I'm asked to talk about Latin American revolutions: a big subject, which I'll make bigger by referring to revolutions elsewhere. After all, if, as Chust observes (2015, 21-42), historians of revolutions habitually neglect Latin America, that's no reason for Latin Americanists to follow their bad example. And looking elsewhere can help sharpen our own analytical scalpels—refining concepts, noting contrasts and comparisons, perhaps formulating, if we boldly follow in the footsteps of Tolstoy or Trotsky, the “laws of motion” governing revolutions.
For brevity, I adopt a bullet-point approach: a brief preamble, then six key points: categories; causes; process; violence; plot and actors; and, at the end, the end.
First, I come to analyze revolutions, not to praise them. Revolutions are interesting and important: rewarding to study, risky and unpleasant to live through (assuming you survive). It's not our job as historians to lavish praise or cast blame, that being a job for moralists, preachers, or political activists. I mention this because some (“radical”) accounts are colored by romantic schmaltz, while some (conservative) critiques come laden with moral condemnation. Both approaches—and both, incidentally, exaggerate Latin America's revolutionary record—are misguided: our job is to say what happened, why, and with what results.
First, categories. 'Revolution' belongs to the broad category of what Eckstein called “internal war”: revolts, coups, civil wars, etc. (Eckstein 1964). It makes sense to typologize these diverse phenomena, using specific labels: 'bourgeois' (though that's a real can of worms), “peasant,” proletarian,” “nationalist,” “liberal,” “socialist,” “jacobin,” “popular,” “Eastern,” “Western,” “from above” and “from below.” And, of course, a simple but crucial dichotomy, “failed” versus “successful.” The accuracy and “heuristic value” (the explanatory pay-off) of a label depends on the particular case. Of course, revolutions typically combine elements of several categories, especially where ideological labels (like “liberal” or “socialist”) are conjugated with social designations (“bourgeois” or “peasant”).
The several species of “internal war” cover a wide range, from the rare big beasts of the jungle (“great” or “social” revolutions: my main concern) down to the many small burrowing creatures (revolts, coups, insurrections). “Size” embodies two distinct dimensions, which should not be conflated (compare Mayer 2000, 30): first, the radicalism of the phenomenon (the notional threat it poses to the status quo) and, second, its scope (its impact across society). An explicitly radical challenge (like Mexico's Magonista revolts of 1906-8) may briefly affect a few localities; a moderate—but nevertheless “revolutionary”—movement, like Maderismo (1910-11), can achieve decisive national impact: the fall of the Díaz regime, a clear-cut political revolution. Thus, a swathe of “moderate”, even “traditional,” revolts, sufficiently multiplied, can have radical consequences, even if their proclaimed ideology seems—in the eyes of armchair revolutionaries—tame and pedestrian. We can, again, cite Maderismo; or its tenuous ally, Zapatismo; or flip back to the Independence movements, which have been seen as narrow, localist, prescriptive, and traditional (Van Young 2001)—yet, which nevertheless undermined and finally overthrew the colonial regime. Independence was also a clear-cut political revolution, with socioeconomic ramifications; some subsequent revolutions—the Mexican, Bolivian, and Cuban—were, of course, fully-fledged social revolutions.
Second, it's useful to divide the causes of revolutions into “structural” and “conjunctural”: first, long-term and deeply rooted (social, political, economic, or, in Goldstone's account [1991], demographic); the second, short-term and stochastic (transient wars or crises, the role of leaders, contingent “accidents”). The first correspond to Eliot's “vast impersonal forces,”, the second to Braudel's “l'histoire événementielle.” The first—for example, commercialisation, industrialization, proletarianisation—are inexorable and impersonal; their severity varies and, though they're beyond individual control, their impact can be aggravated or assuaged by individual (conjunctural) decisions. The structural causes of the Mexican Revolution (agrarian commercialization, peasant dispossession, authoritarian social control) were replicated elsewhere (in Peru, Bolivia, Guatemala, though not Argentina or Uruguay); however, they were less concentrated and acute. Also, the Díaz regime, being personalist and “gerontocratic,” failed to mediate and, indeed, exacerbated tensions. The Andean region witnessed plenty of peasant revolts in the period, but they didn't jell into a social revolution (or, in Wolf's phrase [1971], a “peasant war”). But Díaz, with his mismanaged Creelman interview, ostracism of Bernardo Reyes, and promotion of insouciant Porfirian “juniors” like Pablo Escandón, aggravated, rather than assuaged, growing sociopolitical tensions. Of course, personalist authoritarian regimes—from Charles I to Ceauçescu—tend to behave like this, because they lack constraints and are cocooned in comforting webs of yes-men and sycophants. (D. Trump, please note).
Distinguishing structural and conjunctural causes is a matter of judgement (and rarely quantifiable). We can posit a 'risk period'—when structural tensions run high and extraneous crises or poor decision-making can have severe costs—and “normal” periods when—such tensions being muted—regimes can withstand crises. Díaz lost the plot when socio-political tensions were severe; just as the Habsburgs and Romanovs plunged into the First World War at a time when both dynasties—and the regimes they headed—faced severe internal challenges.
The Latin American Wars of Independence are again pertinent. The old notion of purposive, patriotic revolutions ousting an oppressive colonial regime—a notion which, of course, could be both crudely nationalistic and politically self-serving—has given way to a revisionist consensus which stresses that, rather than revolutions, these were “civil wars”; that the wars were caused, not by internal tensions, but the Napoleonic invasion of Spain; that, absent this extraneous event, the empire would have happily survived, perhaps (according to a fanciful counter-factual) mutating into a Hispanic Commonwealth of separate-but-equal kingdoms (note: not “colonies”); and, finally, that patriotism played little part in this process, since the kingdoms-turned-republics of the 1820s lacked nationalist sentiments (which only appeared much later) (Womack 2024, 112, 120, 125, 128, 341). In short, the Napoleonic invasion was not just the trigger, but the basic cause of Independence.
The argument—which I can't address in detail—hinges on the distinction between structural and conjunctural causes. Contra revisionism, I see the Napoleonic invasion as a conjunctural event—a “trigger,” perhaps—which determined the timing of what happened but which, absent deeper structural factors, would not have destroyed the empire. (Compare what happened a century earlier when Spain was engulfed in civil war but there was no colonial breakaway.) Those deeper causes are well-known: mounting taxation, ethnic discrimination, and Bourbon centralization, militarization, and anticlericalism, which provoked serious social protests in the later eighteenth century. And, while a clear-cut nationalist blueprint was absent (as it also had been in the embryonic United States), there were explicit “proto-patriotic” stirrings, allied to powerful “anti-gachupín” sentiments. Furthermore, to expect revolutions to be preceded by clear-cut statements of intent on the part of participants (many, in this case, illiterate and all living under repressive regimes) is historically naive. Revolutions typically confront such regimes and don't follow meticulous, pre-announced plans: they evolve rapidly and chaotically, and actions on the ground count for much more than proclamations before the event. “A revolution,” as someone famously said, “is not a dinner-party or… doing embroidery.”
Independence suggests a second relevant formulation regarding revolutionary etiology: destabilizing pressures from below and permissive factors from above (which correspond roughly to the “structural” and 'conjunctural' causes just mentioned). Independence was triggered (sic) by an external crisis which enabled long-standing internal structural grievances to surface. The old model of the pressure cooker helps: the grievances are the simmering stew, for a time contained by a screwed-down steel lid; but if the lid cracks (as it did in 1808), the stew hits the ceiling. Skocpol's model of social revolution follows this logic (Skocpol 1979): crises, brought on by financial collapse or foreign wars, weaken regimes (i.e., lids) which succumb to revolution (the exploding stew). The key variable is the crisis. As I suggested, the contrast between 1808 (crisis and revolution) and 1700 (crisis and quiescence) calls into question this 'top-down' formulation. The “lid” is no doubt a key variable. But plenty of states experience sporadic crises or costly foreign wars that do not provoke revolution; conversely, revolutions happen that defy Skocpolian etiology. The Mexican, Cuban, and Nicaraguan revolutions were not products of war, defeat, or financial crisis (the Chaco war contributed to the Bolivian revolution of 1952, though more by stoking popular protest than cracking the carapace of state control). Nevertheless, enamored of either Huntington's model or Skocpol's (or both), analysts have still tried to contort the Mexican Revolution according to their inappropriate strictures: exaggerating external pressures or implausibly positing that the 'fall of the old regime' preceded guerrilla warfare and that “the agrarian revolution only occurred after the collapse of the centre.” (Shugart 1989). This is putting the cart before the burro. Guerrilla war preceded the fall of Díaz, and agrarian protest helped cause the collapse of the centre. True, governments can fall for reasons other than popular revolt; occasionally, a state can collapse—very occasionally, a social revolution can occur—in the absence of war, violent protest, and social upheaval: for example, the “velvet revolution” in Czechoslovakia in 1989. The outcome was revolutionary, but the preceding process was not; such phenomena—very rare in history—are perhaps better labelled “implosions” than “revolutions.”
Next, two related points can be briefly mentioned: process and violence. The distinction between “process” (sociopolitical conflict) and outcome (the establishment of a new “revolutionary” regime) is not watertight. Though “outcome”clearly can't precede “process,” the two may intertwine. Thus, agrarian reform can occur unofficially—during decentralised revolutionary upheaval—and, later, officially, when the new regime takes power: a sequence evident in Mexico, China, and Cuba. Again, Huntington's “East/West” typology captures something of this dual sequence, though his nomenclature is misleading. The revolutionary process also creates its own dynamic, which goes beyond purposive plans and manifestoes, yet which may be decisive, leaving a durable legacy: in Mexico, greater spatial and social mobility, hyper-inflation, dearth, disease, demographic downturn, a widespread familiarity with violence and weaponry, and a measure of genuine popular empowerment. Something similar is apparent following the Wars of Independence. Here, it's important to note that the radicalism of the outcome doesn't neatly correlate with the magnitude of the preceding process (the protest, mobilization, and violence). Compared to the Mexican Revolution, the Cuban Revolution was brief and less bloody, but it heralded a more radical revolutionary transformation—including of the basic “mode of production,” something which had not occurred in revolutionary America, France, or Mexico (though it had occurred in Haiti). Again, “facts-on-the-ground” counted for more than “words-on-a-page.”
This raises the question of violence. Is violence diagnostic of “revolution” or can there be “bloodless revolutions?” Experts diverge on this point. But, I repeat, “velvet revolutions”—rapid, peaceful, transformations—are very rare; and could be better termed “implosions,” that is, transformations induced by the non-violent collapse of bankrupt regimes. Recent Latin American transitions from military to democratic regimes perhaps belong to a related category: they did not typically involve popular insurgency or “bottom-up” violence (“top-down” violence, of course, had already happened, perpetrated by the military). Occasionally, peaceful elections can also herald significant socio-economic change: for example, Britain (1945), Argentina (1946), Chile (1970). But the “electoral road to socialism” is a narrow path, rarely trodden (Przeworski and Sprague 1986). The causal link between civil violence, warfare, and revolutionary outcomes is not random: revolutions involve big stakes, vested interests tend to dig in and defend their privileges, and so overturning old regimes is not—to repeat Mao—a dinner party…or doing embroidery.” The correlation between the violence of the process and the radicalism of the outcome is highly variable, but the link between revolution and violence is recurrent and arguably inevitable.
Of course, plenty of political violence has nothing to do with revolution. Political assassinations, military repression, and para-military violence are well-known phenomena, but they are not revolutionary. More important—because the confusion is common—revolutions and civil wars must be distinguished. Both involve violence, and all revolutions (leaving aside the rare “velvet” variety, alias “implosions”) are also civil wars—in some cases prolonged conventional campaigns (e.g., Mexico), in others (e.g., Bolivia) short, sharp, successful insurrections. But not all civil wars are revolutions: history is littered with civil wars—involving rival dynasties, regions, or political factions—which do not embody the threat—or actuality—of revolutionary transformation, that is, a substantial shift in class, ethnic, or political power. When revisionists pointedly refer to the Wars of Independence as “civil wars,” they're stating the obvious. However, they seem to imply that they were not revolutionary (potentially or actually): i.e., they did not involve substantial sociopolitical stakes. Yet that was precisely why people took up arms and risked—and often lost—their lives in pursuit of rival objectives. (A process which, incidentally, can't be adequately explained by rational actor models.)
Penultimately, plot and actors. Theorists of the “natural history” of revolutions have discerned distinct stages through which revolutions pass—for example, Brinton's influential three-stage model, which posited an initial “moderate” phase, followed by a radical lurch to the left, and ending in the “Thermidorian reaction” (Brinton 1965). Brinton's model was heavily influenced by the French sequence (just as François-Xavier Guerra [1985] later built his questionable interpretation of the Mexican Revolution on French foundations). Brinton was cautious in generalising his “French” stage theory, but he applied it to England and Russia (Russia, of course, became the second big meta-example in mainstream “revolutionary studies”). Stage theory of a more abstract kind surfaced in James Davies’ old notion of the “J-curve” (Davies 1962). And, while Hobsbawm rightly criticised Arendt's simplistic periodization of revolutions, he too ventured his own risky version (Hobsbawm, 1977, 202; 1962, 95).
Revolutions—like other historical phenomena—of course pass through phases; and periodization is a constant preoccupation of historians. But the notion that Revolutions follow—invariably or even typically—a common pattern seems to me mechanistic and mistaken. The Mexican, Bolivian, and Cuban Revolutions all followed different trajectories (not to mention the English, French, Russian, and Chinese). There is, however, at least one plausible rule-of-thumb (certainly not a Tolstoyan “law”): moderate, middle-of-the-road liberalism, often evident at the outset, soon withers in the heat of revolutionary conflict; hence, major revolutions rarely conclude in functioning liberal-democratic states. (The big exception, open to debate, would be England's “Glorious Revolution” of 1688). But the French feuillants, the European liberals of 1848, Madero in Mexico, Kerensky in Russia, Sun Yat-Sen in China, the liberals of the late Ottoman Empire, and their Cuban counterparts a generation later—all experienced defeat and disillusionment. Though not a law, this seems a recurrent tendency, and it can be plausibly explained. Functioning liberal democracy requires a period of bedding in, as the rules of the game (e.g., alternancia) are established and players learn to accept them. (Here, game theory may help historical analysis.) But revolutions are fast-moving, disruptive and violent, thus inimical to such evolutionary processes (Traverso 2024, 31). Indeed, revolutions are not only inherently violent but often associated with foreign wars and armed interventions (France, Russia, Turkey, China, Cuba, Nicaragua). “Inter arma silent leges,” as Cicero stated (“in warfare, laws fall silent”); but liberal democracy depends on the rule of law to consolidate. Finally, revolutions are high-stakes sociopolitical conflicts, pitting rival collective groups (social, ethnic, ideological) against each other. These are often zero-sum games, involving powerful allegiances and irreconcilable differences—wars-to-the-death scarcely amenable to the peaceful horse-trading characteristic of liberal democracy. The latter follows Lockean principles, but revolutions resemble Hobbesian wars of “all against all.”
While revolutions don't display patterned regularities in terms of plot, there's a loose logic to class participation (the actors). (I accept that 'class' is not the sole determinant of revolutionary participation—other factors, including ethnicity, ideology, and regionalism may also be important—but class is often central. I also accept that many of the population may be “pacíficos,” caught in the middle “between two fires” [Stoll 1993]). It's significant that the major “social” revolutions of recent times have occurred in agrarian societies and that—as the classic studies of Wolf (1971) and Moore (1969) emphasized—peasants played a key role. Industrial societies have rarely generated social revolutions. Regarding causality, peasants in so-called “modernizing” societies face sustained threats to their livelihoods, in what is often a harsh zero-sum game (industrial class relations, though certainly conflictual, can allow for some sharing of a bigger pie). Those threats go beyond the purely economic to assume a moral dimension, provoking what Moore (1978) called “moral outrage” and James Scott (1976) famously analysed in terms of the “moral economy” of peasant protest. Such approaches explain revolutionary commitment much more convincingly than any rational-actor model. And, finally, peasants—subordinate rural folk—are better placed to mobilize and resist, guns or machetes in hand, than urban workers or middle classes, who are concentrated in the cities, under the watchful eye of the state. Hence Huntington's so-called—though misnamed—“Eastern” model: rural guerrilla warfare, culminating in a revolutionary seizure of the cities, a sequence evident in China, Mexico, and Cuba (though not France, Turkey, or Bolivia).
Finally, we come to the end: a feature of revolutions which, as Hobsbawm and Mayer observe, has received much less analysis than the question of how revolutions start (Hobsbawm 1986, 21; Mayer 39). Some end in violent counterrevolutions: 1848 in Central Europe or the ouster of Arbenz in Guatemala (1954). In each case, the golpistas sought—with some success—to turn the clock back. Of course, some counter-revolutions fail—and, in failing, they may accelerate the revolutionary process: Huerta in Mexico (1913), Yuan Shih-Kai in China (1916), Kornilov in Russia (1917). Foreign counter-revolutionary interventions can have similar counter-productive outcomes: the Austro-Prussian invasion of France (1792), the Bay of Pigs (1961).
These cases represent clear-cut binary alternatives. But some counter-revolutions are more discriminating and partial. In France, “Thermidor” ended Jacobin rule, while Napoleon's “18th Brumaire” eventually overthrew the Republic, but without wholly subverting the gains of the Revolution. The Bolivian coup of 1964 similarly ousted the MNR, but—out of choice or necessity—the military did not dogmatically turn the clock back: the agrarian reform survived, the mines remained nationalized. Clearly, this was a different revolutionary demise compared to the abrupt counter-revolution that ousted Arbenz a decade earlier.
Lastly—undoubtedly the most interesting and important case (not just in Latin America, but the world)—there's the Mexican Revolution. Like the Russian and Chinese revolutions, the Mexican never suffered a sudden reversal (whether immediate or belated). Rather, having accomplished a series of major reforms through the 1920s and '30s, the regime began to mutate, as political generations came and went and both the global context and domestic class structure evolved. There was no sudden counter-revolutionary caesura but, rather, a prolonged, piecemeal “Thermidor,” which was consummated in the regime of the PRI, its eventual neoliberal turn, and its final fall from power in 2000. By then, of course, “the Revolution” had become an item of rote incantation and official mythology (epitomized by the oxymoronic label of the “PRI”). And MORENA's recent re-cycling of that mythology is mostly meretricious window-dressing. Though I am no expert, I suspect that the story in Russia and China may be similar: no sudden counter-revolution, but a sustained—and perhaps predictable—abandonment of revolutionary aims and ideals. This process—the routinization of revolution—was foreseen by Kafka who, when asked about the trajectory of the Russian Revolution, gloomily replied: “as the flood spreads…the water becomes shallower and dirtier. The Revolution evaporates and leaves behind only the slime of a new bureaucracy” (Janouch 1951, 71).