Contents List of Figures ix List of Tables xi Acknowledgments xiii 1. Introduction 1 2. Matlatzinco: A Forgotten Capital 3. What Is a Terrace? 11 26 4. The Landscape Archaeology of Cerro Tenismo 5. The Stratigraphy of Terrace Construction 6. The Stratigraphy of Terrace Destruction 7. Documentary Glimpses and Guesses 40 85 130 158 8. The Anthropology of Landesque Capital 200 9. Legends Black and White: Aztec Agriculture and Its Transformation 10. Conclusions 233 272 Abbreviations Used in Appendix, Notes, and References 283 Appendix: Sources for Reconstructing the Demography of Calixtlahuaca and Toluca, ca. 1470–2010 285 Notes 289 References 319 Index 357 vii CHAPTER 1 Introduction history of agriculture and the agrarian underpinnings of different societies, from the Neolithic to the present day. What archaeologists are slower to appreciate is that understanding the processes of terrace construction, use, and decay is crucial to conducting surface survey or excavation in a presently or formerly terraced landscape, even if agriculture itself is of marginal interest to the project. This is because terracing exercises a profound influence on site formation processes, affecting the preservation and visibility of all archaeological remains. Fortunately, it does so in a patterned way, and the patterns apply across a broad range of environmental and cultural settings. That an archaeologist should single out “agriculture” or “agrarian history” as a central research interest still raises some eyebrows. There are two main reasons. The first is the deeply ingrained perception that preindustrial agriculture barely changed through centuries and millennia and is thus a rather boring and limited field of inquiry. This perception is in turn rooted in the idea that peasant farmers are and always have been the most stubborn and conservative subset of humanity and that, in any given place, there were only one or two ways to grow food, given the rudimentary tools available in premodern times.1 Belief in the capacity of the environment to determine the overall shape of human cultures and the course of history may have been banished from mainstream thought in the social sciences, yet when it comes to agriculture, environmental determinism is alive and well, not only among the general public but among some academics who see “agri-culture” as a part of culture that is only one step removed from nature and its immutable laws. Once we have described the traditional ways of farming in a certain part of the world, they claim, one can generally extrapolate them across the ages. Needless to say, we disagree. Ethnography, history, and archaeology provide multiple examples of farmers taking risks and innovating and document a dizzying diversity of agricultural ecosystems (agroecosystems) This book explores the agrarian history of a settlement in the Toluca Valley, 50 km west of Mexico City. The settlement, known through the ages as Pintambati, Matlatzinco, Callixtlahuacan, and San Francisco Calixtlahuaca, was once the capital of a powerful city-state that controlled the valley. Its standing then declined, and it became successively a provincial center of secondary importance in the Aztec Empire, a depauperate village within the estate of Hernán Cortés and his heirs, and finally a dystopian suburb of the city of Toluca. Throughout the ups and downs of history, its inhabitants have never ceased to grow maize and other crops, sometimes on the slopes of Cerro Tenismo, a mountain on the south side of present-day Calixtlahuaca, and sometimes away from it. The book is based largely on the results of survey and excavations on the terraced slopes of this mountain, which took place in 2006 and 2007. The Toluca Valley, one of the four major regions of Aztec Mexico, is by no means remote, but the other three regions — the Basin of Mexico, Puebla-Tlaxcala, and Morelos (Figure 1.1) — have for a number of reasons attracted the bulk of archaeological research in central Mexico. Thus, the first, most modest goal of this work is to increase the knowledge of a region that rarely registers on the mental map of our fellow archaeologists. Geoarchaeology and the Study of Landscapes Shaped by Agriculture A second goal, one that transcends the specifics of the Toluca Valley, is to showcase a set of methods that we have found useful for studying landscapes that are terraced, or were terraced, at some point in the past. Terraces are a familiar part of the agricultural landscape in many mountainous areas of the world, from the rice paddies of the Far East, through the olive groves and vineyards of the Mediterranean, to the maize, potato, and coca fields of the Andes (see Chapter 3). To understand how and why such landscapes came to be has an obvious intrigue for anyone interested in the millennial 1 = extensive wetlands Jilotepec t o M e z q u i t a l 5000 4000 3000 2000 1000 m a.s.l. Río B a s i n Cerro Gordo Tolman Xaltocan Jocotitlan Mazahuacan? Cuahuacan? d e T o l u c a l a s Ixtapaluca Xico Xochimilco s Capulhuac Iztapalapa Mexicaltzingo e u c Cr Mexicaltzingo Coatlinchan México (Tenochtitlan) Chalco M e x i c o Calimaya Sierra de Chichinautzin V a l l e y Teotenango v a d a N e CALIXTLAHUACA Sabana (MATLATZINCO) Grande Tecaxic Toluca Ocoyacac Zinacantepec Chicahualco Atenco Tlacotepec Metepec Olopa Tezcotzingo Huexotla S i e r r a Tacuba (Tlacopan) Chapultepec Tepetlaoztoc Texcoco o f Apan u apan rr Azcapotzalco a Cihuatecpan Otumba R í o Zah Sierra de Guadalupe ie Jiquipilco Nevado de Toluca Teotihuacan Cuautitlan S t o rma M i c h o a c a n Le La Laguna (Hueyactepec) T l a x c a l a Cerros Blancos Tlaxcala Tizatlan Ocotlan Amoltepec La Malinche Cholula Puebla Tenancingo Ocuilan Cuernavaca Malinalco Sultepec Lomas de Buenavista Cuexcomate Capilco Figure 1.1. Map of central Mexico with places mentioned in the book. M o r e l o s P u e b l a 20km Introduction created by farmers without recourse to modern machinery.2 With parochial pride, we think that archaeology is unique in that it reveals ways of managing the landscape and producing food that have long been forgotten, including ones that received wisdom used to regard as unviable in a particular environment or did not even imagine.3 The other reason why the idea of an archaeology of agriculture defies the imagination is the difficulty of visualizing what durable remains ancient farming could possibly leave. In a normal sequence of events, the crops will be taken away at harvest, the configuration of the field erased before the next planting, and the food consumed. Crop residues and food obviously fall into the category of “perishables,” as do the majority of agricultural implements made of wood, fiber, or leather. It takes some major or minor misfortune — a flood burying a field under a thick sheet of silt, a capsized canoe, a pot left to simmer for too long — to interrupt the routine and create the somewhat exceptional conditions under which rapid burial, waterlogging, or charring will preserve these largely organic materials. All this is true, but over the timescales routinely considered in archaeology, there were many such fortunate (for us) misfortunes. More importantly, certain types of agriculture leave an obvious, widespread, and almost indelible imprint on the landscape in the form of irrigation works, stone boundary walls, buildings for storing and processing crops, etc. The challenge is to assign a definite age to such remains and to establish their agricultural purpose. The form and decoration of a ceramic vessel is often instantly recognizable as characteristic of a certain period of time. The shape of a silted-up ditch tells us nothing about its age, and its purpose may have been to carry away sewage or receive runoff from a country road rather than to aid any agricultural undertaking. A more subtle imprint of past agricultural activity may take the form of soils modified by plowing and fertilizers, or vegetation patterns altered by intentional burning and the selective removal or encouragement of certain species of plants. More often than not, agriculture will also upset the prior workings of the natural ecosystem. Even a modest level of disturbance will alter the way that charcoal and microscopic plant remains (pollen and phytoliths) disperse and how many of them reach lake bottoms and similar places where they tend to accumulate.4 Higher levels of agricultural disturbance will alter the transfers of water and sediment downslope and downstream in a drainage and may radically transform the appearance and behavior of rivers. Traces may be observable in the erosional scars left on a hillside or in the volume and type of sediment deposited by floods in a valley. Dating and establishing a causal link with past agricultural activity in such settings is yet 3 more difficult.5 Even visiting the right places — referred to as “off-site locations” in archaeological jargon — may not be on the agenda of the average project, which will tend to gravitate toward places where human dwellings once stood and where artifacts are to be found in higher densities, rather than toward more distant fields or waterways. The conceptual and methodological shackles constraining the archaeology of agriculture are particularly cumbersome for those who study the precolumbian Americas. The perception of the immutability of traditional agriculture is compounded by the old mischaracterization of this part of the world as one where social complexity and artistic achievement contrasted with technological simplicity and stagnation.6 If we focus exclusively on the tools used to turn the soil, we may have to admit the charge — only a few types of digging sticks and oar-shaped implements appear to have been in use.7 But, as has long been noted, agricultural technology consists of both tools and techniques.8 Techniques are ways of doing things, tools are the portable equipment for doing them. Tools turn out to be of lesser importance than techniques in the history of preindustrial agriculture. The variety and sophistication of precolumbian farming techniques stands in stark contrast to the relative simplicity of the tools. In Mesoamerica and the Andean culture area, these techniques were successful in creating a permanent infrastructure for agriculture of a scale and sophistication that compares favorably with other parts of the preindustrial world — something that we shall subsume later in the book under the label of landesque capital. Agriculture in the American tropics also seems to score high on most measures of agrodiversity, including the taxonomic diversity of crops and other species present in and around fields, the diversity of techniques, and ways in which agricultural labor was organized.9 Methodological obstacles to the study of precolumbian food production stem from the scarcity of domestic animals and virtual absence of small-grain cereals. Bone is one of the few durable organic materials, and in those parts of the world where animal husbandry played a major role, archaeozoology has a major role to play in reconstructions of food production. In Mesoamerica, it provides information on the raising of dogs and turkeys for food, but beyond these two species it primarily informs our reconstructions of the nonagricultural components of human subsistence. The absence of domestic ungulates also precludes multiple opportunities for the preservation of crop remains. The animals graze on or are fed crop residues and food scraps and, when their dung is used as fuel, the plant parts that remain undigested or are scooped up accidentally from the floor of the barnyard stand a good chance of being charred.10 4 Chapter 1 The way that maize and small-grain cereals are handled also makes a crucial difference. Maize cobs are broken from the stem, shucked, and shelled one by one. In contrast, small grains are reaped, threshed, winnowed, and sieved in bulk. As a result, chaff, the seeds of secondary crops, and weeds are intentionally or incidentally brought back to habitation sites, which again increases their chances of being charred and recovered by archaeologists.11 Another unfavorable point of divergence is the abundance of economically important root crops, fleshy fruit, and vegetables in the Americas. Many require neither cooking nor parching before consumption or storage, yet again reducing the chances of accidental charring. The cumulative effect of all these taphonomic differences is that plant remains tend to be extremely scarce at the average site in the Americas. The recovery of a meaningful sample requires that the archaeologist set aside an enormous volume of earth for flotation and sieving. This is not always productive and not always possible because, after all, archeologists want to recover artifacts too. In a situation where archaeozoology is often secondary, and archaeobotany requires inordinate drudgery, the third branch of environmental archaeology — geoarchaeology — often becomes the most promising avenue of inquiry in seeking to shed light on past agriculture. Geoarchaeology, in the narrowest sense of the term, is the study of soils and sediments (in short: earth) in archaeological context.12 The geoarchaeologist focuses neither on artifacts nor ecofacts (animal and plant remains) but rather the matrix in which they are found. The three-dimensional geometry and the composition of the different layers of earth can provide a wealth of information on what activities took place at an archaeological site or an off-site location, as well as what happened once people ceased to frequent the place. Much of geoarchaeology is thus the study of stratigraphy, i.e., the mutual relationships of different layers of both natural and cultural origin. As we have argued above, much of what remains of past agriculture can be described as “earth”: think of the ridges and furrows of the buried field, the soil modified by cultivation, or the sediment inadvertently mobilized down the drainage. Many traditional farming techniques involve rearranging earth and stones, be it in the process of building terraces, digging canals and ditches, or adding organic and inorganic “amendments” to improve soil quality.13 Therefore, many methods that earth scientists have developed in order to assess the age and origin of natural strata can be adapted to assess the age, origin, and function of agricultural features. A more general virtue of earth science is that it in- culcates the habit of viewing the world at different spatial and temporal scales and seeking more processual or more historical explanations, depending on scale.14 This proves important in understanding agriculture, where attention shifts from the study of soil attributes in a thin-section slide observed under a microscope (see Chapter 5), to the economic interplay of regions growing different kinds of crops (see Chapter 9). Toward the higher orders of the spatial scale, the study of past agricultural systems has been taken up by another branch of archaeology, usually labeled by its practitioners as “landscape archaeology.” If we take again the narrowest, most technical definition of landscape archaeology, it is the study of patterns visible on the surface of the Earth that result from past human activity.15 In contravention of traditional archaeological routine, the landscape archaeologist focuses attention less on habitation sites and more on the fields, pastures, canals, and roads that separate or unite them. Here the routine will often consist of examining pictures taken from airplanes and satellites, as well as old maps, comparing them to the results of archaeological settlement surveys, and ground-checking the observed patterns at crucial locations. Different sets of active and abandoned agricultural features — the classic case being field boundaries — may be ascribed to different time periods solely on the basis of their shape and position with respect to one another and to habitation sites. Landscape archaeology rarely involves excavation, which is a precondition for much of geoarchaeology, especially where “opportunistic” exposures of stratigraphy are unavailable.16 The two complement each other in obvious ways. Wider definitions of geoarchaeology may subsume landscape archaeology, but in practice the two are rarely integrated. Each has had its visionary proponent in Mexico — landscape archaeology in the figure of Pedro Armillas (1914–1984) and geoarchaeology in that of Karl Butzer (1934–2016). But unfortunately, neither found many foot soldiers to implement his ambitious program.17 We have tried both methodological approaches at Calixtlahuaca, though we devoted more effort to geoarchaeology than to landscape archaeology. This is reflected in the organization of the book. Chapter 4 presents the substantive findings of landscape archaeology, Chapters 5 and 6 those of geoarchaeological field- and labwork. In Chapter 7, we step outside the bounds of archaeology altogether, searching for additional data on the agrarian history of Calixtlahuaca in documents, some of them published, others that we consulted in the archives of Toluca and Mexico City. We do not subscribe to the belief that the methods employed to collect empirical data predetermine the theoretical framework in which they will be used. With Introduction the partial exception of Smith, we gave little thought to where the project would lead us in terms of theory building. For both Armillas and Butzer, the study of agricultural landscapes provided, almost inevitably, the building blocks for a theoretical edifice known as cultural ecology or human ecology. Because of our intellectual parentage (the books we have read and the people we have rubbed shoulders with), and because we are concerned with explaining the interactions between humans and the agroecosystems that they create, we speak and write — often quite unaware — the language of cultural ecology.18 We are willing to sign our names under Harold Brookfield’s intentionally nebulous declaration of theoretical intent for his book Exploring Agrodiversity. The source material used in this book is drawn from a wide range of disciplines, especially from anthropology, agricultural science, and geography. [We would add archaeology and history.] Both the anthropology and geography are of a type that in the Americas would be called cultural ecology or sometimes human ecology. This interdisciplinary field evolved in the 1950s and 1960s. . . . It suffered competition for students’ attention from successive waves of logical positivism, neo-Marxism, and postmodernism and from the inroads of a politicized environmentalist ethic that emerged strongly in the 1970s, offering a different meaning for the term ecology. Yet it has survived. Cultural ecology has continued to appeal to a minority of students and to command a wider multidisciplinary attention than some of the mainstream work in geography and anthropology. As a field, it has changed with the incorporation of a developmental perspective and adaptation to modern concern with biodiversity conservation. Its essential characteristics are a strong base in empirical research and in exploration, in depth, of the long-lived people-environment paradigm in the social and natural sciences. Necessarily, it remains eclectic, but it has a clear focus in the management by people of their biophysical resources.19 Brookfield has also been hailed as one of the founding fathers of the currently more crowded theoretical edifice known as political ecology, which stresses the social and political struggles over natural resources, including farmland.20 In Chapter 8, we analyze terraces using the concept of landesque capital, which is eminently political, while in Chapter 9 we explore land improvement and degradation in the economic 5 and historical context of Aztec and Spanish attempts to forge multiethnic empires. We are thus now learning a posteriori that, in our theoretical insouciance, we have also strayed into the field of political ecology. Historical Archaeology and the Study of Societies Shaped by Colonialism The third and most ambitious goal of this book is to provide new answers to some of the questions that have engrossed those trying to comprehend and judge the consequences of the worldwide expansion of Europeans since the fifteenth century. We thus make a foray into historical archaeology and its highly politicized debates on the ills and saving graces of colonialism. As we surveyed and excavated Calixtlahuaca, we realized that many agricultural features and associated strata at the site dated not to its era of prehispanic glory but rather to distinct moments of the last 500 years. Our project thus began to straddle the divide between prehispanic and historical archaeology. There are two well-known but contradictory definitions of the purview of historical archaeology, one focusing on literacy, the other one on colonialism. Both are relevant to what we try to achieve in this book. The first definition states that historical archaeology “studies the remains of literate societies that were capable of recording their own histories.” The second identifies it as “the archaeology of the spread of European cultures throughout the world since the fifteenth century, and their impact on and interaction with the cultures of indigenous peoples.”21 The latter is the definition that most closely approximates the practice of the field, as self-identifying historical archaeologists are found almost exclusively in former European colonies. Archaeologists studying the literate societies of ancient Mesopotamia or the Classic Maya do not use the label and usually walk different professional paths. In the Americas, the strength of the genetic and emotional ties to Europe and Africa are in fact an excellent predictor of the standing of historical archaeology. It is strongest in countries where the indigenous population has been wiped out or reduced to the margins of society, such as the United States, Argentina, or much of the Caribbean. It is consequently weak in Mexico, where the indigenous element is genetically dominant and where the Creole and Mestizo elites who have forged the modern nation-state have met remarkable success in their conscious efforts to excise the 300-year link to the Spanish metropolis from the mental map of the citizens of the Republic. The “high civilizations” of the prehispanic period are the ones that excite the imagination of the general public and of both national and foreign archaeologists. We would estimate that fewer than 5 percent CHAPTER 2 Matlatzinco A Forgotten Capital A Tale of Two Cities The Aztec Empire grew out of a constellation of hundreds of independent polities called altepetl, some of which are identified in Figure 1.1. The term is an apt Nahuatl-language metaphor for a polity that was often centered on a mountain (tepetl) and separated from its neighbors by wetlands (atl being “water”). Modern scholars like to translate altepetl as “city-state,” a concept that we ultimately owe to that of the Greek polis.2 Coming at the end of a period of widespread migrations, the thirteenth and fourteenth centuries were the heyday of independent city-states in central Mexico and a time of great ethnic and linguistic diversity (Table 2.1).3 Nahuatl was one of the dominant languages. Those who spoke it traced their mythical origin to a place called Aztlan, somewhere far to the north. Other people spoke languages belonging to the Otomanguean family, believed to have deeper roots in central Mexico. Marriage and alliance across ethnic and linguistic boundaries, however, was common — at least for the more prominent members of society. Speakers of different languages often resided side by side within the same altepetl, using Nahuatl as their lingua franca. This situation led to the development of common cultural codes shared throughout and beyond central Mexico. The material culture was also strikingly similar across the entire region. Yet — archaeologists have tried very hard and failed resoundingly to find stylistic indicators of particular languages or ethnic groups. People built houses, fashioned stone and clay, and farmed in similar ways. Variation is more readily explained by differences of class, wealth, occupation, gender, or environmental constraints, than by ethnicity.4 The term “Aztec” comes from mythical Aztlan, but as the name of a people or a polity it is largely a scholarly invention of the eighteenth century.5 It came to be used, in its more restrictive sense, to refer to the alliance of city-states led by the altepetl of Mexico-Tenochtitlan that ultimately conquered large swaths of the country we now call Mexico. In this book, however, we follow a competing tradition, which uses “Aztec” as a convenient, Matlatzinco and Toluca are two cities whose intertwined fortunes we will consider. Matlatzinco is not a place-name on any modern topographic map, nor can one find it through Google Earth. This book and a few prior publications argue that the name corresponds to a place on the outskirts of a village called San Francisco Calixtlahuaca.1 That village has the strongest claim to be the descendant community of ancient Matlatzinco, but by no means the only one. If we seek the biological descendants of its original inhabitants, we would probably find as many of them in the state of Michoacan, Mexico City, or Toluca, as in Calixtlahuaca itself. If we consider what city inherited the place that Matlatzinco once held in regional politics or trade, Toluca emerges as the unquestionable descendant. In contrast to Matlatzinco, Toluca is a name generally known to even the least educated Mexicans. With more than one million inhabitants, it is among the ten largest cities in the country and, as a measure of its clout, has its own premier league football club bearing its name. From the earliest written sources that survive to modern-day scholars, commentators have repeatedly confused, substituted, or equated Toluca and Matlatzinco. We are thus compelled to mention briefly what each of the names evoked through the centuries and explain how and why the confusion arose. In doing so, we outline the history of political and ethnic relations in the Valley of Toluca since the thirteenth century and set the stage for the presentation of fieldwork results. We realize that to start with politics, including the intrigues of Aztec emperors and Hernán Cortés, may seem an oddity in a book on “landscapes.” Yet it is indispensable for two main reasons. First, the identification of Matlatzinco with Calixtlahuaca is not well established in the literature. Second, as explained in Chapter 1, one of our aims is to judge whether and how Aztec and Spanish imperialism shaped the landscape. It is difficult to do so if we do not present the political background. 11 12 Chapter 2 Table 2.1. The changing status of languages spoken in the vicinity of Matlatzinco and Toluca. Language Matlatzinca Otomi Mazahua Nahuatl Spanish Synonyms Family Otomanguean Hñähñü Otomanguean Otomanguean Mexican, Aztec Uto-Aztecan Castilian Indo-European 1200–1476 1476–1521 1521–1650 1650–1821 1821–1910 1910–Present Dominant Codominant Codominant Secondary Marginal Extinct Secondary Secondary Secondary Secondary Marginal Marginal Marginal Marginal Marginal Marginal Marginal Marginal Secondary Codominant Codominant Codominant Secondary Extinct Secondary Codominant Dominant Dominant all-inclusive label for the cultures of central Mexico on the eve of Spanish conquest. In this way, we acknowledge that the objects of archaeological study rarely reveal what language their makers or users spoke. Thus, the Early and Late Aztec periods simply refer to certain time intervals, synonymous with the Middle (1150–1350) and Late Postclassic (1350–1520) (Table 2.2). Matlatzinco and Tollocan were two important altepetl some 50 km west of Tenochtitlan. The surrounding valley took its name from one or the other, depending on the speaker and time period. Matlatzinco (often deformed under the influence of Spanish to Matlacingo or Matalcingo) seems to have been the more common geographical descriptor for the valley well into the early Colonial period, continuing to be used long after the altepetl of Matlatzinco had morphed into the lackluster village of Calixtlahuaca. Toluca (a deformation of Tollocan) is the name by which the valley is known today, but the toponym was sometimes used to identify it as early as the sixteenth century. This is the first source of confusion regarding Matlatzinco: when speaking of the prehispanic period, it is often difficult to tell whether a particular source refers to the valley or its most important settlement (Table 2.3). The second source of confusion is that, when speaking of a person, Matlatzinca can refer not only to the inhabitants of Matlatzinco (the city, city-state, or valley) but also to a speaker of the Matlatzinca language or a member of the Matlatzinca ethnic group. Matlatzinca is a language of the Otomanguean family, spoken in the prehispanic and Colonial periods at Matlatzinco, Tollocan, and many other settlements in the valley and farther to the west. Some sources include the Matlatzinca in the roster of peoples who migrated from Aztlan, most of them unquestionably speakers of Nahuatl. Scholars have wondered whether this means that Nahuatlspeakers who settled in the valley of Matlatzinco in the course of these early migrations had assumed a Matlatzinca ethnic identity (without necessarily giving up their language) or if some non-Nahuatl speakers also ended up subscribing to the Aztlan origin myth.6 Either Table 2.2. Archaeological periodizations. Time Name Periods (Central Mexico) 1500 BC–AD150 Formative 150–650 Classic 650–950 Epiclassic 950–1150 Early Postclassic 1150–1350 Middle Postclassic = Early Aztec 1350–1520 Late Postclassic = Late Aztec 1520–1820 Colonial Since 1820 Modern Phases (Aztec-period Calixtlahuaca) 1130–1380 Dongu 1380–1450 Ninupi 1450–1530 Yata Note: Pre-1520 periods after Parsons et al. 1982; Hare and Smith 1996; Smith 2001a. Phases after Huster and Smith 2015. The 1520 and 1820 period boundaries have no support in archaeological research and are instead rounded dates of Spanish conquest and Mexican independence. interpretation provides additional, nonarchaeological support for referring to the Postclassic city and city-state of Matlatzinco as “Aztec.” It was also easy to confuse Matlatzinco and Tollocan, especially for people who wrote of them at a desk in Mexico City or Spain, because the two were proximate: one less than a two-hour walk from the other, though largely hidden from sight by a small mountain chain, for which René García Castro has coined the name Serranía de los Chimal (Figure 4.2).7 As mentioned, we identify Matlatzinco with the archaeological ruins on the outskirts of Calixtlahuaca, at the northernmost edge of the mountains in question. The majority of sixteenthcentury sources refer to Matlatzinco as the most important settlement in the valley, at least before 1475.8 Calixtlahuaca is by far the largest archaeological site of the Aztec period in the Toluca Valley and the one with the largest volume of monumental architecture.9 CHAPTER 3 What Is a Terrace? Riser, Tread, Fill, and a Few Other Definitions the flattening, no matter how small, how crude, or how purposeful.” 2 As they emphasize, the most basic function of a terrace is to reduce the gradient of the slope. But their definition has some shortcomings. We must distinguish between “flat” and “level.” A level surface is one that is contained within a single plane, even if that plane is somewhat inclined, as most terraces indeed are. “Flat” means both level and horizontal. Perfectly flat terraces are a feat of engineering that was rarely achieved in the ancient world, but, when farmers came close to it, we speak, admiringly, of “bench terraces.” 3 The second problem with Spencer and Hale’s definition is that it sets sight on only one of the two components of a terrace that cannot exist without one another. By virtue of simple geometry, if the gradient of a slope is reduced in some places, it must be steepened in others (Figure 3.1). The part with a reduced gradient becomes the “tread” and tends to be the stage for most subsequent human activity.4 The part with a steepened gradient, often faced with a stone wall, becomes the “riser.” Tread and riser are terms an architect would use to describe the horizontal and vertical faces of the steps of a staircase. The metaphor is apt in that terraces usually come in series, forming giant staircases and giving mountains a stepped appearance. But just as the treads are rarely horizontal, so are the risers rarely vertical. Conceptually, a riser is usually grouped with the tread above it. The part of the tread farthest from the riser is the “back,” the part closer to the riser is the “front.” The slope-parallel dimension is the “width” of the terrace; the slope-perpendicular dimension is its “length.” Depending on the use given to a tread, a larger or smaller surface area may have its soil turned ahead of the planting season. We refer to that part as the “cultivation surface.” We sometimes single out the very front of a terrace as the “leading edge.” This narrow strip is rarely part of the cultivation surface because turning the soil there too frequently would undermine the stability of the riser. The leading edge may be under trees or other The dictionary meanings of “terrace” refer to a variety of more or less level surfaces that rise above something else. The surfaces may be of earth (terra in Latin) or some other material. They may be man-made or natural, as in “terraces of the river Thames.” As most native English speakers now live in urban or suburban spaces, the most common everyday usage may be the one that equates “terrace” with a large balcony or porch. The terraces we are concerned with in this book are typically found in more rural places and are sometimes subsumed under the label of “agricultural terraces.” But the designation is unwieldy and somewhat misleading because some of these terraces support houses and other structures apart from or instead of crops. To put it differently, terraces are usually but not always “fields.” Where the distinction is of interest, we will distinguish between “agricultural” and “residential” terraces. Such terraces are always found on sloping land, and before we define them, we need to spell out a few basic terms used to describe slopes throughout the book. The most fundamental attribute of a slope is its angle or “gradient,” which describes how steep it is. We express it in degrees that separate it from the horizontal plane: 60° would be very steep, 3° would be described as “gentle.” There are two possible directions as you move on a slope. “Parallel” to slope means in the direction of maximum gradient. This is the line that a dropped ball would follow. “Perpendicular” to slope means in the direction of zero gradient. If you move perpendicular to slope you always stay at the same altitude. The contours of a map trace lines perpendicular to slope. When human activities modify slopes, we speak of “slope management.” 1 We define terracing as a form of slope management that permanently modifies the gradient. In the words of Spencer and Hale, “The agricultural terrace may be conceived as any artificially flattened surface on which crops are grown subsequent to 26 What Is a Terrace? 27 TREAD cultivation surface preterracing slope gradient leading edge RISE R SOIL preterracing slope deposits FILL FRONT BACK Figure 3.1. The components of a terrace. crops that are “perennial,” i.e., grow for several years and need not be tended as frequently. It may also assume a shape wholly different from the rest of the tread. The third fundamental component of a terrace is the fill, which is the earth moved from its original position during construction and retained by the riser. It is typically much looser than the earth beneath it that is still in its natural position. But some construction techniques create terraces that lack fills (see below). Those of us who have some training in earth sciences make a distinction between “sediment” and “soil” and get upset when archaeologists use the terms interchangeably.5 Sediment is earth that has been moved from one place to another, by gravity, water, wind, in a basket, cart, or any other means. Sediment is deposited and accumulates from bottom to top, often in distinguishable layers or “strata,” a fact of life that forms the basis of both geological and archaeological stratigraphy. Soil is earth in which plants grow. It forms gradually, once the surface of a sediment (or a rock) is colonized by a plant. It slowly develops a profile, a series of layers of different properties, in this case called “horizons.” Soil profiles develop from the ground surface down- ward. The fill of an agricultural terrace, for example, is by definition sediment, because it has been moved about. Under the cultivation surface of a terrace tread, we will find a soil because there are crops growing in it. As usual, the difficulty in using these two terms lies in the fact that the same handful of earth is at the same time “soil” and “sediment.” There are times when even geoarchaeologists do not wish to make the distinction and sidestep the issue by resorting to the more neutral “earth.” In field parlance, we proudly and lovingly say “dirt,” which is the last practical term we wish to introduce at this stage. The Variety and Classification of Terraces Figure 3.2 shows a sample of the bewildering variety of terraces in different parts of the world.6 Terraces vary in size, shape, construction method, actual and perceived function, the use that they are given, and their placement in the landscape. They range from obstacles of brush casually piled across the channel of a mountain stream to elaborate staircases of rice paddies in which the flow of water must be strictly controlled through a b c d e f g h Figure 3.2. Slope terraces of different morphology and purpose: (a–d) agricultural terraces; (a) irrigated bench terraces for rice with earthen risers following contours west of Banaue, Ifugao province, Philippines; (b) strip lynchets — terraces presumed to have been created gradually by plowing in the Middle Ages — near Old Settle, North Yorkshire, England; (c) olive trees on bench terraces with risers of dry-laid stone near Valldemossa, Mallorca, Spain; (d) vineyards on terraces with wide and sloping earthen risers and narrow treads, Douro Valley, Portugal; (e–h) agricultural and residential terraces; (e) dispersed houses among vineyards on sloping-tread terraces with stone risers, Unstrut Valley, Sachsen-Anhalt, Germany; ( f ) dispersed houses among irrigated bench terraces for rice, Bàn Phùng, northern Vietnam; (g) clustered houses on Inca bench terraces with stone risers and spillways at Intipata, Peru; (h) clustered houses on rice terraces at Bangaan, Ifugao, Philippines. What Is a Terrace? feats of hydraulic engineering and complex organizational arrangements among farmers. Many terraces are actively used, but many others lie abandoned. Modern, highly mechanized agriculture favors large expanses of flat land. Where these are in short supply, it prefers slopes that pose few problems for the movement of tractors and other wheeled vehicles. This means slopes that are relatively gentle and continuous, i.e., not broken into multiple parcels by terrace risers. This is why terraces are something that we usually associate with “traditional” agriculture that relied heavily on manual labor. But “traditional” does not mean ancient, and, as we shall see, abandoned terraces that are still in good shape are unlikely to be older than a few centuries. Any of the attributes just mentioned can give rise to multiple classifications of terraces. For the purposes of this book, we need to look only at a few of the simpler classifications and limit our discussion to those categories that are common in highland Mexico. We have already made the distinction between agricultural and residential, which are two overlapping categories defined by use. A simple but widely used classification concentrates on the finished form of the terrace, and its placement in the landscape.7 It distinguishes (1) cross-channel terraces; (2) sloping-field terraces; and (3) bench terraces (Figures 3.3, 3.4). Cross-channel terraces are created by placing an obstacle across the dry channel of a seasonal stream and letting sediment pile up behind it after rainfall events. In this case, the gradient that the farmer reduces is the longitudinal gradient of the stream, striving to slow the flow of water. It is usually impossible to retain all the water within the terrace fill, and a sluicegate or some other allowance must be made for letting the excess water escape downstream. Simple cross-channel terraces require relatively little effort or technological know-how to build and are among the oldest terraces that we know in Mesoamerica, appearing some 3,500 years ago.8 Cross-channel terraces are also found in small, steep ravines or similarly concave sectors of a hillside. Such places naturally concentrate the flow of water — we speak of concave or “water-gathering” slopes — but rarely witness channelized flow and do not develop the morphologies typical of stream channels. Instead, there may simply be water seeping into a grass-covered depression whose long axis is parallel to the slope of the hill. The other two categories are typically found in the straight or convex (“water-shedding”) sectors of a hill. Sloping-field terraces have a tread that is far from flat and often only a little less inclined than the natural preterracing gradient. A form that is particularly common in central Mexico consists of sloping treads separated by risers that take the form of earthen banks (“berms”).9 More often than not, a ditch runs alongside the berm, 29 both parallel to slope and closely following the contour. The berm is the spoil from digging the ditch. The ditch is usually upslope of the berm, on rarer occasions downslope of it. One of the names for this type of terrace is metepantle. The first member of this word comes from metl, the Nahuatl for maguey (the century plant, Agave spp.), which is the crop most commonly planted on top of the berm. From a distance, this presents the picture of neat, slope-perpendicular rows of magueyes. In aerial photos, it is these rows that most readily allow one to note the presence of terracing (Figure 3.3b). Magueyes may be interspersed with or replaced by cacti, in particular nopal (prickly pear, Opuntia spp.), fruit trees, or other useful perennials. In this form of terracing, then, the riser also supports crops, though usually different from those grown on the tread. Bench terraces have riser treads that are nearly flat. Where irrigation waters are distributed over the treads, the terraces usually take this form because it allows close flow management. However, dry-farmed terraces — those that rely exclusively on the rain that hits the surface of the tread — can be bench or sloping-field varieties. In central Mexico, bench terraces are often retained by a wall of unfaced field stones that leans toward the back, interlocking with the fill of the terrace. The wall may be laid dry or with a mud mortar made by wetting the same earth that is used as fill. Magueyes or other perennials are planted along the leading edge of the tread, just as they would be along the berm of a metepantle. In both cases, the plants may be spaced at regular intervals, leaving the intervening spaces empty and assuring that sufficient light reaches the crops, a condition critical for the optimum development of sun-loving succulents such as maguey or nopal. Where their growth is not the major concern of the terrace owner, or when the terrace is abandoned, the perennials propagate spontaneously to fill the empty spaces, together with a range of shrubby weeds, to form veritable hedgerows or live fences along the riser and leading edge of the terrace. How to Make a Terrace: Construction Techniques and Their Effect on Stratigraphy Yet another way to classify terraces is by focusing not so much on their finished form but on the techniques used to create them. Terrace construction perforce requires earth to be shifted about on a slope. This may be achieved with earth-moving machinery, by carrying or throwing earth manually from one place to another, or by encouraging natural processes of sediment transport, by gravity or water. Agricultural terrace use always involves tillage. This term encompasses plowing but also the turning and breaking up of soil with harrow, hoe, digging stick, or another tool. Before the arrival of 32 Chapter 3 topsoil (A horizon) retaining wall subsoil (B horizon) hand-filled self-filled maguey row sediment transfer a. natural slope b. cut terraces c. cut-and-fill terraces d. hand-filled terraces e. self-filling terraces f. ditch-and-berm terraces (metepantles) Figure 3.5. Different construction techniques and their effect on terrace stratigraphy. the Spanish, there were no draft animals in Mexico and consequently no plows. In any given season, tillage may affect the entire surface of a field or only a part, creating furrows and ridges, isolated pockets or mounds, or other spatial configurations. Terrace construction and tillage may seem like two separate processes, but they both involve shifting earth about and in practice sometimes merge into one another.10 Both have profound effects on the stratigraphy of a slope.11 We may start out on an imaginary slope with a soil cover consisting of two horizons (Figure 3.5a). The topsoil, or A horizon, is the one that supports the highest What Is a Terrace? 35 People build terraces in order to: FUNCTION ● Retain water PURPOSE ● Support houses MOTIVE ● Feed extra members of the household REASON ● Respond to population pressure ● Retain nutrients and other soil amendments ● Support other structures ● Pay tax or contribute to a feast ● Respond to social or political pressures ● Increase soil thickness ● Accelerate development ● Support an infield of soil fauna ● Support an outfield ● Accelerate development of favorable soil structure ● Delimit property ● Support a garden ● Store food for years of bad harvests ● Manage risk ● Prevent soil erosion ● Pass property on to heirs ● Reduce incidence of frost ● Make settlement defensible ● Reclaim degraded land ● Make a profit in the marketplace ● Respond to market demand ● Invest in landesque capital Farmers Agronomists (Physical) geographers Landlords Rural sociologists (Cultural) geographers Cultural anthropologists Historians Archaeologists Figure 3.6. “Why do people build terraces?” The question and the answers different groups of people are likely to give. notions of what agriculture is or should be — rarely voice in a direct manner. They are more likely to complain about sterile subsoil surfacing in their plot, a sheet of sterile sand that covered it after a downpour, or a new gully that makes it difficult to move about. Such complaints, however, pale into insignificance in the face of the constant worry that the crops may not get sufficient water to mature. The primary function of an agricultural terrace thus seems to be to hold water, halting or substantially slowing its movement downslope. This is not only the farmers’ appreciation: field surveys of terraces in the Americas have found that their distribution matches closely that of permanently or periodically water-stressed environments.17 The second-most important function may be that of thickening the soil profile, particularly the part that is most actively used by plants. Natural soil A horizons are typically very thin on slopes. They will be substantially thicker under terrace treads, albeit only in slope-perpendicular strips of a certain width behind the terrace riser (see preceding section). The other functional advantages of terraces are to a large extent a by-product of the ability to retain water and increase soil thickness. An added benefit of terracing may be to reduce the incidence of frosts. Terraces may induce wind turbulence that will prevent cold air descending to the level where the crops are growing, while stone risers — especially those of dark-colored volcanic rock — absorb heat during the day and radiate it during the night.18 “Purpose” refers to the intended use of the terrace: to have it support a house (Figure 3.7), shed, or threshing floor, or to grow crops. This is perhaps the question of most immediate concern to the farmer. An agricultural purpose may be subdivided depending on the intent to treat the terrace tread as a garden, infield, or outfield. A garden implies the shared residential and agricultural use of a tread or flight of contiguous treads. Members of the household use it to grow small quantities of plants that are highly valued but do not necessarily contribute much to the diet in terms of the intake of calories or crucial proteins — picture a housewife dashing from her kitchen to fetch a condiment she has forgotten while cooking a meal. The plants may include ornamentals and the garden may be aesthetically pleasing. A field, in contrast, is the place to grow staple crops. The infield is still close to the house and, like a garden, benefits from daily oversight and inputs of kitchen refuse and human and animal waste.19 The outfield is farther away, and the farmer will tend it only at crucial points during the growth cycle of the crop. Yet another purpose of building What Is a Terrace? 39 A Tillage furrows conduct runoff perpendicular to slope and toward tread edges unprotected by side walls. G Neglected slope-parallel terrace drain widens into ravine. B Sheet erosion exposes peds of the hard preterracing substrate at the back of the tread, where the terrace had been cut into the slope and topsoil was always thin. H Slope-parallel access way collects runoff and turns into ravine. C Neglected riser collapses creating a cone of rubble and redeposited terrace fill. I Maguey plants are allowed to flower instead of being castrated and exploited for their sap or hearts. Their dry stumps are removed by people collecting fuel. D Collapse cone locally steepens slope gradient, accelerates runoff, and initiates gullying. J People engaged in construction projects elsewhere begin to purposefully rob stone from risers and stockpile it before taking it away. K In the absence of people guarding their crops and of transitable access ways, livestock graze on the grassy treads and climb over risers, dislodging stones. E Gully breaches terrace riser and connects to ravine. F Piping of terrace fill undermines and breaches riser. STRATIGRAPHY EXPOSED IN SCARPS: A Friable terrace fill C B Hard preterracing substrate D H F E G I K J Figure 3.9. The disintegration of a terraced slope. hold out may take to robbing stone and borrowing fill (Figure 3.9, point J), encouraging erosion upslope of their own plots, and grazing their animals on treads where crops are no longer grown (point K). The result is a proliferation of weak points where terraces fail or are willfully destroyed. In order to discern which causal relationships could plausibly have operated in the past, it is advisable to study both the vestiges of terracing and the deposits that may be the result of their collapse, and to date each independently. It is also a good idea to look at long sequences of agricultural change, rather than concentrating on one’s own favorite period. We have tried to follow these recommendations at Calixtlahuaca. CHAPTER 4 The Landscape Archaeology of Cerro Tenismo Archaeology Done with Your Feet Landscape archaeologists like to speak of “palimpsests,” likening the landscape to an ancient manuscript on parchment, from which generations of scribes erased parts of their predecessors’ writing while making room for their own texts. As with the different handwritings in a manuscript, it is often possible to tell which features of the landscape came first simply by observing their style, degree of completeness, state of preservation, and recording which crosscuts or wraps around which. In some cases, judicious observation may obviate the need for expensive excavation or reduce its scope to surgical interventions at crucial locations. The Mexican countryside may not have enjoyed as much peace and respect for fence lines, but the same basic principles should apply to any landscape shaped by agricultural activities. Pedro Armillas, an exile of the Spanish Civil War, was the first to try them in the radically new surroundings in which he disembarked in 1939. As an artillery officer, he possessed the rudiments of topography and on arrival was tasked with surveying land parceled out to Indians in remote parts of southern Mexico. This background influenced him to take a radical departure from the monument-centered mindset prevalent in Mexican archaeology and to pioneer the use of aerial photographs and off-site pedestrian survey. By the 1960s, he had steeped himself in the writings of British landscape archaeologists and applied them to reading Aztec and more recent imprints on the wetlands and slopes of the Basin of Mexico. He published his wetland observations in Science, a highly visible venue, while relegating those on slopes to verbal communications.2 One of Armillas’s adages, recalled by his students, was that “archaeology is done with the feet, . . .walking” (in Spanish, to do something with your feet — hacer algo con los pies — means to bungle or make a mess of it).3 Pedestrian survey became indeed the bread and butter of Mesoamerican archaeology.4 The British- or Armillas-style landscape archaeology, on the other As one looks up the hillside above modern Calixtlahuaca, past the flat roofs of ugly cinder block houses with their jungle of cables, clotheslines, and plastic water tanks, the eye meets the neat outlines of buildings that dwarf the houses in size (frontispiece and Figure 4.1). The cut stone facing the sides of the pyramids — because this is what they are — is of a shade of gray similar to cinder block, but one cannot deny them monumentality. Between them and farther upslope, the hillside is dotted with thousands of maguey plants and, late in the rainy season, pasted dark green with maturing maize. Cinder block houses, pyramids, and agricultural fields are three rather incongruous elements of the same landscape. Pointing out that they do not belong together may seem an aesthetic value judgment and not a proper scholarly statement. As we will try to show in this chapter, however, perceiving such incongruence is often a necessary first step in formulating hypotheses about the past of a particular landscape. They lie, in fact, at the foundation of “landscape archaeology,” a scholarly subdiscipline developed in its earliest and, so far, fullest form in Great Britain. Because no army has thoroughly burned and plundered the British countryside in about a thousand years, and because woodland has been scarce for at least two, the remains of different time periods — house foundations, fences, animal pens — have piled up instead of being erased and are also readily visible from the air. When photographs taken from airplanes and balloons became widely available as a by-product of military operations in two world wars, people examining them started to note roads that appeared to lead nowhere, strange kinks in otherwise straight hedgerows, and patterns of vegetation that did not respect field boundaries. Fieldwalking, searching historical maps and local archives, and in the end excavation, often confirmed them to be the result of the superposition of human activities of different time periods and allowed each to be dated.1 40 a b c Figure 4.4. Cerro Tenismo and surroundings: (a) aerial photo, May 1999; (b) major subdivisions of the landscape — see also thick white lines in (a); (c) transect from summit to valley floor. The Landscape Archaeology of Cerro Tenismo 47 unclassified FLO PLA5 CIN PLA4 TRA BEN PLA3 BAD MAG PLA2 QUA FOR SUM STE PLA1 TEC FAN unclassified unclassified 100m Figure 4.5. Landscape units. Compare to aerial photo, Figure 4.4. For code explanation see Table 4.1. The intermediate gradient of the Lower Slope would confer it characteristics intermediate to those of the Upper and Middle Slopes. The Valley Floor would be the locus of most intense deposition, fed by the wasting of Cerro Tenismo at all altitudes, as well as by overbank flooding of the Río Tejalpa and its tributaries. The aerial photo, however, brings out relatively few neat patterns that respect this coarse subdivision. In keeping with the prediction of thin soil cover, there are indeed several outcrops of basaltic bedrock on the Upper Slope, within the STE, MAG, and BEN units. Most are curvilinear ridges roughly perpendicular to the general direction of slope. They stand out as sets of dark and often overlapping arcuate lines, with their convex side always facing downslope. This reflects the morphology of the different lobes of lava fanning out from the center of the volcano. Because there is no soil to support tree growth, the black patches that elsewhere mark tree crowns and their shadows are notably absent. But even the STE unit, defined largely on the basis of its steepness and lack of obvious man-made features, lacks the hallmarks that usually identify places of active erosion such as the very light colors indicative of recently exposed subsoil. Instead, much of it is under grassland, which can be attributed to a combination of soil conditions and grazing pressure. Moreover, the only 571b 575 20m 571f 571b 573 499c 571c 571d 572 551 499a 430 499b 429 583a 583c 583d 445a 529 583f 583b 321 583e 445b 528 446a 446b 448b 447 448a 449 448c 543 544 N 525 445c The Landscape Archaeology of Cerro Tenismo robbed from a wall that originally stood on top of the scarp such that the original riser was taller; or (3) the terrace was built by cut and no fill methods (see Chapter 3) and was never meant to have a stone wall. Cutting back the scarp and exposing its stratigraphy is often our only chance to falsify some of these hypotheses.24 Side walls are decidedly less common than riser walls and more frequently discontinuous. Many are rather flimsy and do not seem to give much structural support to the terrace fill. The back-of-tread ditches are one of the most original features of the terracing system of Cerro Tenismo and deserve careful consideration. At first sight, we thought they could be part of a network of irrigation canals that had fallen into disrepair. Some of the most sophisticated bench terraces in the world are fully integrated with irrigation networks that allow water to flow from one tread to the next (Figure 3.2a, f, h). In the Basin of Mexico, on the north slope of Cerro Gordo, William Sanders and Thomas Charlton described terraces in use in the 1960s characterized by ditches laid out in the same manner at the back of each tread.25 Water was diverted from a slope-parallel shallow canal to enter each tread at a back corner. The diversion was achieved by building the sides of each terrace in the form of earthen banks projecting into the canal. The ditch served to break the force of the water and to prevent gullying. The treads are described as flat and “self-flooding,” but neither Sanders nor Charlton specify how water was distributed from the ditch over the cultivated surface of the tread. The similarity of form is striking, but there are several problems with viewing the ditches on Tenismo as irrigation canals. At the back corners, the mouth of the 67 ditch is often suspended at some height above the bottom of the slope-parallel drains or gullies. Water would thus tend to exit rather than enter the terrace. It is of course imaginable that in the past the drains were shallower or that water was impounded with earth, brush, or planks so that it rose to the level of the ditch. In their present form, many ditches are partially silted up, discontinuous, and even disconnected from the drains. All of this could be blamed on the lack of maintenance under the current regime of intermittent cultivation. The features most incongruous with irrigation are the low berms, lines of stones or rows of magueyes bordering the ditch, for they obstruct the spread of water over the cultivation surface, whether this distribution is done by surface flow or splashing the water with coas (prehispanic hoes), basins, ladles, or scoops. The only way it may reach the crops is perhaps by gradually seeping through the terrace fill. Conversations with farmers also suggest that their main concern is to keep potentially damaging channelized flow out of their terraces. Given the subhumid climate of the Toluca Valley, the possibility of insufficient water supply seems to be of less concern, and the distribution of ditches substantiates this outlook. They are most common on the steep Upper Slope where the erosive power of runoff is high. But we did not document any on the narrow treads of the FOR unit, which is close to the summit and therefore has a very reduced catchment. The practice of leaving ridgeand-furrow in place for the duration of the fallow may also be a way of stemming slope-parallel flow over the surface of the tread. In conclusion, we are led to believe that farmers dig these ditches to channel away an excess of water that Figure 4.12. (opposite) Landscape sketch of the surroundings of El Panteón and Location 321. Observations: Field 429 is the platform supporting the Aztec architectural complex known as El Panteón. The ditches in that field are very recent creations of the INAH guardians of the archaeological zone. The platform obstructs the straight course of the Panteón Drain, which swerves sharply at the corner of Field 572 and is redirected along the western edge of the platform toward a different concave sector of the slope. If the drain were to continue on a straight course, it would follow, farther downslope, the boundary separating Fields 529, 528, 448a, and 543 from Fields 445a–c and 448b–c. The upper part of that boundary is marked instead by stone walls and an earth-and-rubble berm. Water seeps to the surface, however, within Field 448, and a minor drain makes its appearance at the downslope end of the seep. It is possible that this set of field boundaries marks the original, straight course of the major drain, diverted to the west only once the Panteón was built. As in the previous sketch, there are several wide treads likely merged from narrower treads of an older generation, some still preserving topographical irregularities along the suspected “seam” (e.g., 551, 529, 446a). The large “bulges” at the back of Fields 430 and 551 may be the remnants of Aztec mounds, the result of tread mergers, or terrace fill slumped after massive riser failures. The large gap in the retaining wall of Field 572 suggests a riser failure played a role in forming the bulge in Field 430. In a process also referred to in observations on the previous sketch, some treads (e.g., 572, 583d–e, 446a) have inherited concave sides from a once wider water conduit. In Fields 448b and 447, a different solution was adopted: the field located in the concave sector of the original slope (447) was not filled in but separated from the convex sector (448b) by a continuous wall. The water seep in Fields 447, 448c, and 544 is conditioned not only by the concavity of the slope but also by the presence of a bedrock outcrop. CHAPTER 5 The Stratigraphy of Terrace Construction Techniques of Excavation and Stratigraphic Analysis In our field nomenclature, a “location” is simply one of the places where we dug, contained within the same field or set of contiguous fields, aligned on the same local grid, and tied to the same local benchmarks.2 Other archaeologists would call this an “operation” or an “excavation area.” In each location, we excavated one or several rectangular blocks, some contiguous, others not. We call each of these a “locus.” Other archaeologists would call this an “excavation unit.” Some architectural features, middens, or sets of samples were separated out as additional loci. Finally, a “lot” is any volume of earth that we removed separately in the course of excavation, recording its three-dimensional shape with respect to the grid and benchmarks. The lot is thus our basic provenience unit. Additional lot numbers refer to collections made on the surface, usually ahead of the excavation or removed from sections, usually at the end of excavating a particular location. The location-locus-lot system is reflected in all the labels that we use to refer to excavated contexts. Single three-digit numbers from 303 to 329 refer to locations; two-part numbers (e.g., 326-2) refer to loci; three-part numbers (e.g., 326-2-4) to lots. In the “terrace” excavations, we usually laid out the initial loci to form a trench 1.5 m wide. When targeting terrace fills in convex sectors, the trench would be perpendicular to the riser and preferably spanning the width of at least one complete tread. This is advisable because the stratigraphy and soil properties at the front and the back of a terrace tread are always radically different. It is also of great advantage to be able to dismantle the stretch of riser found in the path of such a trench in order to examine any differences in stonework and other stratigraphic details. Dismantling stone walls is obviously a thorny issue with most landowners and often requires either protracted negotiation or a “hitand-run” excavation. When targeting water conduits or terrace fills in concave sectors, the trench would normally be perpendicular to the conduit and hillslope, as this is the axis We excavated in 27 locations on and around Cerro Tenismo, numbered 303 to 329 (Figures 5.1, 5.2). Borejsza and Frederick selected seven of these with a view to exploring the stratigraphy of terraces and water conduits. Smith and other project members selected the remaining 20, hoping to find buried Aztec houses and to excavate each in its entirety. We thus made an initial distinction between stratigraphic and extensive, or between terrace and house excavations, with Borejsza in charge of the former, Smith of the latter, each employing somewhat different excavation and recording techniques. In the end, the distinction became blurred, as most house excavations also took place on terraces and exposed interesting stratigraphic sequences, several of which Borejsza and Frederick were called on to record. Conversely, terrace excavations exposed some vestiges of house architecture and burials. Houses were present in 11 locations, but we managed to explore extensively only 9. In the end, we excavated more than 900 m2 and recorded a total horizontal length of some 500 m of stratigraphic sections.1 We present only a selection of these in this chapter and the next. Here, we focus on the locations that turned out to be most useful for understanding the construction techniques and stratigraphy of terraces. Four are located in convex sectors of Cerro Tenismo; two span both convex and concave. In Chapter 6, we turn to locations on or near the water conduits. Four are in part or in whole in concave sectors, with two more on or near the alluvial fan of the Palace Drain. Because much of the sediment in these locations is derived from terrace fill and the former soil cover of the mountain, the stratigraphic sequences presented in Chapter 6 shed light on the processes of terrace destruction as well as the changing form of the water conduits. We occasionally refer to things that we discovered at the remaining 15 locations but do not present them in any detail. 85 Figure 5.1. Monumental buildings excavated or identified by García Payón (s1 to s17 and cr. = Cruciform Structure), locations excavated by our project (303 to 329), and proposed master drains (white lines and inset). (e) (a) (b) (f) (c) (g) (d) Figure 6.10. Reconstructive drawing of water conduit evolution in Location 326: (a) natural concave sector of slope; (b) erosion of topsoil and scouring of ravine; (c) infilling of ravine with household trash (Dongu phase); (d) infilling completed, digging of drain, terrace construction, cultivation (Dongu through Yata phases?); (e) abandonment, erosion of terrace tread, silting-up of drain (early Colonial?); ( f ) ravine deepening and widening (Colonial?); ( g) redigging of drain, reconstruction of terrace, and renewed cultivation (Modern period). CHAPTER 7 Documentary Glimpses and Guesses The Promises and Limitations of Documentary Sources Archival sources on the Toluca Valley are vast, varied, and still partially unexplored, especially from the standpoint of historical archaeology. Because of its proximity to Mexico City, agricultural wealth, and position on routes to several mining centers, the valley drew the sustained attention of royal officials from the 1520s onward. Their 300-year-long legal war of attrition with the Marquesado’s own army of lawyers and accountants has left a paper trail of tens of thousands of pages scattered through the archives of Mexico, Spain, and the United States. From the moment of the villa’s foundation, notaries took up residence and recorded an astonishing amount of the private dealings of Toluca’s Spanish inhabitants. Handed down from one generation of notaries to the next, much of this record has survived. The historians who set about cataloging it in 1978 estimated in 2012 that they had looked at less than 4 percent of approximately 550,000 deeds now held in a public archive.1 To the dismay of several sixteenth-century commentators, the Indians also readily embraced the Iberian culture of litigation, setting down their disagreements, as well as their noncontentious legal proceedings in both Spanish and Nahuatl. In the 1820s, the arrival of the executive, judiciary, and legislative powers of the State of Mexico, as well as the founding of the Instituto Científico Literario2 contributed to the growing torrent of paperwork flowing from Toluca. The historiography of the Toluca Valley has risen to the challenge, especially in the last few decades. Toluca now has two institutions of higher education with exceptionally large and productive history departments.3 In addition, the PRI oligarchs have an established tradition of patronage of “cultural” endeavors, and their largesse has extended to the publication of works on regional history, facsimiles and transcriptions of docu- ments, as well as the compilation of exhaustive archival indices that few other states can boast.4 The compatibility of the historical and archaeological records is a less rosy matter. The first problem facing anybody wishing to reconstruct landscape history is that the bulk of the documentary sources deals with land tenure and land-based taxation, not land use. It is common to read reams of litigation records without ever discovering what the owner actually grew or grazed on the land under dispute, let alone how he went about it.5 As should be clear from the preceding three chapters, archaeological exploration is the exact opposite, giving us much more immediate insights into land use than land tenure, though some aspects of the latter can be inferred from well-preserved boundary markers — terrace risers being a good example. The second problem is that many fields and tracts of land never turn up in the historical record. This is especially true of land under stable ownership or land judged to be of little value. As we shall see, Cerro Tenismo unfortunately met one or both of these criteria for protracted periods of time. It is easier to study broad regional patterns in multiple villages over a century or more than to focus on a single village or to reconstruct the regional structure of landholdings at a particular point in time.6 The scope of archaeological fieldwork, even on a relatively large and well-funded project such as ours, is usually a single tract of land within the larger holdings of a village. With regard to excavation, our results are pegged to subdivisions of certain fields and can be extrapolated to the entire tract of land only after a judicious consideration of sampling biases. A microhistory fully compatible with archaeology would thus have to descend to the level of specific tracts of land and fields, a quixotic endeavor for most places in most time periods. When individual properties — typically made up of several cultivated fields or grazing enclosures — appear in the historical record, they are usually 158 Documentary Glimpses and Guesses situated (1) by referring to major natural or man-made features (e.g., “on a hillock,” “by the road”); (2) by listing a toponym that refers to a tract of land (e.g., “La Cumbre Chica,” “Banco Nova”); or (3) by naming the owners of neighboring plots. The last format is the most common, from sixteenth-century-land grants to twentiethcentury bills of sale. Accompanying plans are rare and schematic. It is thus nearly impossible to pinpoint the location of the fields. Our first solution has been to make the best of any glimpse of Calixtlahuaca that we could catch in the sources. Even where land was not the main subject, we read on, hoping to form an idea of what kind of village Calixtlahuaca was at a particular point in time. There are enough such glimpses to reconstruct, for example, the major demographic ups and downs (Table 2.4, Figure 2.2), which were obviously a major factor in decisions regarding land use. Glimpses of Cerro Tenismo itself are exceedingly rare until the twentieth century, but documents about neighboring tracts of land are crucial in documenting the choices made by all the actors involved in land exploitation in and around the village. Our second solution has been to widen the scope of the historical inquiry to the lands of several other villages in the vicinity, paying special attention to those situated in the Serranía de los Chimal (Figure 4.2), which grappled with similar natural and social constraints as Calixtlahuaca. This allows us to make reasoned guesses as to what was likely to go on with land that belonged to Calixtlahuaca, even where specific documentary references are not available. Extrapolating from the general to the specific case is a probabilistic line of reasoning, open to error. However, we think it is legitimate and fruitful as long as we remain scrupulously honest as to what is a “glimpse” and what is a “guess,” critically weighing the extent to which the latter is likely to apply to our case. Another kind of extrapolation that we engage in leads back and forth between land tenure and land use. In this, we follow some regularities that are historically specific. For example, when reading about an eighteenth-century private estate that requested water rights, we may affirm that it did so because the owners wanted to sow winter wheat on the flat valley bottom. Even though the latter is not explicitly stated in the document, the argument is plausible because there are numerous cases in which all parts of the equation, including the crops sown, are spelled out in documents from highland regions of New Spain. Other regularities are cross-cultural and rooted in a body of theory in economic anthropology that establishes correlations between a “gradient of land use” and a “spectrum of land tenure.” 7 As Robert Netting puts it, 159 Though we are accustomed to thinking of land tenure as a set of jural concepts or legal rules externally formulated and enforced by political bodies, I shall examine property rights here as part of a local agroecosystem, testing the hypothesis that, other things being equal, land use by and large determines land tenure.8 This is a very far-going proposition and one that lawyers and traditional historians may frown upon. But it has enormous appeal for archaeology, given how difficult it is to study land tenure by archaeological means. The appeal is greatest for time intervals such as the prehispanic period in the Toluca Valley, from which no written records survive. Aztec Agriculture and the Burdens of Empire, 1200–1550 One of Netting’s most cherished predictions is that in densely populated areas permanently cleared for laborintensive agriculture, we shall find a class of peasant smallholders with strong private rights of use (usufruct rights, dominio útil). They are arranged in a two-tiered system with rights of transfer and administration (residual rights, dominio eminente) held by the elites or certain corporate groups. The residual right to reallocate vacant land is vigorously exercised and ensures the continued maintenance of an intensive agroecosystem. What we know of Aztec land tenure on the basis of sixteenth century sources provides a perfect fit for this model. By and large, land under houses and crops was privately held, managed, and inherited by both commoners and nobles.9 The lease and sale of land were accepted forms of transfer of property rights. Residual rights rested either with certain noble households or with the calpolli, the corporate group that was the basic building block of the altepetl and consisted of the residents of a certain neighborhood.10 Nahuatl had a bewilderingly complex vocabulary that described the different tiers of property rights allowing certain natural or artificial persons to farm, transfer, or tax land. Aztec pictorial writing reached the pinnacle of its sophistication in cadastral registers and maps that demarcated irregularly shaped fields, described soil quality, and identified the owners and the kind of claims they held.11 As far as we can tell from sixteenth-century and later custom, property rights to maguey plants were conceived as separate from the rights to land on which they grew.12 We are extraordinarily lucky in that one of the most cited narrative passages describing Aztec land holdings originated in depositions made by informants who spoke of the lands of Matlatzinco and Tollocan.13 The origin of the passage is rarely noted, perhaps because it Documentary Glimpses and Guesses extent of archaeological site COMMON LANDMARKS (see Table 7.2): 171 “cerca, cerquilla” “arroyo, barranca, barranquilla” “cerrillo” Juan hacienda boundaries and owners de Sojo “calle, camino, senda, vereda” “mojón” Tecaxic Indian villages “casa” “zanja, zanjilla” contours every 10m D Al iego ba d rr e án 1 km . .M . R A Ayacac (deserted) ig.A. Matías Alonso de Figueroa Cuetlaxticpac Juan Fernández Maldonado (2) pm. m.F. ig.Cx. Calixtlahuaca cs.F. cs.G. cs.1 A. Álvaro de Sandoval Tecaxic Diego García de Figueroa ig.Tx. . C.Cx. tej. C.Tx. ig.C. P.Tx. C.Ms. z. C.H. Alonso Sa lvador (1) Sa r( do . lva m 2) B. cs.M. González cr. Domingo Juan de Sojo cm. pd.g. R.M pn. . C.A. Azcapotzalco ig.S. C.Tl. ig.Tl. Tlaxomulco .S cs so on m.M. Francisco de Cañas Tepeitic (deserted) P.Ms. Miltepec C.Mo. Al t.g. ig.Tp. pt. t.p. .Pb Cx cs.2 Guadarrama Francisco de Fuentes & Isabel González ig.Th. Tlahuililpan (deserted) Cx.Pb. Bartolomé C.Mn. Huexopan (Otompan) Maldonado (3) Juan Fernández Maldonado (1) Tlantzinco (deserted) Oxtotitlan B. Antonio Sánchez VILLA DE TOLUCA Figure 7.2. Haciendas and landmarks in the vicinity of Calixtlahuaca in the Villavicencio papers of 1635–1636; based primarily on AGN-HJ, vol. 15, exp. 1. Compare to Figures 7.3, 7.4, and Tables 7.2, 7.3. land to Spaniards in the Toluca Valley but largely outside the Marquesado, with only a few estates on the Sabana Grande encroaching onto the Marquis’s domain. It was the Fourth Marquis, Pedro Cortés (1602–1629), who seized on the idea of emulating the viceroy and raising revenue by issuing grants (mercedes) of land vacated by the epidemics. A convenient way to chart the progress of Spanish landholding around Calixtlahuaca is to start with the exceptionally thorough and well-documented inquiries of the oidor (royal judge) Agustín de Villavicencio into landholding around Toluca in 1635–1636. These allow us to take a snapshot of the extent of different properties at this particular point in time. We will then follow judge Villavicencio in inquiring into the origins of the estates and finally ask what became of them after 1636. Figure 7.5. Map showing haciendas and villages irrigated from the Río Tejalpa in 1757. AGN-Tierras, vol. 2476, exp.1, f. 260. Photo courtesy of the Archivo General de la Nación. CHAPTER 9 Legends Black and White Aztec Agriculture and Its Transformation The Need for a Wider View The previous chapter dissected the site-specific case of Matlatzinco-Calixtlahuaca with the aid of theoretical constructs developed mostly in economic anthropology and economic geography. The focus was on certain regularities of human behavior that recur in different cultures and at different points in time. In this chapter, we shift attention to longer diachronic and wider regional reconstructions and debates that have engrossed an only partially overlapping circle of scholars. Its protagonists have for the most part been anthropological archaeologists with an evolutionist bend and an interest in the Aztecs, as well as historical geographers and environmental historians with a research interest in the “Columbian encounter.” In a way, we are simply picking up loose threads from the previous chapter. The Achilles’ heel of the arguments developed there was our inability to fully assess the options available to farmers in the immediate vicinity, a problem shared by most site-specific archaeological studies of agriculture.1 Within an hour’s walk from Cerro Tenismo, there are other places that could presumably be farmed by people who lived at the site. Some are similar, including the slopes of the Serranía beyond Tenismo or of cinder cones (small volcanoes) to the north and west. Other nearby places are ecologically very different, including the floodplain of the Tejalpa river and the plain that stretches to the north on the river’s left bank. Today, the latter is dotted with many small reservoirs, and we imagine that in Aztec times there were some natural wetlands in this area. Colonial sources analyzed in Chapter 7 repeatedly mention traces of prehispanic cultivation ridges on the valley floor between the Tejalpa and Lerma rivers, as well as in the Serranía west of Cerro Tenismo. There were thus many places in the vicinity susceptible to agricultural reclamation and intensification. In this regard, we know that in other parts of central Mexico the Aztecs diverted water courses for irrigation and transformed wetlands into systems of drained or raised fields. We have no archaeological evidence for such works anywhere in the Toluca Valley, but it seems almost unimaginable that they should not exist, especially given the reputation for bountiful harvests that the valley has had since Aztec times and the existence of many valleyfloor villages with prehispanic roots.2 The problem is not only the scarcity of field research in the region, but also the extremely low visibility — on the valley floor — of archaeological remains in general and of agricultural features in particular. Fields that existed in naturally flat places, unirrigated, and unbounded by ditches or walls, are archaeologically almost invisible.3 The settlement patterns recorded for the last prehispanic period show a high density of sites in the Serranía, and almost none on the valley floor to the north of it,4 even though written sources indicate the existence of several sixteenthcentury settlements in this area including Tlahuililpan, Ayacac, and Cuetlaxticpac, all of them likely outlying barrios of Matlatzinco itself. It is thus difficult to judge the extent to which the immediate vicinity of Aztec Matlatzinco was a place crowded enough to leave some farmers no other option than to intensify production on Tenismo itself. In Chapters 3 and 4, we pointed out waterlogging, heavy soils, and frequent frosts as impediments to valley floor agriculture in Aztec times, but we do not think that they excluded its entire surface area from cultivation. We saw that farmers overcame many of these impediments in the Colonial period, thanks to new technologies, new crops, and the acquisition of grazing animals. But this is something we know from the written sources that appear at that time, not from archaeology. For the Aztec period, we must consider other archaeological sites to assess whether the agriculture practiced on Cerro Tenismo was common or exceptional. We shall draw mostly on examples from the other three major regions of the Aztec realm in central Mexico (Figure 1.1), especially the intensively researched Basin of Mexico, but we will also look to parallel developments in highland Oaxaca. 233 234 Chapter 9 The Colonial period in Latin America and European colonialism worldwide are frequently blamed for the most catastrophic degradation of landesque capital or, as some scholars prefer to put it, the destruction of indigenous agroecosystems. Other scholars refute this claim with equal vehemence, placing the most severe degradation in the prehispanic period. The history of agricultural change between the Aztec and Colonial periods is at the center of this “Columbian” debate, as anticipated in Chapter 1. The Calixtlahuaca project and other Mexican case studies published in the many years since the Columbian quincentennial have produced new insights into Postclassic and Colonial agriculture. Now the time is ripe for a new synthetic assessment. This chapter will therefore use a regional perspective and new data, both archaeological and historical, to arbitrate the unresolved disputes regarding the nature of Aztec agriculture and its transformation under the influence of Spanish colonialism. Prior Reconstructions of the Aztec Landscape: A Critique There is widespread agreement that, on the eve of Spanish Conquest, central Mexico had reached an unprecedented demographic climax, unsurpassed for several centuries afterwards. Aztec Mexico was probably the most densely settled part of the American continent.5 The extent of its cultivated fields and the intensity and sophistication of the agricultural practices could be rivaled only by those of the Inca Empire in the Andes and coastal Peru.6 The scientific study of Aztec agriculture began in earnest in the mid-twentieth century, with a group of archaeologists and anthropologists with Marxist or neoevolutionary leanings who professed an interest in the material “infrastructure” of prehispanic societies. The group’s oldest member was Pedro Armillas, joined by Ángel Palerm, Eric Wolf, and William Sanders.7 Sanders deserves credit for identifying most explicitly the three pathways of agricultural intensification available in central Mexico and exploited to full advantage by the Aztecs: agricultural terracing, canal irrigation, and raised (or drained) fields in natural wetlands.8 The evidence for the workings, location, extent, and productivity of this triad, on which both these pioneers and their successors drew, included, in decreasing order of importance, ethnographic analogy with agriculture in similar settings in the twentieth century, the backcasting of information contained in sixteenth- and seventeenth-century documents, and archaeological fieldwork.9 The latter was predominantly surface archaeology, which involved the prospection and mapping — often in conjunction with archaeological settlement surveys — of those elements of the agricultural landscape that archaeologists encountered in a derelict state or, if still in use, those they deemed to be of prehispanic origin. Excavation of some of those agricultural features took place as the main wave of settlement surveys in the Basin of Mexico drew to a close around 1980, so excavation results did not feature too prominently in the grand syntheses written at the time.10 Developers and archaeologists uncovered several more terraces, canals, and wetland fields in the 1980s, 1990s, and 2000s, but these discoveries drew relatively little attention from the great synthesizers of this more recent age who often scorn “culture ecological” themes and the description-heavy and predominantly Spanish-language output of salvage projects.11 Targeted excavation of agricultural features of (suspected) Aztec age remains a rare occurrence. The geographical coverage achieved so far is thus very thin and very unequal. The greater part of archaeological information on Aztec agriculture comes from the Basin of Mexico, followed distantly by Puebla-Tlaxcala. In the other two major regions of the Aztec realm, Morelos and the Valley of Toluca, we can probably aspire to the title of pioneers in the archaeology of Aztec agriculture. We share, or at least value, the anthropological and neoevolutionary mindset of most scholars who have written on Aztec agriculture but are at the same time aware of its distortions. Our predecessors have been fond of treating traditional farming practices as a “toolkit” from which social evolution chooses according to the demands of population growth, ambitious elites, or other variables beyond subsistence. Yet, because they unanimously placed Aztec society at the apex of all evolutionary trajectories, they tended to assume that Aztec farmers always chose the most sophisticated agricultural technology appropriate to a particular agroecological niche. At its worst, this mindset brushed aside any notion of technological change or regional variation conditioned by nonecological factors. To those not particularly keen on fieldwork, it seemed to obviate the need to seek archaeological evidence indicating the type of agriculture practiced at any specific locale specifically during the Aztec period. The grave danger of such an approach is that it can blind us to forms of agriculture that were overlooked or did not exist in the ethnographic and ethnohistoric past. It also denies us the chance to put agriculture in social and historical context, which obviously varied from one Aztec citystate to another and changed between the twelfth and sixteenth centuries.12 The text that Whitmore and Denevan wrote for the Columbian quincentennial is a good example of the kind of synthesis of Aztec agriculture that relies heavily on analogy and the publications of Sanders and his associates on the Basin of Mexico Survey. Table 9.5. (cont’d.) Studies most relevant to the assessment of the Melville-Butzer controversy. Id. Region Areas of most intensive study Method (geoarch./ Most severe archival) degradation 22 El Oriental Aljojuca maar geoarch. Colonial M archival both M B 23 Libres y Perote regional coverage 24 Central Veracruz regional archival coverage; geoarch. along Antigua river 25 Mixteca Alta Coixtlahuaca 26 Mixteca Alta Cerro Jazmín, Nochixtlan Valley 27 Mixteca Alta Yanhuitlan, Nochixtlan Valley B/M References Colonial, Modern pre-“Aztec” both Colonial geoarch. pre-“Aztec,” Colonial M M geoarch. pre-“Aztec,” Colonial M Bhattacharya et al. 2015; Bhattacharya and Byrne 2016 Gerez Fernández 1985 Sluyter and Siemens 1992; Sluyter 1997a, b, 1998, 1999 2002 Rincón Mautner 1999 Pérez Rodríguez et al. 2011; Pérez Rodríguez and Anderson 2013; Pérez Rodríguez 2015 Mueller et al. 2012; Joyce et al. 2012 The “B/M” column indicates whether the results are easier to reconcile with the narratives of Melville (M) or the Butzers (B); the former emphasizing significant land degradation in the early Colonial period and the negative impacts of pastoralism; the latter emphasizing the opposite and tending to place the most significant degradation in prehispanic times. The interpretation and verdict are ours. The quality of the supporting data is variable and the verdict consequently precarious in many cases. “Aztec” outside central Mexico refers to the interval equivalent to the Aztec period of AD 1150–1520. northern frontier of Mesoamerica AD 1520 modern state boundaries 4 5 Butzer 6 3 Melville 7 15 11 12 13 16 9 10 19 20 18 17 21 24 25 27 26 2000 500 m a.s.l. study 1 (S. Cook) 23 22 14 100 km Figure 9.4. Studies most relevant to the assessment of the Melville-Butzer controversy (see Table 9.5). Geoarch. Archival 8 CHAPTER 10 Conclusions We have presented two separate summaries of the profound landscape changes that accompanied the metamorphosis of the city-state capital of Matlatzinco into the modern village of Calixtlahuaca. We placed these summaries, based respectively on geoarchaeological and documentary evidence, at the end of Chapters 6 and 7. Rather than offer a third chronologically ordered reconstruction of events, we now return to some of the broader issues raised in the introduction, briefly reviewing what we believe to be the most salient contributions of our research. On one hand, we see our contributions adding to the methodology for studying terraced landscapes, and on the other adding to debates on the legacy of colonial encounters for those members of society whose livelihoods depended on the improvement or degradation of farmland. Methodological Insights We believe our explorations at Calixtlahuaca to be one of the most intensive, site-specific archaeological studies of agricultural and residential terracing in Latin America. Other studies surpass ours in regional coverage;1 in the scope and detail of mapping surface-visible remains of terrace risers and associated features;2 in sophistication of classifications of surface soils on currently or formerly terraced slopes;3 or in mapping and modeling erosive processes.4 We think our project stands out by virtue of the breadth and depth of stratigraphic exposure; the degree of chronological control (see Table 9.1 for both); and the amount of effort put into postexcavation analysis of soil and sediment samples, artifacts, and documentary clues. All four authors of this book have worked extensively on terraced landscapes in other places, including Tlaxcala, Morelos, Oaxaca, and Greece.5 Reflecting on both the strengths and shortcomings of the Calixtlahuaca project (and with the other projects at the back of our minds), we would encourage any fellow archae- ologist interested in understanding a terraced landscape to keep in mind the following exhortations: 1. Excavate. Studies based exclusively on surface evidence result, at best, in detailed but ahistorical descriptions of the last or most conspicuous phase of terracing, which emphasize the supposedly unique and “logical” adaptation to the regional peculiarities of the natural environment. Much more often, they assign erroneous ages to the terraces on the basis of spurious associations with surface scatters of artifacts or with nearby settlement sites. Except where standing architecture or easily datable foundations occupy intact treads, giving a ter­ minus ante quem for terrace construction, we regard such age assignments with utmost suspicion. Studies that rely heavily on surface survey tend to exaggerate the length of terrace occupation, missing episodes of decay and abandonment. Without excavation, relatively little can be said about construction technique, less still about labor inputs or the effect of terracing on agricultural outputs. How far surface archaeology would have taken us at Calixtlahuaca can be judged by referring to the closing section of Chapter 4. Without excavation, we would not know that the Aztec-period terraces were built on a slope significantly altered by previous farming activity and that it involved the labor-demanding importation of earth selected for its textural properties rather than simple cut-and-fill. We would also have no conclusive evidence of the catastrophic terrace decay and degradation of farmland during the Colonial period. At the outset of the project our best intimation that something of that nature had indeed occurred came not from surface observations but from García Payón’s 1930s excavations of the palace buried at the foot of Cerro Tenismo. Worst of all, before fieldwork began in earnest, we were tempted to dismiss nineteenth- and twentieth-century activity on the slope as simple rehabilitation of certain 272 Conclusions parts of the Aztec system. We were ready to apportion a major share of our resources to producing a site-wide map of the maze of stone walls visible on the surface, hoping to interpret it in terms of Aztec urban planning, to delimit the different neighborhoods of Matlatzinco, and to model the movement of people through this built environment. We aborted that avenue of research as the first excavation units revealed the complexity and depth of transformations that the slope had undergone in the last 500 years. 2. Excavate a lot. Because terraces are landscape-scale features, it is difficult to make sense of their stratigraphy by means of a few “telephone-booth” test pits. On slopes that have experienced multiple cycles of terrace construction and decay, the preservation of the older phases of terracing and associated occupation and cultivation surfaces tends to be very fragmentary, often following narrow slope-perpendicular belts (Figure 5.9). Moreover, as one moves in the slope-parallel direction, there tend to be drastic differences in the nature and fertility of soils (Figure 8.2). In order to maximize the chances of discovering ancient terracing and documenting it in all its diversity, we therefore like to start by digging slopeparallel trenches that span the full width of at least one of the treads visible at the modern ground surface. We implemented this strategy successfully at several of the locations discussed in Chapter 5, discovering buried terrace risers and houses of which there was no hint at the modern ground surface. Where terrace maintenance has lapsed, there are usually many “opportunistic” exposures of stratigraphy in gullies, roadcuts, or behind risers from which stone has been robbed (Figure 3.9.). It is tempting to do just enough digging to create a vertical face around such an exposure, thus minimizing the volume of earth to be moved. The downside of this approach is that the distribution of opportunistic exposures in a terraced landscape is not random. They are most frequent along the lateral edges of terrace treads, often occupied by slope-parallel access routes and water conduits and along risers. Both loci are often beyond the edge of the cultivated surface and atypical of the tread as a whole. Where risers have been refurbished, the part exposed just behind the stone wall is often a wedge of unstratified earth of mixed provenience, emplaced more recently than the rest of the terrace fill (Figures 4.17; 5.6, Zone 111; 5.26c). While we do not disdain opportunistic exposures, we advocate that they be approached selectively and enlarged into proper excavation units whenever possible. In the end, there is no escaping the fact that terraced landscapes are rarely amenable to “surgical” interventions.6 273 3. Target downslope portions of treads. Whatever the construction technique, terrace fill comes in wedges that thicken downslope (Figure 3.5). The front, or downslope belt of each tread tends to be the place where multiple phases of terracing can be distinguished from one another. They are often the only place where vestiges of earlier phases of terracing are preserved (Figures 5.8, 5.9). Where choices are to be made, especially after digging the initial tread-spanning trench, it often makes sense to target the front of a tread. Paradoxically, this often means digging precisely where the surface densities of ancient artifacts are the lowest. It is at the back of a tread where tillage and thinning terrace fill tend to concentrate artifacts at the modern ground surface, typically comminuted, commingled, and worn by multiple episodes of redeposition. At the front of the same tread, artifacts are likely to be buried at depth, larger, less battered, and, most importantly, closer to their original place of discard. There are two caveats to this proposition. Houses — among the most reliable indicators of terrace age and often targets of archaeological exploration in their own right — tend to be placed at the back of a terrace tread. Where the location of risers has shifted up and down a slope, or where their spacing has changed, it is difficult to anticipate where the front and back of former treads lay (though subtle changes of slope gradient sometimes give a hint; Figures 4.7, 4.10–4.14). Some of the best opportunities to study ancient houses are to be found precisely in situations where the changing morphology and placement of a riser makes it encroach onto the upslope portion of the tread of a previous terracing phase (Figure 5.26). 4. Cut across terrace risers. To some extent, this is simply an extension of the previous point. Apart from this, contrasts in the type of masonry or other morphological characteristics of the riser, as well as subtle shifts in alignment often provide the best clues to the different phases of terracing (Figures 5.14, 5.25, 5.26, 5.28). Cutting across a riser is usually necessary to note such changes, and to observe if and how different courses or levels of the riser articulate with different wedges of terrace fill. Of course, targeting the very front of a terrace during excavation and cutting across the riser is the perfect recipe for undermining the stability of the extant terrace. Moreover, it often requires the felling of fruit trees growing along the leading edge or removal of other perennial crops that took years to mature (a few magueyes fell victim to our excavations at Calixtlahuaca). Farmers, landowners, and conservationists understandably do not like the idea, but it is well worth a negotiation, a payment for damages, or even risking somebody’s ire. References * Works produced by the Calixtlahuaca Archaeological Project Albores Zárate, Beatriz 1985 El desplazamiento de las lenguas indígenas en la antigua zona lacustre del Alto Lerma. Cuicuilco 16(1):23–35. 1995 Tules y sirenas: el impacto ecológico y cultural de la industrialización en el Alto Lerma. Colegio Mexiquense, Zinacantepec. 1998 Origen pre­mexica de las chinampas de la zona lacus­ tre del Alto Lerma mexiquense. Colegio Mexiquense, Zinacantepec. 2011 Pasado lacustre y cambio económico en el Alto Lerma Mexiquense: dos vías de su industrialización. In Historia general ilustrada del Estado de México, Vol. 6, El periodo institucional (1930–2005), edited by Paolo Riguzzi and Luis Jaime Sobrino Figueroa, pp. 273–301. Colegio Mexiquense, Zinacantepec. Alcántara Gallegos, Alejandro 2004 Los barrios de Tenochtitlan. Topografía, organización interna y tipología de sus predios. In Historia de la vida cotidiana en México, Vol. 1, Mesoamérica y los ámbitos indígenas de la Nueva España, edited by Pablo Escalante Gonzalbo, pp. 167–198. FCE, CDMX. Alexander, Rani T. 2004 Yaxcabá and the Caste War of Yucatán: An Archaeo­ logical Perspective. University of New Mexico Press, Albuquerque. Aliphat Fernández, Mario, and Gerd Werner 1994 The Tepetates of the Mexican Central Highlands: Prehispanic and Modern Impact of Agriculture and Water Management. In Transactions of the 15th World Congress of Soil Science, Acapulco, Mexico, July 1994, Vol. 6a, pp. 528–540. Sociedad Mexicana de la Ciencia del Suelo, Chapingo. Allan, William 1965 The African Husbandman. Barnes and Noble, New York. Allen, Bryant J. 2001 Boserup and Brookfield and the Association between Population Density and Agricultural Intensity in Papua New Guinea. Asia Pacific Viewpoint 42: 237–254. Allen, Bryant J., and Chris Ballard (eds.) 2001 Beyond Intensification? Reconsidering Agricultural Transformations. Special Section. Asia Pacific View­ point 42:157–333. Abascal Macías, Rafael 1980 Riego y control de agua en los cultivos prehispánicos de Tlaxcala. Revista mexicana de estudios antro­ pológicos 26:115–167. Abascal Macías, Rafael, and Ángel García Cook 1975 Sistemas de cultivo, riego y control de agua en el área de Tlaxcala. In Balance y perspectiva de la antropología de Mesoamérica y del norte de México, Arqueología 1, pp. 199–212. Sociedad Mexicana de Antropología, CDMX. Aeppli, Hans, and Ernst Schönhals 1975 Los suelos de la cuenca de Puebla­Tlaxcala: investiga­ ciones acerca de su formación y clasificación. Steiner, Wiesbaden. Aguilar Robledo, Miguel 2009 Contested Terrain: The Rise and Decline of Surveying in New Spain, 1500–1800. Journal of Latin American Geography 8:23–47. Aguirre Beltrán, Gonzalo 1967 Regiones de refugio: el desarrollo de la comunidad y el proceso dominical en Mestizoamérica. Instituto Indigenista Interamericano, CDMX. Ahrndt, Wiebke (ed.) 2001 Edición crítica de la “Relación de la Nueva España” y de la “Breve y sumaria relación” escritas por Alonso de Zorita. INAH, CDMX. Aimers, James J. 2007 What Maya Collapse? Terminal Classic Variation in the Maya Lowlands. Journal of Archaeological Research 15:329–377. Aiuvalasit, Michael J., James A. Neely, and Mark D. Bateman 2010 New Radiometric Dating of Water Management Features at the Prehistoric Purrón Dam Complex, Tehuacán Valley, Puebla, México. Journal of Archaeo­ logical Science 37:1207–1213. Alanís Boyso, José Luis 1976a Corregidores de Toluca: apuntes para su estudio, 1590–1810. BEEM, Toluca. 1976b Corregimiento de Toluca, pueblos y elecciones de república en el siglo XVIII. Historia mexicana 99:455–477. Alanís Boyso, José Luis, Leopoldo Sarmiento Rea, and Rodolfo Alanís Boyso 1978 Archivo General del Estado de México: catálogo y síntesis de documentos manuscritos relativos a pueblos del Estado de México, 1542–1823. Oficialía Mayor del Gobierno, Toluca. 319 320 References Ammerman, Albert J., Gary D. Shaffer, and Nicholas Hartmann 1988 A Neolithic Household at Piana de Curinga, Italy. Journal of Field Archaeology 15:121–140. Anawalt, Patricia R. 1981 Indian Clothing Before Cortés: Mesoamerican Costumes from the Codices. University of Oklahoma Press, Norman. Anderson, J. Heath 2009 Prehispanic Settlement Patterns and Agricultural Production in Tepeaca, Puebla, Mexico, aD 200–1519. PhD dissertation, Pennsylvania State University, University Park. Andrews, Anthony P. 2012 Historical Archaeology in the Maya Area: A Working Bibliography. Electronic document, sites.ncf.edu /andrews, accessed April 8, 2020. Anselmetti, Flavio S., Daniel Ariztegui, Mark Brenner, David A. Hodell, and Michael F. Rosenmeier 2007 Quantification of Soil Erosion Rates Related to Ancient Maya Deforestation. Geology 35:915–918. Aranda Sánchez, José María 2000 Conformación de la zona metropolitana de Toluca, 1960–1990. UAEM, Toluca. 2011 La urbanización, 1930–2005. In Historia general ilustrada del Estado de México, Vol. 6, El periodo in­ stitucional (1930–2005), edited by Paolo Riguzzi and Luis Jaime Sobrino Figueroa, pp. 99–129. Colegio Mexiquense, Zinacantepec. Armillas, Pedro 1971 Gardens on Swamps. Science 174:653–661. 1987 El paisaje agrario azteca. In La aventura intelectual de Pedro Armillas: visión antropológica de la historia de América, edited by José Luis Rojas, pp. 66–107. Colegio de Michoacán, Zamora. Armillas, Pedro, Ángel Palerm, and Eric R. Wolf 1956 A Small Irrigation System in the Valley of Teotihuacan. American Antiquity 21:396–399. Arnáez, José, Noemí Lana Renault, Teodoro Lasanta, Purificación Ruiz Flaño, and Juan Castroviejo 2015 Effects of Farming Terraces on Hydrological and Geomorphological Processes: A Review. Catena 128:122–134. Arroyo Abad, Leticia, Elwyn Davies, and Jan Luiten van Zanden 2012 Between Conquest and Independence: Real Wages and Demographic Change in Spanish America, 1530– 1820. Explorations in Economic History 40:149–166. Assadourian, Carlos Sempat 1991a Estructuras indígenas en transición. In Tlaxcala, una historia compartida, Vol. 9, Siglo XVI, edited by Andrea Martínez Baracs and Carlos Sempat Assadourian, pp. 69–148. Gobierno del Estado, Tlaxcala. 1991b Estructuras económicas coloniales: el sistema de las haciendas. In Tlaxcala, una historia compar­ tida, Vol. 10, Siglos XVII–XVIII, edited by Andrea Martínez Baracs and Carlos Sempat Assadourian, pp. 13–90. Gobierno del Estado, Tlaxcala. Assadourian, Carlos Sempat, and Andrea Martínez Baracs (eds.) 1991a Tlaxcala, textos de su historia, Vol. 6, Siglo XVI. Conaculta, CDMX. 1991b Tlaxcala, textos de su historia, Vols. 7–8, Siglos XVII– XVIII. Conaculta, CDMX. Aston, Michael 1985 Landscape Archaeology and Local History. Routledge, London. Athens, J. Stephen 1999 Comments on “Intensification in the Pacific: A Critique of the Archaeological Criteria and Their Application,” by Helen M. Leach. Current Anthropol­ ogy 40:321–322. Ávila López, Raúl 1991 Chinampas de Iztapalapa, D.F. INAH, CDMX. 1995 Excavaciones arqueológicas en San Gregorio Atlapulco, Xochimilco. Report on file at the Archivo Técnico, INAH, CDMX. 1998 Investigaciones del Proyecto Arqueológico San Luis Tlaxialtemalco. 6 vols. Report on file at the Archivo Técnico, INAH, CDMX. 2006 Mexicaltzingo: arqueología de un reino culhua­ mexica. 2 vols. INAH, CDMX. Avni, Gideon, Naomi Porat, and Yoav Avni 2013 Byzantine-Early Islamic Agricultural Systems in the Negev Highlands: Stages of Development as Interpreted through OSL Dating. Journal of Field Archaeology 38:332–346. Avni, Yoav, Naomi Porat, and Gideon Avni 2012 Pre-Farming Environment and OSL Chronology in the Negev Highlands, Israel. Journal of Arid Environ­ ments 86:12–27. Avni, Yoav, Naomi Porat, Joseph Plakht, and Gideon Avni 2006 Geomorphic Changes Leading to Natural Desertification versus Anthropogenic Land Conservation in an Arid Environment, the Negev Highlands, Israel. Geomorphology 82:177–200. Bakels, Corrie 1996 Growing Grain for Others or How to Detect Surplus Production. Journal of European Archaeology 4:329–336. Balkansky, Andrew K. 1998 Urbanism and Early State Formation in the Huamelulpan Valley of Southern Mexico. Latin American Antiquity 9:37–67. 2006 Surveys and Mesoamerican Archaeology: The Emerging Macroregional Paradigm. Journal of Ar­ chaeological Research 14:53–95. Balkansky, Andrew K., Verónica Pérez Rodríguez, and Stephen A. Kowalewski 2004 Monte Negro and the Urban Revolution in Oaxaca, Mexico. Latin American Antiquity 15:33–60. Ballesteros Arias, Paula, Felipe Criado Boado, and José M. Andrade Cernadas 2006 Formas y fechas de un paisaje agrario de época medieval: a Cidade da Cultura en Santiago de Compostela. Arqueología espacial 26:193–225. The Geoarchaeology of a Terraced Landscape From Aztec Matlatzinco to Modern Calixtlahuaca Aleksander Borejsza, Isabel Rodríguez López, Charles D. Frederick, and Michael E. Smith “Employing geoarchaeology with a healthy dose of ethnohistory, this book uses a terraced hill in central México to demonstrate—masterfully—how to unravel the subtle nuances and complexities of the creation of landesque capital. It is a major contribution to Mesoamerican studies and stands to revolution- The toil of several million peasant farmers in Aztec Mexico transformed lakebeds and mountainsides into a checkerboard of highly productive fields. This book charts the changing fortunes of one Aztec settlement and its terraced landscapes from the twelfth to the twenty-first century. It also follows the progress and missteps of a team of archaeologists as they pieced together this story. Working at a settlement in the Toluca Valley of central Mexico, the authors used fieldwalking, excavation, soil and artifact analyses, maps, aerial photos, land deeds, and litigation records to reconstruct the changing landscape through time. Exploiting the methodologies and techniques of several disciplines, they bring context to eight centuries of the region’s agrarian history, exploring the effects of the Aztec and Spanish Empires, reform, and revolution on the physical shape of the Mexican countryside and the livelihoods of its people. Accessible to specialists and nonspecialists alike, this well-illustrated and well-organized volume provides a step-bystep guide that can be applied to the study of terraced landscapes anywhere in the world. The four authors share an interest in terraced landscapes and have worked together and on their own on a variety of archaeological projects in Mesoamerica, the Mediterranean, Poland, and the United Kingdom. ize the field of landscape archaeology.” —William E. Doolittle, professor, Department of Geography and the Environment, The University of Texas at Austin “This is a milestone publication in geoarchaeology and a book that every serious ecologically oriented archaeologist needs to read and have on their bookshelf for consultation in future fieldwork. The authors employ a multi-disciplinary approach to the study of terraced landscapes, integrating ethnohistoric, ethnographic, archival research, and standard archaeological survey and excavation, with geoarchaeological investigation of both prehispanic and modern landscapes. It is one of the Aleksander Borejsza is professor of archaeology at the Universidad Autónoma de San Luis Potosí. Isabel Rodríguez López works as a freelance archaeologist in Mexico. Charles D. Frederick is a consulting geoarchaeologist and research fellow at the University of Texas at Austin. He is the coeditor of Landscape and Land Use in Postglacial Greece (2000). Michael E. Smith is professor at the School of Human Evolution and Social Change at Arizona State University. He is the author and editor of several books on the Aztecs, including Aztec City-State Capitals (2008) and The Postclassic Mesoamerican World (2003). most comprehensive and scholarly treatments on the subject available today.” —Kenneth Hirth, professor of anthropology, Pennsylvania State 392 pp., 8½ x 11 University and author of Obsidian Craft Production in Ancient 92 illustrations, 7 maps Central Mexico ISBN 978-1-64769-022-9 Hardcover $85.00 To order: phone: 800-621-2736 fax: 800-621-8476 email: orders@press.uchicago.edu online: www.uofupress.com For publicity inquiries, contact Hannah New, Marketing Manager, 801-585-9786 or hannah.new@utah.edu www.UofUPress.com