents from the FPL (dated February 1984) explicitly address the need for evaluation, training, augmentation, and specialization of its cadres (FPL 1984d).

143. Díaz 1982–1984.

144. Secretariado Central del PRTC, n.d., 4.

145. The General Command representative of the RN Fermán Cienfuegos addresses the issue directly: “We have a repopulation process in which the refugees return to their place of residence, raising a new organizational challenge that we must face and resolve” (Cienfuegos 1986, 86–7; emphasis added).

146. FPL 1983; FPL 1984f.

147. FMLN 1987, 26; Documentos Propuestas y Projectos a Reunión CG (Junio / 84), Nidia Díaz Papers, box 1, folder 1, Hoover Institution Library & Archives.

148. In one internal document, for example, the authors write, “Recruitment campaigns should be launched for personnel who can use typewriters, for professors, radio operators, draughtsmen (people trained in technical drawing), theater artists, and auto and bank mechanics. . . . Yet, it is politically and strategically fundamental that they all come from the working class” (PRTC, n.d.c, 1).

149. Wood 2000, 49; 2003, 166–167.

150. Binford 1998.

151. Wood 2003, 167.

152. Binford 1998, 30. The disparate evolution of the ERP and other OP-Ms further highlights the importance of distinguishing between organizational outcomes and behavioral ones.

153. Binford 1998, 29.

154. Orero 2016.

155. Regarding the PRTC’s health provisions, Ilja Luciak notes that four cadres were “sent to Mexico to receive training in basic dental care” (2001, 104). This expedition is helpful context for understanding a common joke among the PRTC leadership: “The best way for Salvadoran security forces to identify FMLN collaborators in the countryside would have been to look for the peasants with the cleanest teeth” (Luciak 2001, 104).

156. Comandancia General del FMLN 1985a, 36.

157. Viterna 2013, 118. Emphasis added.

158. Huang 2016b, 39.

This chapter draws extensively on Zaks 2025.

1. Manning 2007; de Zeeuw 2008; Klapdor 2009; Söderberg Kovacs and Hatz 2016.

2. Across the rebel-to-party literature, Klapdor makes this distinction most explicit, arguing that “transformation is both an outcome and a process” (2009, 13).

3. The quote is from Klapdor 2009, 13. Jereon de Zeeuw specifies two attitudinal changes: the “democratization of decision making” and the “adaptation of strategies and goals” in service of their “new political tactics.” The two structural changes he proposes are the demilitarization of combat structures and the “development of a party organization” (2008, 13–15.). On internal elite negotiations, see Manning 2007.

4. Hannan and Freeman 1977, 1984, 1989; Cummings and Huse 1985; Greve 1998.

5. Scott and Davis 2007, 35.

6. The quotes, in order, are from de Zeeuw 2008, 13; Manning 2007, 57; and Close and Prevost 2007, 4.

7. Fligstein and Dauber 1989, 84–85.

8. Hannan, Pólos, and Carroll 2003.

9. Hannan and Freeman 1977, 957.

10. Wantchekon and Neeman 2002, 439. Emphasis added.

11. To be sure, characterizing the new environment as tumultuous should not raise too many eyebrows. Volatility among new democracies is well known and multidimensional. From a domestic vantage, scholars note the significantly increased risk of regime collapse and authoritarian reversals (Svolik 2008). From an international vantage, scholars also observe that new democracies are more prone to interstate conflict (Mansfield and Snyder 1995). Ralph Sprenkels echoes this critique, arguing that “what democratic transition theory tends to interpret as highly positive steps in the process—the demobilization of guerrilla troops, for example—raised for many of those directly involved complex and uncomfortable questions about the future of their movement” (2018, 2).

12. Kaufman 1985, 46.

13. Panebianco 1988.

14. As Rebecca M. Henderson and Kim B. Clark (1990, 118) argue, major organizational change is difficult because the “knowledge of the organization—particularly its communication channels, information filters, and problem-solving strategies” are deeply embedded. Consequently, organizations have a hard time imagining or correcting for their own obsolescence. On this point, also see Scott 2003, 287.

15. Levitsky 2001a.

16. Wildavsky 1988.

17. Specifically, adaptation is defined as a specific type of organizational change that “reduces the distance [or increases congruence] between the organization and its institutional environment” (Sarta, Durand, and Vergne 2021).

18. Sutcliffe and Vogus 2003.

19. Dresden 2017; Ishiyama and Widmeier 2019.

20. In his exploration of party adaptation, Steven Levitsky (2001a) takes a similar tack. Namely, he identifies that it was specifically weak institutionalization that enabled the Partido Judicialista to adapt—flouting conventional assumptions that strong institutionalization was a uniformly beneficial trait associated with party longevity.

21. Amburgey, Kelly, and Barnett 1993, 52. Emphasis added.

22. Minkoff 1999; Sewell 1992; Haveman 1993a; Sutcliffe and Vogus 2003.

23. Hannan and Freeman 1989, 8.

24. Hamel and Välikangas 2003, 12–3.

25. Haveman 1992, 48.

26. Manning and Smith 2016.

27. Hamel and Välikangas 2003, 5

28. Levitsky (2001b, 33–34) elaborates on this assumption and the implications for studying the evolution of party organizations.

29. Gerlach and Hine 1970, 34.

30. Jets are equipped with a tailhook to catch the high-tensile arresting wires, which absorb the shock of landing and allow the aircraft to come to a stop in the space allotted.

31. Rochlin, La Porte, and Roberts 1987, 78.

32. (ibid.) P. Rochlin, La Porte, and Roberts 1987, 83.

33. Rochlin, La Porte, and Roberts 1987, 83.

34. Sutcliffe and Vogus 2003, 10; Wruck and Jensen 1994.

35. Often, though not always, this moment occurs at the negotiating table. After the dance of luring other parties into negotiations, a now-common item on the agenda is whether a settlement will include electoral participation provisions (Matanock 2018). In other cases, the two sides do not engage in direct negotiations, but as the fighting subsides, militant groups decide to exploit existing openings in the political system and undertake party formation in the absence of formal provisions. I do not differentiate among these options. In addition, who makes the decision to exploit that opportunity varies by organization. In some cases, the decision is spearheaded by a single leader; in others (and more often) it’s made in consultation with a broader set of top leaders and commanders in the group.

36. Manning 2007; Close and Prevost 2007; de Zeeuw 2008; Ishiyama and Batta 2011b.

37. To be sure, routine assessments occur throughout the group’s life cycle. Usually, the needs are small and the corresponding changes are incremental: more weapons, different smuggling networks, more recruits for a given unit, or altered combat strategies, to name just a few. With demobilization and elections on the horizon, however, this task is infinitely more complex.

38. As Herbert Kaufman argues, organizational changes demand “reasoned assessments of the relevant conditions and of the changes considered most likely to achieve the desired ends” (1985, 46).

39. When this internal assessment is addressed, it is most commonly portrayed as a black box of internal negotiations (Manning 2007). To be sure, this depiction is not entirely inaccurate, just—by the author’s own admission—incomplete.

40. Grzymala-Busse 2002, 21.

41. Elsewhere throughout the organizational literature, these paths to organizational change are known as “exploration” and “exploitation” (March 1991, 71). They refer to an organization’s decision to expand into new markets either by building structures to explore new territory or by exploiting and refining existing structures to take advantage of a related opportunity.

42. Evolutionary-learning theories of party formation hold that without quickly adapting during the trial-and-error periods of early party formation, parties will face extinction (Kitschelt 1989, 44).

43. Hannan and Freeman 1984, 159.

44. Hannan and Freeman 1984, 156.

45. Many people who are hired in the party-building stage are either total outsiders to the organization or were part-time affiliates during the war (Manning 2008b), and these outsiders are often confronted with intense resentment on the part of long-standing members.

46. Dresden 2017; Ishiyama and Widmeier 2013; Manning and Smith 2016.

47. Stinchcombe 1965, 148.

48. As Stinchcombe goes on to note, creating new structures “relies heavily on social relations among strangers” where “trust is precarious” (1965, 149).

49. Padgett and Powell 2012; Parkinson 2013.

50. Sewell 1992.

51. Vines 1991; Manning 2002; Manning 2008b.

52. Internal documents as well as secondary sources reveal that the OP-Ms were more likely to work together in this period—carrying out joint operations, meeting at lower levels (below the General Command), and even meeting up socially (Díaz 1982–1984). Moreover, in a diary entry dated September 20, 1984, Díaz (1982–1984) pens a to-do list ahead of a joint-command meeting, including the need to “work on propaganda at the FMLN-level” (as opposed to the level of the PRTC, of which she was a commander). Cienfuegos 1986; FPL approx. 1984e.

53. For example, if rebel-to-party transformation unfolded by massive hiring initiatives, the ex-rebel party’s ability to recruit skilled personnel would likely be a function of two core factors. The first factor is the group’s repertoires of violence against local populations during war (which, if high, would disincentivize joining). The second factor is the structural composition of the labor market (availability of skilled workers, literacy rates, etc.).

54. Bracamonte and Spencer 1995, 35.

55. Wade 2008, 38.

56. UNSC 1992, 38.

57. Pearce 1986, 282.

58. Union leader interview, quoted in Pearce 1986, 242.

59. Union leader interview, quoted in Pearce 1986, 242.

60. Wood 2003, 167.

61. Comandancia General del FMLN 1985b.

62. Hoover Green 2018, 119.

63. Bracamonte and Spencer 1995, 116.

64. Bracamonte and Spencer explicitly attribute the failed strike to the injudicious repurposing of political cadres into militant roles (1995, 116).

65. Comandancia General del FMLN 1985a, 36, 76.

66. PCS 1993.

67. de Zeeuw 2008, 13–15.

68. Díaz 1982–1984.

69. FMLN-FDR, n.d.

70. FPL, n.d.c.

71. Partial document (PRTC, n.d.b, 4).

72. FPL 1984d.

73. In another example, Díaz wrote a formal document criticizing the organization for adopting a “circular” promotion strategy—that is, promoting people into command positions based on personal friendship networks, rather than expertise.

74. Political memorandum (FMLN-FDR 1991). Emphasis added.

75. FPL, n.d.a. Emphasis added.

76. PCS 1991. Emphasis added.

77. PRTC 1992. Emphasis added.

78. ERP, n.d.b. Emphasis added.

79. Villalobos 1993. Emphasis added.

80. ERP, n.d.b, 5. While other OP-Ms also sought to integrate grassroots leaders further into the party, the difference is that for the FPL and PRTC, those structures were not new—nor were their political relations with grassroots leaders.

81. Wade 2008, 40.

82. Luciak 2001, 98, 99.

83. Spence 1997, 18. Emphasis added.

84. Final three quotes in the paragraph are from Montgomery 1995, 254; Sprenkels 2018, 115; and PCS 1993.

85. FMLN 1992, 2; ERP, n.d.; see also Luciak 2001, 99.

86. FMLN 1992, 4.

87. Sprenkels 2018, 88.

88. The PCS also recounts similar dynamics in its account of the organizational direction ahead of party formation (PCS 1991).

89. FMLN-FDR, n.d. Emphasis added.

90. As Sprenkels notes, “The socioeconomic strategy laid out in the internal documents of the FPL essentially consisted of reconverting the civil-political front into a channel for development projects to benefit the historical OP-M constituents” (2018, 115).

91. Wood 2003, 178.

92. Land reform was one of the FMLN’s central priorities in the negotiating process. As Wood documents, “Insurgent cooperatives and the FMLN forced a transfer of approximately eight percent of the nation’s farmland” (2003, 182).

93. PRTC 1992.

94. Montgomery 1995, 249

95. Tommie Sue Montgomery further observes that the UN “had to flatter, cajole, and bully the commission to do the job it was created to do” (1995, 247–251).

96. Montgomery 1995, 232; May, Schneider, and Arana 2018.

97. May, Schneider, and Arana 2018, 59.

98. Viterna 2013.

99. Equipo Maiz, Cuaderno No. 4: Los Acuerdos de Paz, 1992.

100. Sprenkels 2018, 116.

101. Luciak 2001, 78.

102. Vigil 1991, 231.

103. Sprenkels 2018, 125.

104. Leonhard 1999, 30–32.

105. Allison 2006, 145.

106. Díaz 2006.

107. Sprenkels (2018, 135–136), for example, observes that “many former insurgents relied on interpersonal ties developed during the war to survive in peacetime. Their ability to do so [however] depended on the particular skills they had developed during their revolutionary career, some of which allowed for peacetime reconversion more easily than others. All OP-Ms formally dissolved as organic entities, but this did not mean that they also disappeared as relevant interpersonal networks.”

108. FPL 2009; FPL 2007.

109. Ramos, López, and Quinteros 2015, 12.

110. Ishiyama and Widmeier 2019, 130; Sprenkels 2018, 123.

111. Huang 2016b.

112. Luciak (2001, 98) and Montgomery (1995, 254) clearly document the ERP’s political lag in this period and how that lag shifted the balance of power.

113. Wood 2003, 174.

114. Sprenkels 2018, 122.

115. Sprenkels 2018, 122.

116. Spence 1997, 14; Close and Prevost 2007, 4.

1. Comandancia General del FMLN 1985a. Yes, a thousand eyes should really only trans late into five hundred votes, but good math doesn’t always make for pithy quotes.

2. Allison 2006; de Zeeuw 2008; Ishiyama and Batta 2011b; Sindre 2016b.

3. Söderberg Kovacs and Hatz 2016; Manning and Smith 2016.

4. Barnett and Carroll 1995.

5. Zaks 2024.

6. According to Ishiyama and Batta, transition into electoral competition may beget transformation of the party organization (2011b, 372). See also Söderström 2016, 215, and Sindre 2018, 23.

7. Of seventeen articles and books, only seven stick with one term; among those, many conflate indicators, which makes it difficult to even recognize the few works that employ consistent usage, such as de Zeeuw 2008; Sindre 2016a, 2016b; and Söderström 2016.

8. This section draws considerably on Zaks 2024.

9. Lyons 2005; Norris 2008; Hartzell and Hoodie 2015; Marshall and Ishiyama 2016; Matanock 2017b.

10. Hannan and Freeman 1989.

11. Olson 1998; Levitsky 2003; Levitsky, Loxton, and Van Dyck 2016.

12. For example, see Shugart 1992; Acosta 2014; Manning and Smith 2016; Söderberg Kovacs and Hatz 2016; Matanock 2016; Ishiyama 2019; Acosta and Rogers 2020. In contrast, de Zeeuw (2008) and a few of his contemporaries (e.g., Klapdor 2009) consistently use “transformation” to describe the outcome of their analyses, and their definitions focus largely on internal structural changes, which are outside the scope of this discussion.

13. According to Robert Adcock and David Collier (2001), conceptual frameworks must capture the “broad constellation of meanings” associated with the concept we wish to pin down.

14. On party registration, see, for example, Acosta 2014; Manning and Smith 2016; Acosta 2019. On ballot appearance, see, for example, Matanock 2016; Söderberg Kovacs and Hatz 2016.

15. Acosta 2014; Manning and Smith 2016; Söderberg Kovacs and Hatz 2016; Matanock 2016; Ishiyama 2019; and Acosta 2019 rely on a binary outcome variable. Relying solely on registration to demarcate transition creates a set of positive cases that does not differentiate between groups that registered but never even ran in elections and rebel successor parties with long-standing careers in political office. See Zaks 2024 for a deeper theoretical and methodological elaboration of this problem.

16. Shugart 1992.

17. Sartori 1970; Goertz 2006.

18. Acosta (2014, 2019) and Matanock (2016) take this approach in the analyses that accompany their respective datasets. Similarly, Manning and Smith (2016) and Söderberg Kovacs and Hatz (2016) add a rebel-to-party variable to different UCDP datasets.

19. The Legion of Doom is a now-defunct group of hackers that operated out of Boston in the 1980s. The Animal Liberation Front is a transnational organization dedicated to freeing animals from scientific and cosmetic testing facilities. Their recent communiques include “17 Chickens Liberated (Midlands, UK)” and “Shoplifters Liberate 51 Lobsters (USA)” (Animal Liberation Front, 2022). Neither have expressed the desire to take political control of a specified country.

20. On the distinction between disarmament and party formation, see Matanock 2016; Matanock and Staniland 2018. On disarmament and transformation, see de Zeeuw 2008; Klapdor 2009. On disarmament and transition, see Söderberg Kovacs and Hatz 2016.

21. Söderberg Kovacs and Hatz 2016.

22. Of course, some settlements mandate demobilization as a prerequisite to party registration, but that constraint is itself a separate variable.

23. This logic echoes E. E. Schattschneider’sadage that “democracy is not to be found in the parties, but between the parties” (1942, 60).

24. Kenneth A. Bollen (1980, 375) observes a similar problem with measures embedding “stability” in definitions of democracy.

25. ORDEN stands for Organización Democrática Nacionalista.

26. Shugart 1992, 122; de Zeeuw 2008, 5; Acosta 2014, 671; Manning and Smith 2016, 973; Söderberg Kovacs and Hatz 2016, 7; and Matanock 2016, appendix p. 2, all use this definition.

27. Sartori 1976, 54.

28. Sartori 1970, 62.

29. Sartori 1970; Adcock and Collier 2001; Goertz 2006.

30. Sartori 1970, 58. I use failed transitions because it is an intuitive shorthand for identifying the zeroes. However, success and failure are loaded terms, the definitions of which can change depending on the question, so I also use nontransition.

31. They may have reverted to violence, disintegrated, or transitioned into another sector.

32. Allison 2006.

33. Ishiyama and Widmeier (2019) make this argument directly. Huang (2016b, 39) similarly argues, “Where rebels tap into civilians as a significant war-making resource, the latter become politically mobilized.” Elements of the argument are also evident in Wickham-Crowley 1987.

34. Manning and Smith 2016, 2019. Additionally, while de Zeeuw (2008) primarily engages rebels’ postconflict transformation, he echoes this logic in his discussion of postconflict success.

35. Even in the most straightforward case—the mobilization of ex-combatants—demobilized members do not always come through as loyal voters, as Michael E. Allison posits. He argues that the size of insurgent groups predicts electoral performance because rebel successor parties can rely on ex-combatants to mobilize on election day (2006, 153). Yet, in Sierra Leone, the newly formed RUFP (successor party to the Revolutionary United Front) not only failed to win more than 2 percent of the popular vote, but more damningly, notoriously failed to mobilize even its former members (Mitton 2008, 202).

36. Shefter 1977; Lawson 1980; Kitschelt 2000; Levitsky 2003; Roberts 2016.

37. Of course, some might argue that the party, new as it is, has simply not had time to create linkages. But the previous two chapters demonstrate that this argument would vastly oversimplify the plurality of rebel-citizen relations forged during war.

38. Mampilly 2011, 6.

39. Duverger 1951.

40. Matanock 2017b.

41. The four datasets are Acosta’s 2014 and 2019 Revolutionary and Militant Organizations Dataset (REVMOD); Söderberg Kovacs and Hatz’s 2016 expansion of the Uppsala Conflict Data Program (UCDP) Peace Agreement Data (hereafter cited as SK&H); Manning and Smith’s 2016 expansion of the UCDP Conflict Termination data (hereafter cited as M&S), and Matanock’s Militant Group Electoral Participation (MGEP).

42. As Zaks 2024 further notes, the number of rebel-to-party transitions ranges from 33 to 91, and the number of rebel-to-party failures ranges from 60 to 660. Yet, the datasets all cover roughly the same years.

43. See the empirical discussion and appendix in Zaks 2024 for comprehensive documentation of the disparities and the problems that follow.

44. In contrast, both Manning and Smith (2016) and Söderberg Kovacs and Hatz (2016) expand datasets from the UCDP to conduct rebel-to-party analyses. Manning and Smith add a rebel-to-party variable to the conflict termination data, for which observations are conflict centric. Söderberg Kovacs and Hatz expand the peace agreement data, in which the observations are all peace agreements signed to conclude a conflict.

45. Conventionally, civil war is defined as “a contested incompatibility that concerns government and/or territory where the use of armed force between two parties, of which at least one is the government of a state, results in at least 25 battle-related deaths in one calendar year” (UCDP 2021).

46. The UCDP, for example, differentiates among conflicts over “political” versus “territorial” claims. However, in light of the stakes of rebel-to-party transition, I choose to include both—especially since we observe transition in both kinds of cases. For example, the FMLN made purely revolutionary claims on the Salvadoran government. In contrast, territorial autonomy was a core issue for both the Provisional Irish Republican Army (and its successor party, Sinn Fein) as well as groups that lay claim to occupied Palestinian territory, such as Fatah, Hamas, the Democratic Front for the Liberation of Palestine (DFLP) and the Popular Front for the Liberation of Palestine (PFLP).

47. Most of the groups omitted on this basis express no desire to participate in politics. They tend to view the system itself as the problem and expressly prefer to operate outside of legal bounds. For example, anarchist groups like Black Star (Mavro Astari) in Greece, whose actions are focused on the violent dismantling of the system itself. Black Star opposes what it sees as worldwide US imperialism and it most often fights US presence by throwing Molotov cocktails at diplomats’ cars.

48. See, for example, Acosta 2014. Of course, al-Qaeda’s local offshoots may be viable contenders. Relatedly, transnational groups are different from rebels with a base in a neighboring country but who are making specific political claims on a target government.

49. Matanock 2018, 657.

50. As I demonstrated in the 2022 article, MGEP includes al-Qaeda Somalia and codes it as “not having participated” in elections, yet party formation was not legal in Somalia during the scope of its inclusion in the dataset. Conversely, Söderberg Kovacs and Hatz (2016) rely on electoral provisions in negotiated settlements as a scope condition, when party formation is often legal outside these bounds.

51. Matanock 2016, 846; Acosta 2019. Though Acosta uses this criterion as well, he assembles the universe of cases (N = 2,322) then selects a random sample (N = 536), of which 457 are assigned values for rebel-to-party transition (Acosta 2019, appendix).

52. Matanock 2012, 87; 2016, 847.

53. King and Zeng 2001; Zaks 2024.

54. Söderström (2013), Marshall (2019), and Manning and Smith (2019) also speak to the theoretical relevance of distinguishing short-term and long-term contenders in the electoral system.

55. Analogously, Suazo (2013) argues that existing conceptions of “inclusive peace processes” miss key implications for security by omitting the duration of inclusiveness.

56. This number is empirically derived: parties tend to either fail after the third election or persist indefinitely.

57. LeBas 2011, 26.

58. LeBas 2011.

59. While health care structures are not directly translatable into party structures, health services rank highly among the critical needs of populations in war-torn areas. As such, I consider maintaining a health division as part of the local infrastructure that falls under the governance umbrella.

60. I use the name requirement as a proxy for consistency. The goal is to separate episodic round-ups of students for the sake of recruitment from lasting engagement with civilian communities.

61. To further ensure that one of the domains was not driving the results, I created a series of attenuated indices, which systematically omitted each of the components, and reran all of the tests. The differences between the results with the full index and the attenuated ones are insignificant.

62. See similar discussions from Huang 2016b, 61; Heger and Jung 2016; Stewart 2018; Albert 2022, 626.

63. Albert (2022, 626) notes this concern as well.

64. Huang 2016b, 61.

65. The divergent accounts between Pearce (1986) and Binford (1998) illustrate the potential problems that could arise from aggregating based on a geographical snapshot.

66. Matanock 2018; Söderberg Kovacs and Hatz 2016.

67. Haveman 1993b; Minkoff 1999.

68. Allison 2006.

69. Cunningham, Gleditsch, and Salehyan 2009.

70. Staniland 2014.

71. Duverger 1951; Lijphart 1994.

72. Acosta 2014.

73. While Acosta’s dataset is significantly larger, it also exhibits a high rate of missing data. Due presumably to the random selection, REVMOD is missing a substantial number of groups from transition-heavy states and regions: Afghanistan (missing six groups), Bosnia and Herzegovina (missing two groups), Burundi (missing all four groups), Cambodia (missing both major parties), Central African Republic (missing two groups), Chad (missing five groups), Congo (Brazzaville) (missing four groups), Democratic Republic of Congo (missing three groups), Djibouti (missing two groups), Ethiopia (missing three groups), Liberia (missing two groups), Sri Lanka (missing three groups) Uganda (missing two groups). Of these cases, twenty-nine successor parties have won seats in at least one postconflict election, twenty-two of those have won seats in three or more elections, and fourteen have achieved nominal participation.

74. Söderberg Kovacs and Hatz 2016, 5; see also de Zeeuw 2008, 19–20.

75. Nominal participation aligns with the coding used in Acosta 2014; Manning and Smith 2016; Söderberg Kovacs and Hatz 2016; Matanock 2016. That said, Söderberg Kovacs and Hatz (2016) also include disarmament as a necessary condition for success, which would exclude a few groups from consideration here.

76. An alternative explanation for the significance of size is actually a bias in the measure itself. It is possible that estimating the size of rebellions is easier for groups with higher levels of cohesion. As such, higher estimates on group size may actually be reflecting something important about the organization itself.

77. Matanock 2018.

78. This result further corroborates the propositions and conclusions laid out in Zaks 2024.

79. The municipalities with evidence of FMLN presence but without clear evidence of which OP-M was in charge are Zacatecoluca, Guatajiagua, Sensembra, San Gerardo, Berlin, and Nueva Granada.

80. Binford 1997.

81. Ishiyama and Widmeier 2013, 2020.

82. In the first analysis, the p-value on the results is 0.059; in the second analysis the p-value is 0.047.

1. Hoover Green 2016.

2. Söderberg Kovacs and Hatz 2016.

3. Weinstein 2007.

4. Staniland 2014; Stewart 2021.

5. Abdullah 1998, 209.

6. Hanlon 1990, 46.

7. Manning 2002, 51.

8. Morgan 1990, 610.

9. Vines 1991, 5.

10. Funada-Classen 2013, 382.

11. Cabrita 2001, 116; Manning 2002, 51.

12. Hanlon 1990, 51; Emerson 2014, loc. 485.

13. Sesay and Ukeje 2009, 27. According to Kwesi Aning and Angela McIntyre (2004, 27, 67), Sierra Leone had a well-functioning Westminster-style parliament, already established political parties, and the “promise of a budding democracy.”

14. Sesay and Ukeje 2009, 27.

15. Sesay and Ukeje 2009, 8–9, 30.

16. Sesay and Ukeje 2009, 30.

17. Aning and McIntyre 2004, 68.

18. Sesay and Ukeje 2009, 30.

19. Abdullah 1998, 208.

20. Rashid 1997; Savage and Rahall 2003, 49. The quote is from Abdullah 1998, 208. The terminology “lumpen youth” derives from lumpenproletariat.

21. Fauvet 1984, 115.

22. Cabrita 2001, 139–40.

23. Cabrita 2001, 146.

24. Fauvet and Gomes 1982, 114.

25. Cabrita 2001, 157.

26. Alexander 1995, 9.

27. Quoted in Alexander 1995, 9.

28. Cabrita 2001, 140.

29. Hanlon 1990, 228–9.

30. Abdullah 1998.

31. Aning 2010, 286.

32. Abdullah and Muana 1998.

33. Hazen 2013, 83.

34. Fauvet and Gomes 1982, 12.

35. Specifically, the CIO suggested that Renamo be split under two commands, one under each potential replacement: Afonso Dhlakama and Lucas M’lhanga (Fauvet and Gomes 1982).

36. Abdullah and Muana 1998, 177.

37. Denov 2010, 62. Given the critical role that experienced radio operators have played in both the FMLN and Renamo, having a communications specialist in a leadership role should have been advantageous for the centralization and cohesion of the RUF.

38. South Africa had two primary motivations for sponsoring Renamo and using the group to continue a war against the Machel regime in Mozambique. The first interest was security related. Frelimo tacitly supported the African National Congress, which made frequent incursions into South Africa from its bases in the south of Mozambique. South Africa’s second interest was economic. Any extent to which Mozambican ports or railways were out of commission was the extent to which South Africa retained a monopoly on regional port access. Understanding South Africa’s motivations behind its sponsorship helps to contextualize Renamo’s actions under its leadership.

39. Martin and Johnson 1986; Cabrita 2001; Emerson 2014.

40. Emerson 2014, loc. 1226, 1890.

41. Emerson 2014, loc. 1226.

42. Fauvet and Gomes 1982, 15; Fauvet 1984, 117.

43. For example, in 1980, Dhlakama appointed Renamo’s first “political commissar,” Henrique Sitoe. Sitoe’s role would be to help communicate the political goals of the organization to the foot soldiers and the local people inside Mozambique. This early venture into establishing a contingent of political commissars failed, as Sitoe admitted to having no knowledge of politics and defected shortly thereafter (Fauvet and Gomes 1982, 12).

44. Emerson 2014.

45. Minkoff 1999, 1677.

46. Manning 2008a, 61.

47. Manning 2008a, 61.

48. Carrie Manning corroborates this interpretation when she argues that “the establishment of formal political activities and structures within Mozambique began as a direct outgrowth of attempts to negotiate an end to the conflict beginning in 1984” (1998, 177).

49. Manning 2008a, 61.

50. Evo Fernandez, quoted in Hall 1990.

51. Chart in Manning 1998, 164. For additional evidence on this point, see Vines 1991. This source is especially compelling since Vines’s work focuses almost exclusively on the humanitarian atrocities committed at the hands of Renamo soldiers. His account seems otherwise committed to painting Renamo as little more than well-organized bandits, and indeed, the subtitle of the book is “Terrorism in Mozambique.” Additionally, see issue 10 of the newsletter Africa Confidential.

52. Alexander 1995, 29.

53. Manning 2002.

54. Sesay and Ukeje 2009; Denov 2010.

55. Richards and Vincent 2008, 83

56. Richards and Vincent 2008, 83.

57. As Paul Richards and James Vincent (2008, 92) note, he infused the otherwise hollow rhetoric of “emancipation” with grounded tenets of the Bunumbu curriculum, including “rural self-reliance and cooperative empowerment.”

58. Cohesion was always going to be difficult since the RUF included the Special Forces contingent of Liberian soldiers, who had fewer direct stakes in a political revolution in Sierra Leone.

59. Peters 2011, 127.

60. Richards and Vincent 2008, 88.

61. Peters 2011, 83.

62. Abdullah and Muana 1998, 178.

63. Peters 2011, 144.

64. Richards and Vincent 2008; Hazen 2013.

65. Interview with a former RUF combatant, quoted in Peters 2011, 93.

66. Abdullah 1998, 226.

67. Michael Ganawa, the public relations officer of the RUFP, and two former commanders, Edward Kamara and David Vandi, quoted in Richards and Vincent 2008, 93.

68. As Jennifer M. Hazen notes, the SLA’s military offensive successfully ejected the RUF from mining areas and cornered the group along and over the Liberian border (2013, 91).

69. Abdullah and Muana 1998, 183.

70. Abdullah and Muana 1998, 183.

71. Hazen 2013, 92.

72. Abdullah and Muana 1998, 183.

73. Harnecker 1983, 84.

74. Abdullah and Muana 1998, 191.

75. Peters 2011, 85.

76. One ex-RUF commander estimates that across the whole organization (which numbered around twenty thousand at its peak), approximately forty people were in charge of ideological training (Peters 2011, 130).

77. While many rank-and-file members were involved in illicit mining, the RUF’s official policy was to stay out of this sector until Sankoh shifted his stance.

78. Hazen 2013, 79–80.

79. This situation is in some ways comparable to the patronage structures under which Renamo emerged. While Renamo had its own agenda, we consistently observe the organization being forced to bend to the will of its Rhodesian sponsors and, later, to that of its South African sponsors.

80. Reno 2004; Hazen 2013.

81. Abdullah 1998, 227–228.

82. Abdullah 1998, 229.

83. Two of the men, Fayia Musa and Ibrahim Deen Jalloh, were held in RUF captivity for over two years before being released in November of 1999; the fate and whereabouts of the other two remain unknown (Hirsch 2000, 56).

84. Richards and Vincent 2008, 90.

85. Hazen 2013, 84.

86. Peters 2011, 151.

87. Former RUF commander, quoted in Peters 2011, 151.

88. Peters 2011, 153–159.

89. Peters 2006, 79.

90. Former RUF commander, quoted in Peters 2006, 79.

91. Peters 2011, 152.

92. Manning 2002, 92.

93. Manning 2002, 94.

94. Manning 2002, 94.

95. Manning 2002, 94.

96. Alexander 1997, 14.

97. Alexander 1997, 14.

98. Manning 2008c, 188.

99. Manning 2002.

100. Specifically, he oversaw the appointment of delegates in Zambezia, Tete, Manica, and Nampula provinces (Manning 2002, 96).

101. Manning 2002, 96.

102. Vincente Ululu in 1994 addressing potential international donors in the run-up to elections (Manning 2002, 105–6).

103. Afonso Dhlakama, quoted in the Waterhouse and Lauriciano 1993.

104. Manning 2002, 114.

105. Manning 2002, 114.

106. Manning 2002, 111.

107. Manning 2008b, 55.

108. Frelimo won 129 seats in the assembly with 44.33 percent of the vote, and Renamo won 112 seats with 37.78 percent of the vote (Africa Elections Database).

109. Governance provisions for the RUF are detailed in Articles III, IV, and V of the Lomé Peace Accord (UNSC 1999).

110. UNSC 1999, 7.

111. The demobilization packages entailed counseling, a small “reinsertion allowance,” and relocation to a community in which the ex-combatants would receive vocational training in any of a variety of fields (Humphreys and Weinstein 2007, 539).

112. Mitton 2008, 200.

113. Keen 2005.

114. The quote is from Richards and Vincent 2008, 88. As I discuss in the introduction, a number of recent works code this transition as a success because the RUFP appeared on the ballot in the 2002 national elections (Acosta 2014; Söderberg Kovacs and Hatz 2016; Manning and Smith 2016), and thus “participated in competitive electoral politics” (Acosta 2014, 671). This argument is largely a response to others who claim that the RUF was either fighting solely to control diamonds (Collier 2000), or that it was little more than a proxy force against ECOMOG troops run by Charles Taylor of Liberia (Gberie 2005).

115. Hannan and Freeman 1984.

116. Richards and Vincent 2008, 93.

117. Richards and Vincent 2008, 93.

118. Richards and Vincent 2008, 102.

119. Peters 2011, 151.

120. Mitton 2008, 202.

121. Mitton 2008, 202.

122. Interviews with RUF commanders reveal not only the nature of the differences but also the fact that higher-ranking members of the organization knew to some extent how distinct Kailahun’s sect was from RUF units elsewhere. One commander recalls, “In Kailahun . . . there were no food-finding missions, because the people were producing it themselves” (Peters 2011, 170).

123. Peters 2011, 170.

124. Peters 2011, 106.

125. Peters 2011, 170.

126. Peters 2006, 82.

127. Peters 2006, 169.

1. Söderberg Kovacs and Hatz 2016; Allison 2006, 2010; Manning and Smith 2019.

2. Rudolph 2008, 92.

3. Starting in 2018, the FARC party was allocated five seats in the House of Representatives and five in the Senate, irrespective of vote share. This quota is set to remain in place until 2026.

4. On who joins rebellions and why, see Gurr 1970; Gates 2002; Humphreys and Weinstein 2008. On how and when rebels deploy types of violence, see Weinstein 2007, Gutiérrez-Saním and Wood 2017. On how, when, and why rebels engage in behaviors beyond violence, see Mampilly 2011; Arjona, Kasfir, and Mampilly 2015; Arjona 2016; Kasfir, Frerks, and Terpstra 2018; Stewart 2021.

5. Collier and Hoeffler 2004; Humphreys 2006; Weinstein 2007.

6. Dresden 2017; Ishiyama and Widmeier 2020.

7. Staniland 2012; Arjona 2016; Staniland 2021; Mampilly and Stewart 2021.

8. Autesserre 2009; Arjona 2016.

9. On revolutionary groups, see Lipset and Rokkan 1967; Kitschelt 1994; Zielinski 2002. On religious organizations, see Kalyvas 1996. On labor and trade unions, see Collier and Collier 1991.

10. Hale 2006, 8.

11. Huang 2016b.

12. Hannan and Freeman 1989, 3.

13. O’Donnell and Schmitter 1986; Huntington 1991; Karl 1990.

14. Bratton and Van de Walle 194.

15. Grzymala-Busse 2002.

16. Knight 2008; Muggan and O’Donnell 2015.