A Reevaluation of the Role of War Captives in the Aztec Empire

Abstract

Traditionally, historians believed that taking captives was a major goal in Mexica warfare, and this tendency has even been given as a reason why the Spanish conquistadors defeated the Mexica. Although historians have largely revised these conclusions, the perception that captives were important to Aztec strategy and warfare persists. In this article I argue that the need for captives was not great enough to affect Aztec military strategy or battlefield conduct. First, rituals only needed a small number of victims, which could easily be acquired through the normal course of battle, and thus did not constitute a specific objective. Second, Mexica strategy focused on economic objectives, rather than captive taking. Finally, individual warriors were not well equipped to take prisoners. Although captives played a vital role in Mexica society, the practice should be thought of as opportunistic, rather than strategic.

Tradicionalmente, los historiadores creían que la toma de cautivos era un objetivo importante en la guerra azteca, y esta tendencia se ha dado incluso como una razón para que los conquistadores españoles derrotaran a los aztecas. Aunque los historiadores han revisado en gran medida estas conclusiones, persiste la percepción de que los cautivos eran importantes para la estrategia y la guerra aztecas. En este artículo sostengo que la necesidad de cautivos no era lo suficientemente grande como para afectar a la estrategia militar azteca o a la conducta en el campo de batalla. En primer lugar, los rituales sólo necesitaban un pequeño número de víctimas que podían adquirirse fácilmente a través del curso normal de la batalla, por lo que no constituían un objetivo específico. En segundo lugar, la estrategia azteca se centraba en objetivos económicos, más que en la toma de cautivos. Por último, los guerreros individuales no estaban bien equipados para tomar prisioneros. Aunque los cautivos desempeñaban un papel vital en la sociedad mexica, la práctica debe considerarse oportunista, más que estratégica.

Historians have argued that taking captives was an important part of Mexica warfare, that it served as an overarching goal for political expansion, and that it explained how the Mexica were able to find enough victims for their sanguinary religious practices. It even partly explained why Spanish conquistadors defeated the Mexica: the Mexica fought to take prisoners, whereas the Spaniards fought to kill. More recent scholarship has challenged many of these assumptions, revealing the Mexica's political and economic motivations for conquest and redefining the role of bloodletting in their religion.

Nevertheless, the idea that the Mexica were bloodthirsty, taking captives for human sacrifice, persists among nonexperts. As late as 2019, Camilla Townsend (2019) felt the need to address their violent reputation, directly challenging the idea that the Mexica went to war to obtain sacrificial captives. No recent scholarship has comprehensively examined the reasons why the captive theory is no longer prominent within the historiography. To that end, this article first argues that Mexica sacrificial rituals only needed a small number of victims, not enough to require a deliberate strategy of captive taking. Second, Mexica political strategy focused on achieving economic goals and political stability, not on taking captives. Third, I argue that Mexica warriors were not especially well equipped to take captives and that their battlefield tactics were not conducive to taking prisoners. Although captives were still valued by warriors and society, seizing them was opportunistic, rather than a specific goal.

Much of our information on Mexica life comes from conquistadors, principally Hernán Cortés (1986), Bernal Díaz del Castillo (1963), Andrés de Tapia (Fuentes 1963), and the Anonymous Conqueror (1917). Priests—including Bernadino de Sahagún, creator of the General History of the Things of New Spain (1959, 1975, 1979, 1981), also known as the Florentine Codex; Diego Durán (1971, 1994); and Toribio de Benavente Motolinía (1950) provided even more detailed accounts of Mexica history and society. Combining these sources has enabled historians to get a good idea of what life was like for the Mexica before the conquest.

Nevertheless, although the information provided in these accounts still forms the basis of most historical studies, they have some significant limitations. The Spanish were prone to misunderstanding Mexica religion and religious practices. Furthermore, political motivations shaped the writings of most, if not all, Spanish chroniclers. The friars, for their part, were tasked with dismantling Aztec religion and replacing it with Christianity, and so they had an incentive to denounce traditional religion. With respect to war captives, the key problem is that Spanish sources exaggerated the frequency and scale of human sacrifice, creating the impression that the Mexica needed an endless stream of prisoners. Díaz del Castillo (1963:276) himself states that, even during his imprisonment, Motecuhzoma “never ceased his daily sacrifices of human beings.” Subsequent historians, who relied heavily on these sources, largely accepted such claims. Consequently, the idea that warfare was aimed at providing victims for the Mexica's religious needs was believable.

The writings of William H. Prescott, whose retelling of the Spanish conquest set the benchmark for subsequent narratives, neatly illustrate this problem. Although a dedicated and thorough historian, he relied extensively on Spanish sources, thereby glorifying Cortés and his achievements. Mexica perspectives are lacking, and Prescott frequently derides the “native superstition” and “immorality” of Mesoamerican religion. Consequently, “in battle, they [the Mexica] did not seek to kill their enemies, so much as take them prisoner” (Prescott 1972:34), ostensibly for sacrifice. Prescott clearly viewed the practice as detrimental to the war effort. This interpretation has been echoed by later historians, such as Tzvetan Todorov (1982) and Hugh Thomas (1993), who, at least partly, blamed the Mexica's defeat on their stronger desire to take captives rather than kill their enemies.

Miguel León-Portilla (1992:99–104) posited that a high-ranking, imperial official named Tlacaelel instituted significant political and ideological reforms that took place mostly during the reigns of Itzcoatl (1427–1440) and Motecuhzoma Ilhuicamina (1440–1469). The key development was the institutionalization of the Flower Wars against Tlaxcala, which happened sometime during the reign of the aforementioned Motecuhzoma and directly led to an increase in the scale and frequency of sacrifice among the Mexica. He characterized the core concepts of this new state ideology as “mystico-militaristic,” which demanded that Mexica warriors capture prisoners in battle for sacrifice. The most comprehensive explication of this interpretation comes from Geoffrey Conrad and Arthur Demarest (1984), who outline the ideological origins of the Mexica Empire and Tlacaelel's role in its creation. This divine mission not only drove Mexica expansionism but also left them politically unstable and militarily inflexible, contributing to their eventual defeat.

In the early twentieth century, George Valliant (1978) emphasized the Mexica's spiritual view of warfare. Richard Townsend (2009:228) stated, “It was an Indian warrior's priority to capture an enemy alive,” and Inga Clendinnen (1991:116) claimed, “The Mexica warriors sought captives, not corpses.” Although these scholars acknowledge economic motivations for expansion, a close examination of captive taking was beyond the scope of their work. Interestingly, general histories and conquest narratives tend to portray captive taking as a common practice among Aztec peoples, contradicting the mystical-militarist interpretation that described seizing captives as a specifically Mexica practice.

Several historians have looked closer at Mesoamerican warfare, producing detailed texts on warriors, military strategy, and armaments. Ross Hassig (1988) examined the practical and strategic aspects of Mexica warfare, including the Flower Wars. Barry Isaac (1983a) challenged the believed role of war captives by demonstrating that Mesoamerican battles were bloody affairs, with high casualties for both sides. Ian Heath (1999) provided an overview of Mesoamerican warfare in Armies of the Sixteenth Century 2. More recently, Isabel Bueno Bravo (2007) and Marco Antonio Cervera Obregón (2011) studied Mexica warfare and warriors, with the latter analyzing their weapons and equipment as well. Generally, these historians present captive taking as just one of several goals of warfare, alongside economic expansion, protecting trade networks, and eliminating political opponents. This contextualization of war captives is an important step forward. Nevertheless, although war captives are no longer presented as the sole or overarching motive for warfare, the core idea of their significance has yet to be fully refuted.

Religion and War Captives

The Mexica's need for prisoners depended on the number of victims killed during their yearly ritual cycle. If they sacrificed large numbers of captives, they would have to deliberately engage in captive taking to meet that need. However, if they had a smaller number of victims, the Mexica would not have needed specific battle tactics or techniques to take prisoners. Therefore, the question becomes: How many people did the Mexica sacrifice?

Cortés (1986:36) suggested that 50 people were sacrificed in each temple, with an overall total between 3,000 and 4,000 people annually. Andrés de Tapia alleged that the great tzompantli (skull rack) in Tenochtitlan contained the skulls of 136,000 sacrificial victims (Fuentes 1963:42). Durán (1994:339) frequently commented on the scale of Mexica sacrifices. He claimed that the largest number of sacrifices at one time occurred during the AD 1487 dedication of the Templo Mayor, when 80,400 victims were killed on Ahuitzotl's orders. Some scholars agree with these high numbers, suggesting that the Mexica killed between 20,000 and 50,000 captives each year (Heath 1999:37).

However, none of the chroniclers had an accurate way of measuring the number of victims slain. Their figures are guesses, often made decades after the fact and with little physical evidence to support them. Furthermore, some of their claims defy logic. Take Durán's impossibly large figure of 80,400 victims. We must question how the Mexica could have managed the logistics of such an event. How did they feed and house so many captives? And what did they do with the bodies afterward?

Or take Tapia's description of the great tzompantli. Bernard Ortiz de Montellano (1983:404) calculated that the skull rack could have only held 60,000 skulls at most, less than half of Tapia's estimate; it probably held far fewer than even this reduced number. Durán's (1994:341) account is equally unreliable: he claimed that the great tzompantli was constructed to carry the skulls from Ahuitzotl's great ceremony. The skulls of 80,400 victims, plus those from 30 additional years of human sacrifice, could not possibly have fit on the rack.

Accounts of human sacrifice are often connected to accusations of cannibalism. Durán (1994:233), for example, states that Tlacaelel initiated the Flower Wars, at least partly because he had “acquired a taste for human flesh since the lords ate it so frequently.” The cannibalism argument is attractive because it explains why the Mexica would need so many victims. Some scholars, such as Michael Harner (1977), even claimed that sacrifice, cannibalism, and warfare were a Mexica attempt to cope with their protein-deficient diets. If consumption were the goal, the quantity of “flesh” mattered.

The belief that Aztecs were prone to cannibalism, however, was the result of Spanish stereotypes about Indigenous people, and not an accurate assessment of their beliefs and practices. For example, Bernard Ortiz de Montellano (1978:616) noted that mentions of cannibalism and “fattening up” victims in Sahagún's Spanish annotations are absent from the accompanying Nahuatl text in the Florentine Codex; Sahagún added the text because he assumed that cannibalism was occurring. Even Díaz del Castillo (1963:225), commenting that human flesh was served to Motecuhzoma, noted, “He had such a variety of dishes, made from so many different ingredients, we could not tell whether a dish was of human flesh or anything else.” The idea that the Aztecs were cannibals was a rumor created by the Spanish cadre.

Cannibalism seems to have been limited to a few specific rituals, such as Tlacaxipehualiztli (Agricultural Festival), where portions of the victim were given to the captor's family and to Motecuhzoma (Sahagún 1981:49). Sahagún (1959:67) only mentions cannibalism once more, in relation to the sacrifice of bathed slaves. Outside these scenarios, the Mexica abhorred cannibalism (Isaac 2005). Furthermore, the nutritional explanation for cannibalism was rejected by most historians almost as soon as it was proposed (Price 1978). Yet, although historians have largely dispatched the cannibalism myth, the idea derived from it—that the Mexica fought to take captives—still largely remains.

The Templo Mayor's discovery in 1978 gave archaeologists an opportunity to see the material remains of Mexica rituals. Although they unearthed a vast array of offerings, copal, precious stones, ceramics, and animal bones in buried offering caches, human remains were relatively rare (López Luján 1994). From 1948 to 2011, only 142 victims were found under Mexico City (Chávez Balderas 2014:180). Excavation of the Huey Tzompantli, located near the Templo Mayor, uncovered an additional 445 skulls, far short of the figures claimed by Tapia and Durán (Matos Moctezuma et al. 2017:54). There were other tzompantli at Tlatelolco in Tenochtitlan. However, these skull racks were considerably smaller and held correspondingly fewer victims. Although some remains have probably been lost to time, that so few have been discovered strongly indicates that Spanish accounts were heavily exaggerated and that the Mexica sacrificed hundreds, not tens of thousands, of victims (Chávez Balderas et al. 2015; Luján and Olivier 2010).

Durán's and Sahagún's accounts form the basis for modern understandings of the central Mexican ritual cycle, and their records can give some indication of the number of victims slain yearly. Of the two, only Durán regularly gives numbers, although Sahagún's descriptions are more structured and comprehensive.

Most sacrifices were part of the yearly ritual cycle, with small batches of victims killed roughly every 20 days (Sahagún 1981). The largest sacrifice, Tlacaxipehualiztli, involved around 60 victims (Durán 1971:174), but most rituals were far smaller. Toxcatl (Dry Season Festival) had 12 victims at most (Sahagún 1981:66), whereas Ochpaniztli (Cleansing Festival) had only six (Sigal 2011:156–157). Some festivals, such as Izcalli (Earth Festival), featured a sacrifice once every four years (Sahagún 1981:162). New Fire occurred once every 52 years and had only a single victim (Townsend 2009:134–137). Both Durán and Sahagún report that some ceremonies did not involve human sacrifice, although their accounts are sometimes inconsistent. Clearly, the number of victims bore no relation to the importance of the ritual. Tlacaxipehualiztli, Toxcatl, and New Fire were important but differed greatly in the number of victims required.

Durán's (1971) figures suggest that Tenochtitlan conducted between 250 and 300 sacrifices a year. Caroline Dodds Pennock (2012:283) gives an estimate of between 300 and 600 victims a year, though she leaves open the possibility that the true number of victims was larger. Although the number of slain varied, these figures suggest that the city's yearly victim requirement could have been much lower than the thousands reported.

Even then, not all those sacrificed were warriors captured in battle: victims included enslaved criminals, women, and even children. Pennock (2011:25) concluded that 16 female ixiptla (a person or object representing a deity) were sacrificed yearly. Her assessment is supported by archaeological evidence from the Templo Mayor, which indicates that approximately one in four victims were women (López Luján and Olivier 2010:31). Interestingly, a similar issue regarding female victims can be identified in Postclassic Maya sacrificial practices. Although Maya texts and paintings almost never depict women as sacrificial victims, skeletal remains indicate that they were also sacrificed (Vail and Hernández 2007:157–158). This suggests that captured male warriors may be overrepresented by Maya artists. A similar problem may exist in Mexica accounts, where warrior sacrifices attract a disproportionate amount of attention, giving the impression that they were more common than they really were.

Children were sacrificed to Tlaloc during rain festivals, such as Atlcahualo, and the remains of children were found in the Templo Mayor (López Luján 1994:192). Chávez Balderas (2014:190) noted that women, children, and noncombatants made up a greater portion of Mexica sacrificial victims than in other Mesoamerican cultures. Because children, women, and enslaved people were unlikely to have been captured on a battlefield, the Mexica did not need a specific war strategy to obtain them.

Although enslaved people comprised a minor part of the economy, they played an essential role in rituals and were frequently identified as sacrificial victims in the Florentine Codex (Sahagún 1981). There were several ways for slaves to end up on Tenochtitlans’ sacrificial altars. Some of the victims may have been taken as war loot after the pillaging of defeated towns. Others were bought by merchants from slave markets, the most notable of which was at Azcapotzalco (Sahagún 1959:45). These slaves were ritually bathed, adorned with jewelry, and then sacrificed at the Temple of Huitzilopochtli. The proportions of the victims who were enslaved individuals and warriors are unknown. However, analysis of oxygen isotopes in sacrificial victim remains indicates that approximately half of those recovered from the Templo Mayor, and almost all from Tlatelolco, had lived for 10 years or longer in the Valley of Mexico (Moreiras Reynaga et al. 2021:13, 16). Given that warriors were likely sacrificed soon after capture, this finding suggests that the sacrificed individuals were enslaved people who lived in the vicinity of Tenochtitlan for some time before death. This corroborates earlier isotope studies that suggested that many remains came from slaves, not warriors (Chávez Balderas 2018:152).

Most ceremonies involved a range of activities, of which human sacrifice was just one element. Toxcatl focused on the activities of Tezcatlipoca's ixiptla for an entire year (Sahagún 1981). The Mexica had exacting standards about his physical fitness and appearance (Sahagún 1981:66–68). The individual chosen for this role was taught to play the flute, adorned with jewels, and then ritually “married” to four women. Other rituals included elaborate dances and even competitive games (Durán 1971). Ochpaniztli involved ritual dances, mock battles, and the ritual scattering of seeds (Sigal 2011); human sacrifice was a comparatively brief part of the event. For some ceremonies, such as Ochpaniztli, it was not even the ritual's climax. This does not mean that sacrifice was unimportant. It was a critical part of both the Mexica's religious and political practices. The point, however, is that sacrifices were not simply an exercise in butchery: the victims came from diverse backgrounds, and war captives comprised only a portion of the total.

Generally, the need for war captives has been viewed as an extension of the Mexica's “pessimistic” worldview. Soustelle (2002:101) stated, “The ancient Mexicans had no real confidence in the future.” Clendinnen (1991) and Pennock (2011) also referenced Mexica fatalism. Sacrifices and, by extension, captives were needed to stave off cosmic disaster. According to Matthew Restall and Amara Solari (2011), this pessimistic view of Mesoamerican culture owes more to European theology than Mesoamerican philosophy. The Franciscans were deeply influenced by apocalyptic medieval millennialism and spread these beliefs to their Indigenous converts, who in turn reflected them back at their teachers (Phelan 1970).

More recent studies have developed better understandings of Mesoamerican rituals and their meanings, often by including Indigenous art and writing in their analysis. In Cacería, sacrificio y poder en Mesoamérica, Guilhem Olivier (2015) examines the close relationships between hunting, worship, and nature in Mesoamerican rituals. He describes how legends justified warfare and sacrifice, and then how sacrifice legitimized political power among the ruling class and helped establish hierarchical relationships both between and within societies. Another relevant study is Carlos Javier González González's (2011) Xipe Tótec: Guerra y regeneración del maíz en la religión mexica, which examines the relationship between Xipe Totec, agriculture, and warfare. Pete Sigal (2011) explores Nahua ideas about rituals, sexuality, and regeneration in The Flower and the Scorpion. Although difficult to summarize, this work supports the idea that Mexica religion focused on balance, growth, renewal, and regeneration, which explains how their ritual practices interacted with their society.

Economics and Political Stability

If seizing captives were a major objective for the Mexica, we would see evidence of it in their military strategy. However, instead of destabilization, raids, and surprise attacks that would provide them with opportunities to take captives, the Mexica built trade networks, extracted tribute, and exercised political control both directly and indirectly. Captives and tribute were not mutually exclusive goals, but they could conflict in practice, and the Mexica favored economic gains over taking captives.

Consider these examples. During the reign of Motecuhzoma I, Cuetlaxtla revolted against Mexica rule by murdering a group of imperial emissaries. The enraged Motecuhzoma then sent armies to crush Cuetlaxtla. After a brief battle, the Mexica entered the city in victory. Cuetlaxtla's citizens, who did not support the war, surrendered to the Mexica without further resistance (Durán 1994:198). The Mexica then could have sacked the city and taken the population captive. Instead, they spared the citizens, merely executing the rebellious lords, and increased their tribute demands (Durán 1994:200). The Mexica preferred to accept the surrender of a targeted city without armed confrontation, foregoing the opportunity to take captives.

A second example is the Mexica conquest of Tlapan. This region, in eastern Guerrero, had endured decades of warfare between local Aztecs and Tlapanecas. The Mexica entered the conflict in AD 1486. By analyzing the Lienzo de Chiepetlan 1, Gerardo Gutiérrez (2014:162) infers the Mexica's strategy. They formed an alliance with local Aztecs whose help allowed them to outflank the many fortifications constructed throughout the region, thereby avoiding the sites that were too heavily fortified to take easily. Thus, Mexica strategy was shaped mainly by practical considerations such as how to manage the topography of the land, how to exploit local rivalries, and how to best avoid the enemy. As was typical of Mexica conquest, allies played a key role: the empire depended on cooperation from regional actors for its success (Bueno Bravo 2007:342). Yet, every ally was someone who would not be taken prisoner. Certainly, the Mexica could have taken captives when the opportunity arose, but all this suggests that the need to take captives did not dictate strategy. If it did, then the Mexica would have simply attacked everyone.

Instead, the goal of Mexica conquest was tribute. The two main sources for Mexica tribute demands are the Codex Mendoza (Berdan and Anawalt 1992b) and the Matricula de Tributos (Códice de Moctezuma; Berdan and Durand-Forest 1980). Both texts reveal the variety and quantity of goods received by the Mexica. Of the Mexica's 38 provinces, 21 sent maize, beans, amaranth, and chia to Tenochtitlan (Berdan and Anawalt 1992a), enough to feed 200,000 people. Even more provinces sent fabric goods; only Xoconocho and Tepeacac were exempt. Still others sent luxuries such as gold, feathers, cacao, cochineal, wood, and rubber. Durán (1994:417) even recorded that Mexica artisans instigated a war just so they could get their hands on a type of sand used in gem working. All these goods were either essential for Tenochtitlan's food security or were an important component of the Mexica's political reward system.

If captives were as important as believed, they should be depicted more frequently in the tribute rolls. Yet, the Codex Mendoza indicates that only Tepeacac sent captives to Tenochtitlan (Berdan and Anawalt 1992a:98–101). Conversely, Durán (1994:329) claimed that subject cities routinely sent captives to Tenochtitlan. How do we account for this discrepancy? Possibly, the prisoners Durán mentioned were actually slaves. Enslaved people, as previously noted, came from a variety of backgrounds, such as criminals and debtors, and thus were not necessarily war captives. Regardless, their scarcity suggests that captives were sent to Tenochtitlan on an ad hoc, rather than regular, basis and that they were less important than other goods.

Although trade was always important in Mesoamerica, markets became most developed during the late Postclassic, Tlatelolco being the prime example. Although small traders conducted most commerce, it was the professional pochteca who were most closely allied to the state, because they brought in the luxury goods coveted by the nobility. They were even sent into enemy territory to conduct espionage against potential rivals (Sahagún 1959). This was a dangerous task, and merchants were sometimes killed as spies. The Mexica would retaliate against such attacks, using the killing of merchants to justify their aggression. Trading did not necessarily stop warriors from taking captives, but its prominence makes it clear that the Mexica's goals were economic, rather than religious.

The Mexica Empire has been described as “conquest without consolidation” (Conrad and Demarest 1984:53), but the Mexica developed several strategies to ensure stability. Although the threat of violence played a role, the Mexica could not have retained their conquest by fear alone, and rebellions were less frequent than expected if this were the case (Smith 1986). Primarily, the Mexica used marriage alliances to build a network of relationships that bound other dynasties to Tenochtitlan's ruling elite (Hassig 1992). The rulers of Tenochtitlan, Tlatelolco, Texcoco, and Tlacopan were all related (Chimalpahin Quauhtlehuanitzin 1997). These leaders were all closely involved in imperial governance and were even consulted when a new tlatoani was elected, thereby integrating them into the Mexica elite (Durán 1994:247). Representatives from conquered territories came to live in Tenochtitlan where they received lavish gifts (Smith 1986). The Mexica relied on the cooperation of these nobles who became important agents in their own right. Most city-states supplied logistical support for imperial armies and sent warriors to fight alongside them (Hassig 1992).

Where they lacked that support, Mexica imperial leaders resorted to more direct means of control. Although garrisons and governors were rare, some were established in restive cities like Chalco and Cuetlaxtla (Bierhorst 1992:110). Fortifications were also built in some key strategic areas. The most comprehensive examination of imperial administration is probably The Tenocha Empire of Ancient Mexico by Pedro Carrasco (1999). His analysis reveals the complex relationship between vassal states and the Triple Alliance capitals. However, he also elaborates on the role of the calpixque in collecting taxes, the role of the lords in settling disputes, and the installation of governors. Although these structures were still developing, the Mexica Empire was becoming increasingly organized, and by the early sixteenth century, Mexica officials directly administered many Tepaneca cities (Berdan et al. 1996:36). Ultimately, these strategies allowed them at least some control while keeping their administration costs low; they also minimized warfare within the core empire, which would have resulted in fewer opportunities to take captives.

We must still reckon with the Flower Wars, which have traditionally served as proof that taking captives was a major goal of the empire. The Flower Wars were allegedly triggered by a famine that occurred during Motecuhzoma Ilhuicamina's reign in the mid-fifteenth century (León-Portilla 1992). Tlacaelel blamed the famine on a failure to conduct enough human sacrifices and started the Flower Wars against Tlaxcala and Huexotzinco to ensure a steady supply of victims, as per the mystico-militarist theory. However, there are problems with this explanation. Susan Schroeder (2016) noted that Indigenous sources portrayed Tlacaelel's motives for reforms as political and economic, not religious. His apparently religious motivation for instigating the Flower Wars would be unusual. Motecuhzoma offers an alternative interpretation to Tapia, telling him that they spared Tlaxcala so they had a place to train their warriors (Fuentes 1963:33). Frederick Hicks (1979) agreed, noting that accounts of earlier Flower Wars were not connected with human sacrifice and may have started as a training exercise. Florine Asselbergs (2008:40) found the idea that the Mexica deliberately spared Tlaxcala to be dubious, arguing that the potential tribute that could be extracted from the province would have been too lucrative to pass up. Lastly, Hassig (1981) argues that the Mexica's response to famine was to expand the agricultural system within the Valley of Mexico, rather than embarking on further conquests, which ultimately occurred for other reasons.

Hassig (1988) and Isaac (1983b) argued that the Flower Wars were an attrition strategy, intended to wear down Huexotzinco. The Mexica had more manpower and could tie down Huexotzinco's army while encircling the surrounding territories. Bueno Bravo (2007:165, 169) concluded that, although prisoners were a benefit, the Flower Wars’ real aim was to display Mexica power and intimidate other cities into surrendering without the cost of a real war. Alternatively, the Flower Wars may have been a way to excuse the Mexica's inability to conquer the formidable Tlaxcala state. Tlaxcala was surrounded by mountains, giving it a strong defensive position, and it is not surprising that the Mexica struggled to subdue it. The Tlaxcalans themselves explained that their survival was not due to any Mexica policy but to their own stubborn defense of their lands (Díaz del Castillo 1963:179).

Putting the Flower Wars into historical perspective can help adjudicate between these interpretations. Human sacrifice was common to most, if not all, Mesoamerican societies. Nor can the Mexica be held responsible for increasing the scale of human sacrifice. Their rituals were no bloodier, at least relative to their resources, than those of their contemporaries and predecessors, such as the Teotihuacan civilization, which largely shared the Mexica's worldview and cultural practices, including human sacrifice (Sugiyama 1993). It makes little sense to single the Mexica out for their sacrificial acts when so many other Mesoamerican political entities engaged in similar or related practices.

We must also consider Mexica's strategic context. Tlaxcala was not simply another province; it was a member of a political bloc also comprising Huexotzinco and Cholula. The Flower Wars were directed mostly against Huexotzinco: numerous battles were fought between the two powers, involving the mass battlefield slaughter of combatants on both sides (Isaac 1983b). Eventually Huexotzinco's power waned, and Cholula defected to the Mexica. At one point, Huexotzinco made a temporary alliance with the Mexica to help them fight off an attack from their former ally, Tlaxcala (Durán 1994:446). The Mexica would not be willing to form an alliance if their aim were simply to take prisoners; they would have had no reason to stop attacking a nearby and convenient enemy.

Tlaxcala was also close to the trade routes to the Gulf Coast and competed with the Mexica for access to the rich resources—cotton, feathers, and rubber—found in Veracruz. The Mexica viewed Tlaxcala as a source of insurrection and instability, regularly accusing it of fomenting insurrections, encouraging rebellions, and giving military aid to the Mexica's enemies (Durán 1994). Furthermore, Tlaxcala had rich agricultural land and was close to the Valley of Mexico. The Mexica leadership would have certainly coveted access to this rich land for their own purposes. Thus, the wars against Tlaxcala were not a religious exercise but a reaction to a specific strategy adopted by the Mexica leadership.

There is one further type of conflict to explore—coronation wars—that were undertaken after the crowning of a new tlatoani. These events were associated with large numbers of sacrifices, in which captives taken during the campaign were offered at the Templo Mayor. These campaigns were intended to display the power and authority of the new tlatoani. Therefore, the large numbers of sacrifices make sense, because they theatrically celebrated Mexica victories while punishing and intimidating enemies. However, even in these campaigns, taking captives was secondary to political aims. Coronation wars were often directed against provinces that had rebelled after the previous ruler's death: re-subjugation was the main goal. Furthermore, descriptions from primary sources show that these wars still featured large numbers of battlefield killings and that the Mexica made no special effort to secure additional prisoners (Durán 1994:326). They also make it clear that the captives included women, children, and other civilians who were captured along with the warriors (Durán 1994:327). Thus, coronation wars were consistent with most other Mexica conflicts.

Warriors and Battlefield Tactics

Conceivably, individual Mexica warriors could have fought to take captives, regardless of the empire's strategic goals. If this were the case, however, we would see evidence of it in their fighting methods. Their weapons would be designed to incapacitate, rather than kill. They would favor tactics that would net the most prisoners instead of achieving a complete victory. And their accounts of battle would emphasize the taking of captives while downplaying fatal casualties.

Beginning with weaponry, some scholars such as Thomas (1993) and Clendinnen (1991) claimed that Mexica weaponry was designed to wound, rather than kill. However, this is not accurate because Mesoamerican arms could inflict horrible injuries on people (Cervera Obregón 2006). Several conquistadores commented on the lethal power of the macuahuitl, noting that it was powerful enough to kill a horse in one blow (Anonymous Conqueror 1917; Díaz del Castillo 1963). The six-foot-long, two-handed version allegedly could cut a person in half. Although the last macuahuitl has been destroyed or lost, Cervera Obregón (2006) tested a replica macuahuitl on a pig carcass, demonstrating that it could cut flesh very effectively. He also noted that the weapon was unable to cut bone, suggesting that Spanish accounts may have been exaggerated (Cervera Obregón 2006: 134). According to Hassig (1992), the earliest macuahuitl were small but grew in size over the Postclassic, until they developed into the human-sized, two-handed version seen by the conquistadores. This suggests a trend toward more powerful and lethal weapons that were less effective for injuring and capturing opponents. It is possible that the blunt tip of the weapon could have been used to bludgeon opponents into submission, but there are no depictions of them being used this way, and Spanish accounts repeatedly emphasis the macuahuitl's cutting edge, not its blunt tip.

Both Spanish and Indigenous sources indicate that bows, slings, and darts were common among the Mexica, their allies, and Mesoamericans in general. Obsidian arrowheads could easily penetrate naked flesh, and slings could damage even metal-armored enemies. This killing power was unnecessary even against the Spanish; few conquistadores wore metal armor. Neither arrowheads nor slings were effective for taking prisoners. Elite Mexica warriors favored the atlatl, or spear thrower. Darts launched with an atlatl had a longer range and were more accurate and damaging. Each launched dart had 60% greater penetration than a hand-thrown dart (Butler 1975). Like the macuahuitl, which was also associated with the nobility, the atlatl would not be an effective weapon for taking captives because its damage was unpredictable. Its power and the type of wound it was capable of inflicting suggest that the atlatl was a weapon designed to kill (Cervera Obregón 2006).

Battlefield organization and tactics can also be analyzed to determine whether they were geared to captive taking. Clendinnen (1991), for example, gives the general impression that Mexica warriors lacked tactical knowledge. Warriors fought according to ritual concerns, and they even lacked officers and battlefield organization (Clendinnen 1985). Most scholars do not go this far, crediting the Mexica with complex tactics and a high degree of military organization (Heath 1999). Nevertheless, taking captives is still seen as a major concern. Often this ritualized inflexibility is contrasted with Spanish pragmatism, which has been used to explain Spanish success during the Spanish-Mexica War (Soustelle 2002). This portrays the Mexica warrior as a lone wolf with little knowledge of tactics, strategy, or unit cohesion, who lacked the motivation and ability to kill his enemy.

We must question the accuracy of these interpretations. Although depictions of Mexica warriors fighting are brief and lacking in detail, they do exist. Durán (1994), for example, describes multiple battles throughout his History of the Indies of New Spain. Mexica warriors are presented as moving in massed, close formations, rather than advancing as individuals to fight their enemies one on one. However, we should not rely on Durán's account alone, because he did not witness these battles. Díaz del Castillo (1963:289), however, did fight against Mexica warriors, and he describes them in a similar manner, stating, “Nor had they ever seen men so courageous as those Indians charging with closed ranks.” He also writes that the Tlaxcalans “were numerous, and in close formation” (Díaz del Castillo 1963:146). The Anonymous Conqueror (1917:23) also provides a brief description, saying that Mexica squadrons “move with perfect order.” Although individual acts of bravery and heroism were still valued and encouraged, these accounts make it clear that, far from fighting as lone individuals, Mexica warriors massed for shock attacks against enemy lines.

Fighting as a unit would not necessarily prevent individual warriors from taking prisoners. However, it does make it harder to understand how that objective would be accomplished: massed charges could shatter an enemy formation but make it physically difficult to take prisoners. In The Aztec Empire, Davies (1987:228–232) examined some of the challenges a warrior faced in trying to take captives, noting the lack of solid evidence describing how such captures were supposed to be achieved. He concluded that prisoners were probably not taken individually but by groups of warriors working together. Furthermore, Mexica warriors only fought for brief periods, before being rested (Heath 1999:35). Díaz del Castillo (1963) noted that whole units rotated on and off the battlefield. Nor were enemy warriors eager to be captured. In the Book of Gods and Rites and the Ancient Calendar, Durán (1971:113) states, “Some, overdesirous of glory, were captured by the enemy or were slain on the field; often they preferred to be torn to pieces rather than be captured.” Capturing prisoners was far from easy.

For an example, take Motecuhzoma's brother Tlacahuepan, who commanded the Mexica forces that fought against the Huexotzincas in the Valley of Atlixco (Durán is not clear on the date of this battle). Eager for glory, he threw himself into the frontline. On realizing that he was surrounded, he “began to do marvelous things with his sword” (Durán 1994:426), killing 50 Huexotzinca. This story may be apocryphal or perhaps embellished to give a dignified slant to a tragedy (for the Mexica at least). Nevertheless, it makes several points clear. Tlacahuepan was both able and willing to kill his enemies outright, and the Mexica admired him for this. This seems unlikely if captives were so highly prized. Even though Tlacahuepan was eventually captured, he was a high-ranking individual who had been completely surrounded by enemies and was probably exhausted after his heroic exertions. Until that time, the Huexotzincas were unable to hold him, which demonstrates how difficult taking prisoners must have been.

When it came to tactics, Mexica generals focused on defeating their enemies, often through shock attacks supported by missile fire (Hassig 1988). If that failed, the Mexica would outflank their enemies and induce a rout. Conquistadores, including both Bernal Díaz del Castillo (1963) and Hernán Cortés (1986), report that Mesoamericans frequently employed ambushes and feints. Durán (1994) also records that the Mexica made extensive use of spies and scouts to support their military activities. These tactics would not necessarily prevent a warrior from taking a captive, but they did not help him achieve that aim. Largely absent are tactics and strategies that would facilitate seizing captives, such as raids, nighttime attacks, and attacks against civilians. This does not mean that the Mexica never engaged in these types of warfare, but they were not a core part of their military strategy. This strongly suggests that battlefield conduct was not aimed at taking war captives but on achieving military objectives.

We can advance this argument by examining the Mexica's reaction to Spanish technology. Guns, cannons, and horses presented a problem for Mesoamerican generals. Nevertheless, the Mexica quickly responded with their own technological solutions. They countered Cortés's brigantines by putting stakes in the lakebed, effectively controlling the brigantines’ movements (Díaz del Castillo 1963). To neutralize the horses, the Mexica made captured Spanish swords into pikes, a strategy that required great insight and discipline to be effective. To resist guns and crossbows, they armored their war canoes. These responses show that the Mexica viewed conflicts as a series of tactical and technological challenges that required solutions implemented throughout the army, rather than as contests between individuals trying to take captives.

A look at Mesoamerican military encounters supports this position. Isaac showed that battles were bloody, frequently resulting in high casualties for the losing side, and that Mexica warriors were willing and capable of killing great numbers of their opponents on the battlefield. The descriptions of these battles make it clear that these losses were sustained during the active phase of the battle and were not due to prisoners being sacrificed later (Isaac 1983a). Heath (1999:37) also noted the brutality of Mexica warfare and that the importance of war captives may have been overstated: “however, this aspect of Aztec warfare has perhaps received too much attention, and the fact that the Aztecs were willing and able to inflict massive casualties on an enemy during combat has tended to be overlooked.” Fernando de Alva Cortés Ixtlilxochitl (2012) and Diego Muñoz y Camargo (2012) provide accounts of battles fought by the Acolhua and Tlaxcalans, respectively. Although their descriptions are not highly detailed, they give the impression that these encounters were high-casualty affairs.

Although a full and detailed analysis would be beyond the scope of this article, Indigenous depictions of battle in texts and art can also be revealing. If captives were a key part of warfare, we would expect the taking of captives to be memorialized in these texts, as individual city-states sought to glorify their military accomplishments. However, captives were rarely depicted in these documents. The images in Book 12 of the Florentine Codex, on which the text is based, depict the Spanish invasion and conquest of Tenochtitlan. Yet, despite depicting multiple episodes of street fighting, captives are almost completely absent, with the exception of a four-panel section set during the later stages of the war (Sahagún 1975:Plates 152–155). The Lienzo de Tlaxcala depicts the Spanish-Mexica War from the Tlaxcalan perspective and portrays dozens of battle scenes. However, there are remarkably few clear images of prisoners being taken (García Quintana et al. 1983). A similar issue exists for the Lienzo de Quauhquechollan. This text, depicting the Quauhquecholteca's participation in the conquest of what is now Guatemala contains multiple battle scenes, but captives are rarely depicted and only then when the captives were of high status (Asselbergs 2008:173).

Conversely, dead bodies are portrayed quite frequently. Sometimes the corpses bear brutal wounds or are completely dismembered. This implies that taking prisoners was not important enough to record, whereas the violence of the battlefield was remembered. Because these documents (and many surviving Indigenous texts) were written after the conquest, it is possible that the scribes may have eliminated references to captives and therefore to human sacrifice. However, given that Aztecs continued to depict historical rituals in other texts and many of these accounts were not intended for a Spanish audience, this explanation seems unlikely.

One last issue is the promotion of warriors based on the number of enemies they had captured. A few sources suggest that this did occur. Book 8 of the Florentine Codex is an early source for such claims (Sahagún 1979:75). Another is the Codex Mendoza, which contains multiple depictions of warriors of various ranks in the process of taking captives (Berdan and Anawalt 1992b:135–136, Plates 64r, 65r). However, these sources should be treated with caution. Neither the Florentine Codex nor the Codex Mendoza provides definitive explanations for how the Mexica promotion system functioned. Furthermore, some of the details are questionable. Sahagún's informers report that capturing as many as 10 Huastecs merited no renown. This almost certainly does not reflect the actual difficulty of taking so many prisoners: it probably reveals the contempt that the Mexica had for a rival ethnic group, rather than describing an actual part of the promotion system. Furthermore, the entire passage takes place in the context of Motecuhzoma's personally rewarding all warriors involved. Given that military campaigns involved tens of thousands of combatants from multiple cities, this seems improbable. Therefore, it is likely that the described ceremony is an idealized or imagned one, rather than a depiction of a real event.

The images in the Codex Mendoza must also be regarded with caution. In addition to being incomplete, they are not depictions of historical battles (Berdan and Anawalt 1992b:135, Plate 64r), and so it is unclear how these warriors acquired their captives. Their prisoners may even be symbolic, in the same way a burning temple represented a conquered city. In addition, the Codex Mendoza contains depictions of generals and other high-ranking officers but does not explain how they obtained their rank or whether captives were involved (Berdan and Anawalt 1992a:195–197).

Conversely, Durán (1971:197) writes that “brave deeds” were also accepted by warrior societies. Motolinía also implied that Mesoamerican societies rewarded many brave actions. Writing about Tlaxcala, he states, “The reason for this was that Tlaxcallan was larger and had many more warriors, who were bolder and more valiant in killing the enemy and in taking prisoners” (Motolinía 1950:83). Warriors could also win material rewards through scouting and espionage (Durán 1994). This suggests that captives were only one among many ways to earn rewards.

Captives may only have been a target for novice warriors. The Mexica distinguished between at least two different phases of battle—the fight and the rout—and marked the point when enemies began to flee with a war drum, signaling a pursuit (Durán 1994:268). These first-time warriors were only allowed into battle to take captives after the enemy was already fleeing. Sahagún's informants would have been young men during Tenochtitlan's supremacy who did not get the opportunity to advance up the military hierarchy. Therefore, they may not have known the exact criteria for promotion to higher ranks. These novice warriors may have mistakenly thought that they shared the goal of taking captives with elite warriors, especially decades after the Mexica military system was destroyed.

Mexica warriors, however, did have some motivation to take captives. Captives could lead to personal rewards and military promotion, and they fulfilled a religious need. However, there is little in the warriors’ conduct, tactics, and weapons that indicate that this was a major goal. Mexica weaponry was not, given the available technology, suited to disabling or trapping enemies. Nor did the warriors fight in ways that would have allowed them to easily take captives. This is consistent with Barry Isaac's (1983a) analysis of Mexica military history that emphasized that battles were bloody, high-casualty affairs, rather than skirmishes between men trying to take each other prisoner.

The Spanish-Mexica War

As noted, war captives are often brought up in connection with the Spanish-Mexica War. The idea is that the Mexica's unwillingness to kill their enemies in battle allowed the Spaniards to survive otherwise deadly encounters, therefore enabling them to fight another day. This is how both Thomas (1993) and Townsend (2009) introduced the issue. Even Restall (2003:144) states that the practice of taking captives hampered the Mexica's ability to fight the conquistadores effectively. Although these authors are not claiming that the Mexica lost only because they tried to take captives in battle, it becomes a notable factor in their defeat when combined with the image of Motecuhzoma as a weak leader, Mexica fatalistic superstition, and the belief in Spanish military and technological superiority.

Many of these explanations have since been refuted by modern historians. Most modern authors, such as Hassig (2006), note that the most important factor in the eventual Spanish victory was the assistance of native allies. The Tlaxcalans are the best example, providing critical military support to the conquistadores throughout much of the war. This support included fighting men, supplies, medicine, labor, and intelligence. Without their assistance and that of other allied polities, the Spanish would have been defeated. Furthermore, several authors, including Restall (2003, 2018) and Townsend (2003), have argued that Motecuhzoma was a competent and effective ruler. In When Montezuma Met Cortés, Restall (2018) provides a completely revised analysis of the events of the war. Cortés is revealed as a much weaker figure, subject to circumstance and the whims of his captains, while Motecuhzoma is revealed as a commanding figure, unfairly judged by history. Restall provides an incredibly detailed and powerful analysis of the Mexica Empire's disintegration. Rather than resort to simplistic notions of superiority, he explains how the complex political systems working within the empire fragmented as the result of forces unleashed by the chaos of the conflict. In the wake of this book, simplistic explanations for the Spanish victory over the Mexica are no longer tenable. The idea that the Mexica fighting to take captives was a factor in the Spanish-Mexica War should be discarded.

Conclusion

If captives were a key war aim for the Mexica, it would be evident in their military strategy, war conduct, weaponry, and religious practices. They would have sacrificed a larger number of victims and fomented wars, such as the Flower Wars, to provide more opportunities to take captives. They would also have used weapons that were better suited to capturing than killing, and the battles they fought would be less bloody. Although the archaeological record is incomplete, the evidence suggests that the Mexica killed comparatively few victims in their rituals and would not have needed to make captive taking an objective. Furthermore, their strategy was aimed at maintaining this flow of tribute, not creating more wars. Lastly, the Mexica used fearsome weapons, including the atlatl and macuahuitl, and their battles were bloody affairs with massive casualties, especially for the defeated. Although captives did play a role, the seizing of prisoners was probably an opportunistic practice and not a strategic goal.