Webmeister's note: |
For a better view select,
View Fonts Larger
Was the 16th century a demographic catastrophe for Mexico?
Department of History, University of Minnesota
Department of History, University of Minnesota
The "guerra de números"--the debate over the size of native populations at contact and the degree of depopulation which followed--continues in the 1990s (Rabell 1993:35), although not with the intensity as when the matured research of the maximalists, Sherburne Cook and Woodrow Borah, appeared in the early 1960s. The minimalist camp, founded by Angel Rosenblat, has grown in recent years. (I adopt Hugh Thomas' Latinate nomenclature of "maximalist" and "minimalist" as less inflamatory than the vulgar, Anglo-Saxon "High Counters" and "Low".) Subjected to unrelenting criticism by Angel Rosenblat (1967), neither Cook before his death nor Borah since 1966 when he presented a brief paper to the Americanist Congress in Mar de Plata, Argentina, chose to directly engage their opponents (Borah, 1976). In subsequent decades, Cook and Borah's scenario of demographic doom for the native population of Mexico over the course of the sixteenth century was contested on almost all fronts: their data, methods, manipulations, interpretations, and the entire thesis of demographic disaster (Rosenblat, 1967; Sanders 1976; Zambardino, 1980; Henige, 1992; Brooks, 1993). For the Colombian quincentennary, Denevan offered a timely summary of the debate. Then in 1993, the history of the first virgin soil epidemic to strike Mexico, that of 1520-1521, came under frontal attack. The disaster was reduced to "a mild attack of smallpox, such as occurred in contemporary Europe with some suffering, some deaths, and little further effect" (Brooks, 1993).1
Why do Rosenblat, Sanders, and Florescano discount the smallpox epidemic of 1520-21? This thesis is not new. Angel Rosenblat, William T. Sanders, and, now, Enrique Florescano also discount this epidemic. As this essay will show, Rosenblat simply ignores the evidence for the first smallpox epidemic, and all other evidence on the effects of contagious disease. Sanders sees only two major epidemics in the sixteenth-century--none before 1540--but he does not consider any evidence for 1520. Florescano cites "terrible death tolls caused by the epidemics of 1545-1548, 1563-1564, 1576-1581, and 1587-1588," but, in doing so, exludes all smallpox epidemics from his list and any great contagion before 1540. Florescano seems to endorse Sanders' thesis that no major epidemics struck before 1540.
I was surprised to find that Rosenblat, the most unyielding defender of the minimalist camp, also disregards the role of smallpox. My consternation increased when I realized that Sanders and Florescano exclude the smallpox epidemic of 1520 from their lists of deadly epidemics of the sixteenth century. If the orthopox virus did not contribute to demographic collapse in Central Mexico in 1520, then the catastrophe school of contact population history is in error. If the best documented case of a virgin soil epidemic is wrong, extension of the paradigm to other first encounters between Europeans and Native Americans becomes tenuous, if not untenable.2
Fortunately, a re-examination of this case is facilitated by the many extant sources in Spanish and Nahuatl--eyewitness accounts, extensive tax records for a large number of native villages and towns, many inquiries by secular and religious authorities, and chronicles by both conquerors and the conquered. Only in recent decades have many of these sources been subjected to close scholarly scrutiny. To begin my survey of the evidence, I turned to informed historiographical reviews of the epidemiological history of sixteenth-century Mexico.3
The radical implications of the revisionist argument and their tactical, time-saving strategy of favoring philology over quantification, narratives over numbers, were so beguiling that I was lured into sorting out the issues of conquest demography for myself using non-quantitative methods. A cross-examination of published narratives on the first Old World epidemic to strike Central Mexico should help determine whether smallpox was a minor epidemic or a major catastrophe.
I deliberately chose a non-quantitative approach because non-quantifiers show an unrelenting hostility to the quantitative evidence. Thus, Cook and Borah's numbers and tribute data were not considered for this exercise. Since the minimalist interpretation rests mainly on narratives rather than numbers, I studied a wide-ranging body of sixteenth-century documents including annals, genealogies of rulers, reports, chronicles and histories. Responding to Rabell's call for a "nueva lectura" of the sources, I re-examined those used by Brooks--accounts by Hernando Cortés, Francisco López de Gómara, Bernal Díaz del Castillo, and the Franciscans Toribio de Motolinía and Bernardino de Sahagún--and all other published sources in Spanish and Nahuatl which seemed pertinent. Those who know the evidence well will not be surprised by my conclusions. Any comprehensive survey of the most authoritative sixteenth century accounts will show that contemporaries feared that native peoples might become extinct because of raging epidemics.
Since silence is an awkward adversary (most minimalists do not explain why they ignore or discount smallpox in 1520), the following analysis is directed principally at Rosenblat, but my arguments serve as well to counter historians who deny the force of smallpox without having pondered the evidence for themselves.4
Unlike Brooks, Rosenblat, Sanders and others who minimize the importance of smallpox, I am confident that the mortality from the virus in New Spain was demographically significant, several times greater than in Europe. Any alleged similarity with Europe--or, more generally, that the impact of smallpox was minor--is not corroborated by the very considerable mass of evidence published on this subject over the past 470 years. (Like Brooks, Rosenblat, and the minimalists in general, I use only published sources, but my cross-examination is more thorough than theirs.) As nearly as I can determine every sixteenth-century writer who mentioned populations trends of the natives invariably concluded that a precipitous decline was under-way, and that disease was an important, if not the main cause. An extended discussion of this evidence is presented in my essay entitled "Spanish and Nahuatl Views on Smallpox and Demographic Catastrophe in Mexico." Here, I summarize the observations of Motolinía, Sahagún, Pomar, Muñoz Camargo, and López Velasco.5
Around the middle of the sixteenth-century historians and chroniclers began to reflect on the severity of the various epidemics. Motolinía, writing in 1542, saw three great devastations, which he sought to fit to years ending in "one," the most important being the war, pestilence and famine of "1521". Several years after his manuscript was shipped to Spain (and while its author was in Guatemala), the great devastation of 1545 broke out so we cannot know how his numerology would have taken that catastropic epidemic into account.6
On November 8, 1576, as the third great epidemic of the century unfolded, Sahagún, in a rare direct intervention in the General History and for which there is no corresponding Nahuatl text, mused whether the present plague would exterminate the native people. He addressed the question directly and forcefully, leaving no doubt that the smallpox attack of 1520 was exceedingly lethal ("murio casi infinita gente")--more deadly even than the war--but the deadliest was the matlazahuatl epidemic of 1545, "una pestilencia grandisima y universal, donde, en toda esta Nueva España, murio la mayor parte de la gente que en ella habia." In Tlatelolco alone Sahagún claimed to have buried 10,000, before falling ill to the disease himself. As he wrote in November 1576, the number of deaths mounted daily. According to Sahagún, many were dying of hunger, without care, and with no one to provide even a jar of water--charitable relief having been exhausted. He feared that if the contagion continued for another three or four months there would be no natives left, that the land would revert to wild beasts and wilderness. He reasoned that, on the one hand, Spaniards were too few to settle the land and, on the other, the Indians were becoming extinct.7
Pomar, the historian of the city of Texcoco, also singled out three great epidemics of the century, 1520, 1545, and 1576, but characterized that of 1520 as the worst ("haber hecho mayor daño que en las que después acá han tenido, sin otras muchas pestilencias que han tenido de menos furia"). He reported that Texcoco, which surrendered to Cortés without a struggle, used to number some 15,000 citizens ("vecinos") but did not have 600 as he wrote in the 1580s. Many smaller subject villages had disappeared entirely.8
My favorite assessment is the most explicitly quantitatively-reasoned, that by Muñoz Camargo for the province of Tlaxcala also drafted in the 1580s but only published in 1981:
Yo digo que la primera  devio de ser la mayor por que avia mas gente, y la seg[un]da  fue ansi mismo muy grande por que la tierra estava muy entera, y esta ultima  no fue tan grande como las dos prim[e]ras porque aun que murio mucha gente escapo mucha con los remedios que les hazian los españoles y religiosos...9
Evidence from a wide variety of sixteenth-century Spanish and Nahuatl sources point to a single conclusion: the smallpox epidemic of 1520 ranked among the three worst demographic crises of the century. The death rate from smallpox and starvation in 1520-1521 was probably less than for the matlazahuatl epidemics of 1545-1546 and 1576-1577. Nonetheless, if we accept the intelligence offered by one of the most celebrated native chroniclers of the colonial era, the smallpox epidemic of 1520 was the greatest demographic catastrophe of the century for the Nahuatl-speaking people of Central Mexico. Moreover, neither the Spanish nor native accounts are limited to the immediate environs of Tenochtitlan. A comprehensive, quantitative town-by-town survey of population figures, laboriously extracted from Relaciones Geográficas, tribute counts, censuses, or whatever sources were available reveals the broad geographical extent of depopulation.10
Raging disagreement over the size of the pre-contact native population obscures the widespread consensus that the sixteenth-century was a demographic catastrophe for the native peoples of Central Mexico and that disease, exploitation and ecological disruption were the principal agents. All population figures for the early modern era are riddled with errors of every conceivable kind. Yet, for some historians, narrative is easily misread, generalization difficult to elicit, or, perhaps number simply has its own fascination. Table 1) offers a summary of population numbers at two points in the sixteenth-century, 1519 and 1595, for many widely-cited, authoritative efforts by modern historians. The last column of the table reports the percentage of population decline over the century implied by each set of figures. Where authors report ranges rather than a single number, these are included in the table. Figure 1 converts column three into a graph, depicting the percentage of population decline over the century implied by the figures of these authorities. Included are data for "Mexico"--Cook and Borah's area of a half-million square kilometers (but excluding Yucatán, Chiapas, and the North)--the Central Valley and Kubler's 128 towns in Central Mexico, Michoacán and Oaxaca. Figures for the total population of "Mexico" at contact vary from 4.5 million (Rosenblat, Aguirre-Beltrán) to 30 million (Cook and Borah). This enormous range reflects the paucity of data as well as fundamental disagreement over how the few data that are available should be interpreted. If a number for the native population in 1519 is desired, a lengthy chain of supposition and extrapolation is required. Even so, some figures are more nearly guesses than others.11
Rosenblat, the Argentine linguist (he always disclaimed the sobriquet "historical demographer," which others might be tempted to bestow), characterizes his work as nothing more than "de vaga aproximación" "sin fanaticismo" based on "verosimilitud" or a "razonable probabilidad," but "lo único factible" nonetheless. His reasoning is worth citing:
el análisis de esos trabajos [de Cook y Borah] me reafirma en mis cálculos moderados de 1935, que aspiran, sin fanaticismo, a ser solo un índice relativo, de vaga aproximación, lo cual me parece por lo demás lo único factible. Me guiaba entonces fundamentalmente, en cada paso hacia lo desconocido, el criterio de verosimilitud, o de razonable probabilidad, y no creo aún hoy que puede caber otro.12
Unfortunately Rosenblat, for all his diligence, simply failed to carry through this laudable spirit of research, or more importantly, to consider the mass of informed contemporary accounts of Spanish origin or even a single native or mestizo chronicle. Rosenblat, over three decades of publishing on this subject (1935-1967), scarcely revised a figure--nor was his widely cited work subjected to the slightest scholarly scrutiny. For Mexico, he chose numbers compiled before 1570 by the royal chronicler López de Velasco, particularly those for individual towns and villages. For 1492, Rosenblat refers the reader to an appendix of sources, but upon close examination I was shocked to find none listed for Mexico before 1570. In its place, he criticizes the figures of Mexicanists such as Mendizabal and Kubler.13
Rosenblat reasons that his pattern of change is basically in agreement with that of Kubler (who writes of "appalling mortality in the sixteenth-century"), but this is not the case. His empty rhetoric continues to mislead even quantitatively informed historical demographers (Rabell Romero, 20). Figure 1 makes clear that Rosenblat stands alone in defense of a decline in the native population of less than 25%. His position would be unproblematic if he stood on solid ground. His sympathy for Kubler's figures, which point to a fifty percent drop over the century, does not extend to an acceptance of the pattern of decline sketched by Kubler's work. Likewise, Rosenblat accepts López de Velasco's figures, but ignores the chronicler's narrative accompanying the numbers! López de Velasco, after studying a vast array of information compiled by the imperial bureaucracy, concludes that for the realm as a whole "al principio los naturales fueron muchos mas en numero de los que despues ha habido, porque en muchas provincias, donde habia gran multitud dellos, han llegado casi a se acabar del todo." In other places, he sees the population as recovering, but from the relative extent of his text, López de Velasco was clearly more attentive to decline than recovery.14
Rosenblat criticizes the catastrophe school of contact demography for failing to take into account the recovery potential of the native population. His objection went unmet until recently, when Thomas Whitmore published a series of wide-ranging, sophisticated epidemiological simulations. Some will dismiss this work as historical fiction of a quantitative sort, but Whitmore's simulations offer a partial answer to Rosenblat's question about the role of demographic recovery between epidemics. From Whitmore's simulations, based on morbidity and mortality rates obtained from scrutinizing the world-wide historical record, we learn that the likely levels of epidemic mortality probably overwhelmed the homeostatic potential of the indigenous population--even without taking into account deaths due to war or social and environmental disruption. The simulations also reveal what Rosenblat fails to appreciate--the devastation of epidemic disease. Rosenblat reasoned:
Realmente, si en mi estudio llegué a cifras moderadas o bajas para la población de 1492, no fue porque yo me lo propusiera así. Los datos de que disponía sobre la época de la Conquista no me llevaban a más, a no ser que supusiera unas enormes y horrendas matanzas, muy del gusto de una imaginación macabra pero que me parecían inverosímiles con las técnicas de exterminio del siglo xvi.15
Had Rosenblat paid faithful attention to the texts of López de Velasco, López de Gomara and other early chroniclers, instead of "técnicas de exterminio" he would have considered a more powerful, socio-biological mechanism--disease. My study of Rosenblat's sources for Mexico and the considerable body of other texts reviewed here and in my "Spanish and Nahuatl Views" (1995) leads me to reject his vague approximations on precontact populations, and, more importantly, to reject the pattern of demographic decline implied by his figures. Attentive readers of early Spanish and Nahuatl narratives witness a demographic catastrophe, indeed a succession of catastrophes, of unsustainable magnitudes. If numbers are required to comprehend this, they must agree with the narratives. Native reasoning recorded in the Relaciones Geográficas in the last quarter of the sixteenth century confirm this picture. In Michoacán, of 23 villages where the causes of demographic catastrophe are mentioned, 20 blame "pestilencias y enfermedades." The crude figures accompanying these texts point to declines of 60-80%. War is not mentioned once as probable cause, but over-work is noted in several instances as is divine intervention and other less credible factors. While historians continue to insist on the newness of the thesis that disease was the main cause of depopulation, sixteenth century narratives readily testified to the primacy of epidemics, as did many later chroniclers and historians. López de Velasco attributes the decline to war, oppression, and "enfermedades nunca vistas en aquellas partes como fueron las viruelas que les pegaron los españoles." López de Gómara also emphasizes carnage "no a fierro sino de dolencia." Miguel Othón de Mendizábal, in an essay first published in 1939, argued persuasively against the importance of war mortality ("debemos concluir que la mortalidad militar juega un papel poco importante en el decrecimiento colonial de la población indígena") and instead, after reviewing a large number of sixteenth-century Relaciones Geográficas, quoted the following contemporary explanation as typical "'pestilencias grandes que a avido en diversos tiempos y años'."16
At the opposite pole from Rosenblat stand Cook and Borah's figures, but they are not alone. Clavijero, writing at the end of the eighteenth-century, argued that all the chroniclers were in agreement that the pre-conquest population was great--but no one would hazard a guess as to the actual number. He suggested 30 million as a possible figure, but noted at the same time (as Cook and Borah also did almost two centuries later) that any number is subject to a wide margin of error. Cook and Borah's numbers are, however, an order of magnitude greater than most that preceded them and have provoked the greatest controversy and disbelief. The Berkeley empiricists sought to move the debate from the arena of belief to evidence, by using documentation on tribute paid by subject towns. The long chain of assumptions and conversion factors necessary to transform tax payments into tax payers into total population convince many students of the subject that the exercise is untrustworthy. Yet, Cook and Borah's writings are less dogmatic than critics would have us think. Borah noted many years ago: "The more the number of agents and agencies entering into the gathering and processing of materials, the wider are the margins of error. ... For those [estimates of population] of the first century, it seems likely that the most we can hope for is estimates of the order of magnitude." Cook and Borah's figure of 25.2 million for the contact population of Mexico is widely cited, but few cite their range of figures--18-30 million. Their critics--Rosenblat, Sanders, Zambardino and others--might reply that even this range is a wild exaggeration. What I find remarkable in Figure 1 is the correspondence among scenarios of demographic disaster, theirs and their critics--with the exception of Rosenblat.17
The problem is illustrated by Hugh Thomas's prodigiously researched Conquest of Mexico, which wrestles mightly with the numbers on the native population of Mexico in 1520. Thomas, to keep the match fair, stays within conventional historiographical rules and wrestles with all the "inspired guesses" on population size. He recounts the bitter rivalry between the maximalists, the "California School," and the minimalists such as Rosenblat, Zambardino, Kubler, Sanders, and others. I propose that we abandon the old rules on how best to estimate population size, that we focus our attention instead on the question of the magnitude of population decline. Whether the ratio of decline is 1:2 (a decline of 50%) or 1:25 (96%) is important, but unfortunately the quality of the quantitative evidence often does not permit such precision. Rosenblat aside, when the issue is the degree of decline, all the authorities on the demographic consequences of conquest are maximalists. All agree that the native population fell at least by 50% over the sixteenth-century, a demographic catastrophe regardless of the exact percentage. The fury of the "guerra de números" blinds us to an acceptable paz de narraciones.18
For smaller areas, population estimates require less extrapolation, and the range of uncertainty contracts accordingly. All the researchers on this subject who rely on primary sources--Mendizabal, Kubler, Rosenblat, Gerhard, Gibson, Cook, Borah, Percheron and Sanders--encourage research at the local level. Thus, Sanders' estimate for the "Central Mexican symbiotic region" is based on a sampling of archaeological excavations and tribute lists sprinkled over an area of "only" twenty thousand square kilometers. Since Sanders does not consider the epidemic of 1520 as being demographically significant (see 129, figure 4.4), his figures for 1519 should be inflated by some factor (1.1, 1.2, 1.3 ... ?) to take into account the smallpox devastation documented by Nahuatl and Spanish sources alike. Gibson and Kubler also compile figures for a select group of settlements with data at two or three points over the sixteenth-century. Mendizabal developed a detailed portrait for towns in Michoacán, and his figure for "Mexico" is based on bold extrapolations from those data, which he then used to adjust the calculations of the Italian scholar, Dino Camavitto. At finer levels of analysis, the same ghastly picture emerges, a demographic collapse of 50-80%, even when settlements which disappeared entirely are excluded from analysis (the method deliberately chosen by Kubler).19
It seems to me that close attention to contemporary narratives will lead to consensus on the scale, causes, and consequences of the demographic disaster which struck sixteenth-century Mexico. There is agreement that a demographic catastrophe occurred and that epidemic disease was a dominant factor in initiating a die-off, beginning, in Central Mexico, with smallpox in 1520. But the role of disease cannot be understood without taking into account massive harsh treatment (forced migration, enslavement, abusive labor demands and exhorbitant tribute payments) and ecological devastation accompanying Spanish colonization. Killing associated with war and conquest was clearly a secondary factor, except in isolated cases, such as the devastation of Cholula or the leveling of Tenochtitlan.
A fair-minded cross-examination of the broad range of primary sources for the epidemic of 1520 leaves little doubt that smallpox swept through the Central Mexican Basin, causing enormous mortality. The epidemic ranked with the deadliest disasters that native annals customarily recorded. Whether the fraction of smallpox deaths was one-tenth or one-half, we have no way of knowing, but from my reading of the texts discussed here, the true fraction must fall within these extremes, perhaps near the mid-point.
The iconoclast's position on pre-contact population size was recently staked out by David Henige: "despite three centuries of sporadic guessing, culminating in fifty years of intensive investigation, it is still not possible to claim that any number, or any range of numbers beyond a certain irreducible minimum, is significantly more likely than any other number or range of numbers." I agree that the numbers remain controversial, but the narratives are incontestable. If we leave aside the controversy over numbers, there emerges a broad agreement in the Spanish and Nahuatl narratives and in the patterns of decline sketched by historians. After almost five centuries of writing on the subject, there is a consensus that a demographic catastrophe occurred in sixteenth-century Mexico and that it began in 1520 with the first smallpox epidemic.20
For narrative-bound historians, there exists a great library of published Spanish and Nahuatl texts on the demographic misfortunes of conquest and early colonization. For historians who abide quantification, experts point to overall levels of demographic destruction over the sixteenth-century for Central Mexico exceeding 50%, probably ranging beyond 75%, and even topping 90% in some large regions such as the tropical lowlands. Vociferous debates provoked by the maximalists' figures, such as those of Cook and Borah, often obscure the similarities in scenarios of demographic collapse between the maximalists and minimalists, leaving aside Rosenblat. Even Aguirre-Beltrán, who accepts Rosenblat's figure for 1519, discards his numbers for later in the century and thus rejects Rosenblat's pattern of population decline.21 From my reading of the evidence, the revisionist stance is unsustainable because it neglects the devastation of the first epidemic to strike Mexico, smallpox in 1520.
To reduce historiographical uncertainty further will require much additional, careful sifting of archival and archaeological evidence--tasks which, in recent years, few seem inclined to undertake.22 In the meantime, I find convincing the testimony of the Oidor Licenciado Francisco Ceynos, who sums up the opinion of many enlightened sixteenth-century Spanish observers. Ceynos, after five years as fiscal on the Royal Council of the Indies, arrived in Mexico in 1530 to sit on the Real Audiencia of Mexico City. An oidor for more than thirty years, he fought against the widespread practice of enslaving Indians and against the extreme labor and tribute burdens common in that era. On March 1, 1565, he completed a lengthy recommendation on colonization policies suitable for newly conquered regions. As preamble, he reviewed briefly the demographic tragedy of Spanish colonization in Mexico:
y es cierto que del dia que D. Hernando Cortés, marques del Valle, entró en esta tierra, en los siete años, poco mas o menos, que la conquisto e goberno, padecieron los naturales grandes muertes, y se les hicieron grandes malos tratamientos, robos y fuerzas, aprovechandose de sus personas y haciendas, sin orden, peso ni medida; ...disminuyose la gente en gran cantidad, asi por los excesivos tributos, y malos tratamientos, como por enfermedades y viruelas, de manera que en este tiempo faltó muy grande y notable parte de la gente, y en especial en tierras calientes.23
We do not know what number, percentage or ratio Judge Ceynos had in mind for "grandes muertes," "gran cantidad" or "faltó muy grande y notable parte de la gente," but his narrative has the ring of truth. He reported a disaster on a scale unimaginable to contemporary Europeans. If five centuries later this thesis remains beyond the domain of "reasonable probability" for some historians, their number, too, is diminishing as the evidence of demographic catastrophe accumulates.
The nadir of the demographic disaster is usually placed in the seventeenth-century. I chose 1595 for an end-point, not because I believe this to be the nadir of the native population, but to be able to interpolate, rather than extrapolate, comparable figures for the largest number of authors. *Nevertheless, the figure for Sanders is extrapolated from 1568.
Miguel Othón de Mendizábal, "Demografía mexicana. Epoca colonial 1519-1810. Demografía colonial del siglo XVI. 1519-1599," in Obras Completas (Mexico: 1946; first published 1939), III, 309-338.