From Wil Pansters, ed., Violence, Security and the State in Mexico (forthcoming, Stanford 2011)                                                                1 | Who killed Crispín Aguilar? Violence and order in the postrevolutionary countryside Paul Gillingham Crispín Aguilar was a pistolero, one of the violent entrepreneurs who struggled for power across most of Mexico after the armed revolution. During Aguilar’s lifetime his murky biography interested journalists, peasants and policemen, who held “el patrón” responsible for the deaths of hundreds of men, women and children, some of whom vanished into a secret burial ground in the hills above his home town of Actopan in central Veracruz. A handful of hastily pencilled, near- unreadable pages in the Gobernación archives form one of his testaments, part of the “Incomplete list of crimes committed by … Crispín Aguilar”: a catalogue of 94 killings over six years.1 Different types of social scientist could find different reasons to be interested in Aguilar’s life. A sociologist might delve into his struggle to remain a node of multiple political, economic and criminal networks, and to dominate a radical agrarista peasantry.2 A political historian might find his links to the army and President Miguel Alemán worth a look.3 Legal historians might be interested in his trial, acquittal and subsequent jailing; while a criminologist or an aficionado of discourse analysis might like the Lombrosian press descriptions of Aguilar and his peers as “…short, fat, “rechoncho” with a bulky, dangling belly, “like a tear”, which is the work of the beer that they put away like pachyderms do water. The complexion battered by drink; red, it shines and sweats, without the hair of a beard or moustache. The top of the skull, flattened, is not abundantly covered and on the inside, the few good ideas which there may be have their own caciques, which are the bad 2 | thoughts. As is generally seen in that class of subhuman being, the lower lip hangs down, a sign of stupidity and inconfessable degeneration. The bulging red eyes, the low brow, indicative of little intelligence for good, but plenty for evil.”4 Crispín Aguilar was killed in a cleverly choreographed ambush at the noisy height of sábado de gloria in Actopan. More than his life or filmic death, however, it is perhaps its date which is of the greatest interest; for Aguilar did not die in the convulsive agrarian violence of the 1930s, but rather in March 1950.5 Crispín Aguilar was a major figure in the politics of a major state from the late 1930s to the early 1950s. He was a key lieutenant of Manuel Parra: a violent agrocapitalist whose pistolero, political and financial networks underpinned a regional power stretching over 25 municipios.6 But Aguilar was far from unique. On the rich lands of central Veracruz Aguilar jostled up against several rivals, the principal of whom, Rafael Cornejo Armenta of Plan de las Hayas, was capable of mustering large bands of armed men to “invade” Aguilar’s territory.7 Similar competition persisted through the early 1940s in southern Veracruz, where coalitions of ejidos under Juan Paxtián and Nicolás Parra fought it out with cattle raids, canefield ambushes, kidnappings, killings of dependents and hamlet burnings in a “reign of terror” that aimed to build agrarista fiefs and personal livestock businesses.8 Contemporary Veracruz enjoyed a reputation as a failed state. Yet violent entrepreneurs like the Aguilars and Armentas were to be found in many parts of Mexico, variously dubbed caciques, matones, luchadores, valientes, gallos and pistoleros. They were the flagship species of rural societies which remained markedly violent long after the traditional watershed of 1940: the year, so Mexicans and Mexicanists once concurred, when the lead of the revolution was transmuted (by the philosophers’ stone of cardenismo) into the época de oro of political stability and economic development. In 1950 the municipio of Ometepec, Gro., for example, had a homicide rate high enough 3 | to place it among the twenty most violent municipios in Antioquia, Colombia (the third most violent department during the Violencia.)9 The idea that Mexican and Colombian provinces might be compared in terms of violence during the 1940s and 1950s is admittedly counter-intuitive. The metanarrative of modern Colombian history revolves around a peculiarly high degree of violence, making violence “a privileged historical referent…one of the poles of attraction for … social investigation.”10 Mexico, on the other hand, has been repeatedly presented as exceptional in its post-1940 stability. Very different historians have begun questioning the overwhelming weight of violence in interpretations of the Colombian past.11 This study contends that Mexico’s post-revolutionary narrative of rapid PRIísta pacification demands parallel re-examination. Measuring violence then and now: methods and hypotheses The continuity of high levels of violence in the Mexican countryside after 1940 has hitherto been underestimated for three main reasons: a lack of reliable primary material, pronounced obstacles to accessing what documents do exist, and clear disincentives to wandering around Mexican hamlets asking who killed who.12 A primordial difficulty is that contemporary metropolitan elites – politicians, journalists, generals, spies and diplomats – had a distinctly tenuous grasp on provincial reality. Subordinates and citizens routinely provided fictions concerning most aspects of country life, from harvests to homicides. Censorship made the fourth estate little use as an open intelligence source.13 Mexico City newspapers, particularly the “quality” press, were extremely effectively controlled in this period: when rebellion broke out at Balsas, Gro., in 1947, Excélsior ran two short articles on the 4 | “poorly armed cattle rustlers” and then dropped the story.14 Provincial newspapers were less reliably managed, but their contributors experienced pressures ranging from elegant cooptation to outright murder. When Veracruz’s government wished to remove the lurid nota roja from the Diario de Xalapa they gave the editor access to wire service and a deputyship.15 Vicente Villasana, the flamboyant PANista editor of Tampico’s El Mundo, was less fortunate; he was murdered by a state police commander.16 The consequent battle for “cognitive capacity” – the “sustained organization to collect, process, analyse and deliver the types of information about society needed for a modern state to monitor and interpret the impact of its measures” – is a staple of contemporary accounts.17 Realising that neither federal nor state bureaucracies provided reliable reports on the countryside, President Alemán expanded the intelligence agencies, founding the Dirección Federal de Seguridad (DFS) and inflating the Dirección General de Investigaciones Políticas y Sociales (DGIPS) budget some 500%.18 DGIPS agents regularly found out how violent that countryside really was in the 1940s and 1950s: one, dispatched to the coast of Oaxaca, was given an automatic rifle, 300 rounds and two platoons as an escort (and was attacked nonetheless.)19 Yet many of the actors involved in rural violence had strong motives to cover up; as we shall see below, even victims could collaborate in under-representing acts of violence. The DGIPS, moreover, remained a small agency even after its dramatic 1940s expansion. Personnel files reveal a handful of full-time spies who were used intensively; the director could only spare fifteen operatives to monitor the 1952 elections across provincial Mexico.20 Thus the description and quantification of violence, despite the federal government’s interest, remained one of the blanker regions of the state’s generally fuzzy cognitive map. In 1950 the governor of Guerrero reported 44 arrests 5 | for homicide. During the same year Ometepec, a market town of 18 000 people, recorded 23 murders in the Registro Civil.21 Early PRIístas were unsure as to how violent the countryside genuinely remained; they often suspected the worst, and kept their suspicions as quiet as possible; and quite often they seem to have been right.22 Just how right they were is difficult to assess, for studying violence in post- revolutionary Mexico is not straightforward. What primary sources exist are not just unreliable; they are often unreachable. Uncontroversial collections of documents can be of limited use due to periodisation schemes which end history (more effectively than Francis Fukuyama managed) in the 1940s; the cataloguing of even the excellent Veracruz state archive ends in 1949. As for the security archives, the reforms of the Fox sexenio were less sweeping than they seemed. The DFS, responsible together with the Guardias Presidenciales for much presidential dirty work in provincia, opened its doors a decade ago, but only recently has offered a useful search process. The IPS catalogue, helpfully released on disc, is distinctly incomplete. The drawn out liberalisation of the military’s archive has been particularly frustrating. Researchers are now allowed far more than the traditional, carefully-negotiated access to personnel records. Yet the Archivo General de la Nación’s well-ordered, thematically- catalogued military collection has two major flaws: its 470-odd volumes cover little more than a decade, from the late 1960s to the early 1980s; and they omit monthly zone commander’s reports.23 These institutional essentials would give historians new levels of insight into the army’s central role in maintaining internal order, which may be why they are absent. Monthly zone reports occasionally crop up in other archives; in the mid-1940s they are punctiliously divided into “military, political and social activities”; they are extremely revealing.24 6 | These archival missing links have been reinforced by careerist sanctions against scholars and violent sanctions against journalists who are interested in violence.25 The historians of the Special Prosecutor’s Office for Social and Political Movements of the Past were fired without pay and their report, a detailed critique of the military during the counter-insurgency in Guerrero, was edited to remove names and evidence. The perpetrators of Mexico’s disappearances were themselves to be disappeared from public history.26 Manuel Parra – the greatest pistolero of them all – rose from the dead to literally haunt sociologist Antonio Santoyo when he delved into Parra’s life forty years post mortem.27 While the current, extremely high, death rate amongst journalists in northern Mexico is notorious, inquisitive journalists in PRIísta Mexico always ran risks. Finally, such lacunae may also obey theoretical preferences: it is difficult to square high levels of post-1940 political/state violence with either neo-Gramscian interpretations of the revolution or corporatist models of the post- revolutionary state. None of these problems are insuperable. There is abundant qualitative data on violence in the reports of DGIPS agents, more prolix than their (more thuggish) counterparts in the DFS; in political correspondence at the lowest levels of Mexican government, particularly that between comisariados municipales and ayuntamientos; in state judicial archives; in the voluminous protests countrymen sent to the Presidency and Gobernación; in those provincial newspapers which resisted censorship, such as Ignacio de la Hoya’s magnificently plainspoken La Verdad de Acapulco; and in the work of contemporary ethnographers such as Gonzalo Aguirre Beltrán, Paul Friedrich and Marcos Muñoz. (Some of whom had a clear elective affinity for caciquismo; Friedrich was told “You could be a leader here, Paul, you could order killings.”) 28 7 | Quantative data is more complicated, despite Pablo Piccato’s useful compilation of federal and state crime statistics for the twentieth century.29 Federal homicide statistics were not compiled until the late 1930s, providing no valid baseline for post-revolutionary levels of violence. The series that does exist is moreover incomplete. Not all years are covered, and some of the missing data, such as that covering Guerrero in the late 1970s, or Veracruz in the mid-1940s, coincide with periods of unusually intense violence. The lack of a census for 1980 means, additionally, that the population data necessary for calculating homicide rates grows sketchy just as Mexico’s population grew fastest. Above all, there is a fundamental reporting problem: fear of reprisal and mistrust of local government meant that homicides in particular went routinely unreported. Reviewing the Cuajinicuilapa, Gro., registro civil for 1948, Aguirre Beltrán found that only 30% of all deaths went recorded; Veronique Flanet obtained a similar estimate of reporting frequency twenty years later in Jamiltepec, Oax.30 A 1948 protest from the eastern highlands of Guerrero pinned twenty murders on the Salgado brothers – well-connected cattle rustlers – and explained the rationale for victim under-reporting: “There was no investigation by any Authority of these killings, it was kept quiet and the very mourners knew…who they were and said nothing to avoid the same fate….Generals and representatives of the Sr. Gobernador have come and have left Huamuxtitlán with a few thousand pesos in their pockets and they go and mislead the Governor that all is quiet and that there are no such troubles.”31 “It is customary”, one Veracruzano told the president, “that anyone who sees anything and tells tales to the Authorities is murdered.”32 The consequent omissions of reporting inject a tangible unreality into governmental accounts of crime.33 In some cases official homicide counts actually decline as violence peaks; and no one was using household surveys to provide alternative statistics in 1940s Mexico.34 “It is 8 | important”, Moisés de la Peña noted in the late 1940s, “to point out that that the data on recorded presumed crimes does not represent anything but a minimal part of the real number of crimes committed”.35 Yet this is a standard problem: using police or military statistics as an index of the real level of violence is “highly suspect” in many societies.36 Colombian statistical records are similarly variable: the Ministry of Justice homicide rate for the violencia in Antioquia is roughly half that calculated by the department’s governor.37 There is something of a paradox in that the more significant a society’s homicide rate is, the more difficult it becomes to measure accurately. (And the less politically convenient: an instance of Campbell’s Law, which holds that the more politically-charged a statistic is, the less likely it is to be accurate.)38 Thus civilian casualties during the occupation of Iraq, 2003-2006, were estimated to sum between 44 and 48 000 by the occupiers; and 655 000 – an order of magnitude greater – by The Lancet.39 The Piccato dataset is clearly imperfect: it makes Guerrero one of the least violent states in 1940s Mexico. Yet it has multiple, critical applications. If we assume either constant or rising reporting rates since 1940 – which given the growing “degree of stateness” across the period is a justifiable assumption40 – then its finding of a long term decline in the national homicide rate from 1940 onward is reliable.41 Furthermore, acknowledging the crude quality of both Mexican and Colombian statistics should not prohibit their comparison. At a regional level, such an exercise reveals broadly similar rates of homicide in Guerrero, San Luis Potosí and Veracruz in Mexico and all bar the most violent Colombian departments between 1946 and 1960.42 This, it must be stressed, can only be a tentative, suggestive finding. But it is reinforced by what we might call “grassroots statistics”: the homicide rates revealed by interviews and by the libros de defunciones of registros civiles. In La Montaña, Gro., (notably less violent 9 | than Tierra Caliente or the coast) 1953 figures gave results of c. 100 homicides per 100 000 population for Ahuacotzingo and Xochihuéhuetlan, and considerably above that for Olinalá and Huamuxtitlán.43 The high forced migration rates that can be extrapolated from ejidal census results are further indicators of violence. In Tulapan, Ver., in a zone where opportunist agrarian bosses battled with peculiar ferocity, the 1952 census revealed that half the original inhabitants had abandoned the ejido.44 The Mexican countryside remained a profoundly violent place in the 1940s and early 1950s: army officers in Guerrero stressed the “intense” and “very common” incidence of homicide, while in Veracruz an agent reported that “not a day [passed] without underhand murders.”45 Both government and grassroots statistics reinforce qualitative evidence that the PRIísta state was not born of any pax cardenista.46 160 140 Antioquia 120 Cauca 100 Tolima Guerrero 80 Veracruz 60 San Luis Potosí 40 20 0 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 Figure 1.1 Three-year moving averages of homicide rates per 100 000 population, selected Mexican states and Colombian departments, 1948-1960. 10 |